
[{"content":"We have three values. I wasn\u0026rsquo;t always a fan of lab values, but the more I\u0026rsquo;ve worked in larger organizations, the more I\u0026rsquo;ve seen how useful it is to pick a few on purpose. We could list dozens. Choosing three gives us focus.\nHigh-quality work. A good way to think about quality is usefulness. Doing genuinely useful work is really hard, and that\u0026rsquo;s what we should strive for.\nBalance. This runs along several axes. It means learning to say yes and no to things and setting boundaries, including with me. It also means balancing hard science with how that science exists in the world. How do we communicate our results? For software especially, how do we make sure it\u0026rsquo;s useful and used? There\u0026rsquo;s a lot of great software out there that nobody uses.\nCollaboration. Science and engineering are human practices as much as quantitative ones. How effective our work is depends on how it engages with people. So we\u0026rsquo;re social, we\u0026rsquo;re generous with credit, and we\u0026rsquo;re fair. Being generous takes intent, and you often have to create those opportunities by inviting other people to work with you.\nUniversity is a unique time. You\u0026rsquo;re surrounded by passionate people, and that rarely happens again in the same way. Make the most of it.\nEvery student is different, so treat this handbook as a set of defaults we can adapt together.\nIf there\u0026rsquo;s a problem # Disagreements and friction are a normal part of any working relationship, and I want to hear about them. If you can, start with me directly. If that doesn\u0026rsquo;t feel possible, or doesn\u0026rsquo;t resolve things, here are other options, roughly from informal to formal:\nYour supervisory committee members ChemE\u0026rsquo;s Associate Chair, Graduate Studies [TBD] SGS Centre for Graduate Mentorship \u0026amp; Supervision (CGMS): confidential support, facilitated conversations, and mediation UTGSU, the graduate students\u0026rsquo; union [TBD]: confidential advocacy The University Ombudsperson [TBD]: for concerns not resolved through regular channels You never need my permission to use any of these, and using them won\u0026rsquo;t change how I treat you. SGS also has a guide to conflict resolution in graduate studies.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/values-and-culture/","section":"How We Work 🚧","summary":"","title":"1. Values \u0026 Culture","type":"culture"},{"content":"We have become very good at modeling chemical data, but a prediction only matters once it turns into a real molecule or material. Self-driving labs close that gap by pairing robots with models that decide what to try next. Automation also improves the data itself: cheaper, more consistent and easier to replicate. U of T has one of the densest clusters of SDLs anywhere. We expect the line between experimental and computational chemists to fade. The chemist of the future will program a robot to run experiments while modeling the chemistry behind them.\n","externalUrl":null,"permalink":"/research/topics/self-driving-labs/","section":"Research","summary":"","title":"Self-Driving Laboratories","type":"research"},{"content":"Your first week is mostly about getting set up. We try to provide every student with a laptop and a computer screen so you can work comfortably from day one.\nU of T accounts [TBD] Join the lab GitHub organization: chemcognition-lab Zulip [TBD] Compute cluster accounts [TBD] Shared drive [TBD] Go-links [TBD] Set up your laptop and cluster access with hpc_setup Start your first project from the project template Look through our agent skills for AI coding tools Questions can go to the #TechSupport channel in Zulip. The source for this handbook lives in the lab website repo [TBD] (members only), so you can propose edits there.\nWhen you leave # Code lives in the lab GitHub organization and data lives on lab storage from the start of a project. When you leave, your repositories and data stay with the lab so others can build on your work. You\u0026rsquo;re welcome to keep copies of your own code and to open-source it after publication, unless a collaborator or industry agreement says otherwise.\nBefore your last day:\nMake sure every repo has a README explaining how to reproduce your main results Hand off anything another student depends on Papers still in revision: [TBD] ","date":"2 November 2026","externalUrl":null,"permalink":"/culture/getting-started/","section":"How We Work 🚧","summary":"","title":"2. Getting Started","type":"culture"},{"content":"Whenever we care about how things relate, graphs are a natural language: objects become nodes and relationships become edges. Chemistry is full of relationships at every scale, from atoms bonding into molecules to enzymes, metabolic networks and even tequila production. How you set up the graph often decides whether a model succeeds, because it encodes the right priors. We study these choices and how to build them into graph neural networks and transformers alike.\n","externalUrl":null,"permalink":"/research/topics/graphs/","section":"Research","summary":"","title":"Graphs","type":"research"},{"content":"We are getting good at building models that predict. In science, though, a prediction is rarely the goal. We want the mechanism behind it. Explainable AI promises to read that mechanism out of a trained model, but an explanation is not a prediction. There is no test set to score it against, and a heatmap that looks chemically sensible can still be confidently wrong. Pulling knowledge out of information is hard, so we treat interpretability as a measurement problem rather than a visualization one. High predictive accuracy does not guarantee faithful explanations, and two models that score equally well can point to different mechanisms. Our hope is that reliable interpretability tools will let models do more than predict: they should help us form hypotheses, choose experiments and, in the end, do better science \u0026amp; engineering.\n","externalUrl":null,"permalink":"/research/topics/xai/","section":"Research","summary":"","title":"Interpretability (XAI)","type":"research"},{"content":"The expectations I care about most are consistency and communication. Working steadily and telling me where things stand matters much more to me than raw hours. It\u0026rsquo;s one of the most useful habits you\u0026rsquo;ll build during your training, and it keeps deadlines from piling up. Consistent work and open communication also let us iterate, which is how we get to high-quality results.\n1:1s and time with me # We meet weekly for a 1:1, and I try to book them on our shared in-person days (Wednesday and Thursday). The meeting is yours, so don\u0026rsquo;t cancel it just because progress feels slow. Some weeks we\u0026rsquo;ll take a walk and talk through early ideas, and other weeks we\u0026rsquo;ll dig into specifics. If you need more dedicated time, book it through my booking page [TBD].\nWhat you can expect from me # Day-to-day communication is Monday to Friday, roughly 9 to 5. If you send me a message on Zulip, I tend to respond much faster than by email. Don\u0026rsquo;t let me block you: if you need something unblocked, ping me. Don\u0026rsquo;t feel bad about it, really ping me.\nI keep hard boundaries around evenings and weekends, and I don\u0026rsquo;t expect messages or replies outside office hours. I want us to treat these research positions like real jobs and be professional about our time. That\u0026rsquo;s good training for the real world whether you become a professor or go into industry. Emergencies sometimes happen, and for those we have an explicit emergency contact policy in our shared spreadsheet [TBD].\nHow we work together shifts over time. Your first year is really about figuring out how we work together well, so I\u0026rsquo;ll give you more concrete, scoped projects. By the final years of a PhD (and right from the start as a postdoc), I expect you to work much more independently, pursuing your own ideas and leading things.\nAs a newer PI, I\u0026rsquo;m still learning how to juggle everything and stay on top of all the moving parts while making sure everyone gets good mentorship, room to grow, and feedback in a way that\u0026rsquo;s useful. If something isn\u0026rsquo;t feeling smooth, tell me.\nCoding critique # I recommend everyone do at least one coding session or code critique, with me or with other people in the lab. It shows you how someone else reads your code. Even with good AI tools, writing code yourself is worth practicing because it trains how you break problems into small, approachable tasks. For why working alongside others helps, see Codemates.\nBy role # PhD students. For a classic perspective on what a PhD actually is, see Matt Might\u0026rsquo;s The illustrated guide to a Ph.D. I expect at least three works that you lead. The first teaches you how to get into the field, the second builds up the difficulty, and the third should be ambitious enough to help define your direction after graduation. One natural way to make a project more ambitious is to connect it to experiments through collaborators. Leading three works across five years is a reasonable baseline, though doing more will help your career. See Career \u0026amp; Milestones for the year-by-year timeline.\nMaster\u0026rsquo;s students. An MASc is two years with a thesis. Your first year is really about scoping a clear, doable project so you can succeed at it, aiming for one work that you lead. Especially if you\u0026rsquo;re jumping somewhere else next, publishing something is a great public artifact that shows what you built. See Career \u0026amp; Milestones for the two-year timeline.\nUndergraduates. The aim is a well-scoped project, whether a summer position or a course project. Really, the goal is to give you a sense of what research is. If a publication comes out of it and helps you apply to grad school, that\u0026rsquo;s great.\nPostdocs. Publishing and leading ideas independently are expected, and you\u0026rsquo;re much more in control of your career. A big part of a postdoc is defining where you want to grow next, whether that\u0026rsquo;s academia, industry, or starting something, and making sure you get the right opportunities (like lecturing, grant writing, or mentoring). Let\u0026rsquo;s talk early about where you want to go so I know how best to support you.\nGrowing as a researcher # You probably won\u0026rsquo;t know how to propose a project at the start. It\u0026rsquo;s very common not to have ideas at the beginning, since you\u0026rsquo;re still learning the process. The hope is that by the end of your PhD you\u0026rsquo;ll know how, and you\u0026rsquo;ll have tons of ideas. Chris Olah\u0026rsquo;s research taste exercises are a great way to practice: write down ideas and have a mentor rate them.\nScience is open-ended discovery, which means sometimes things don\u0026rsquo;t work. Open-ended projects carry more risk, and that\u0026rsquo;s okay.\nMentoring and helping others is expected, but flexible. Some people want to mentor students and some don\u0026rsquo;t, and that\u0026rsquo;s fine. It\u0026rsquo;s always good practice, though: we\u0026rsquo;ll always be working with people, and learning how to do that well is something worth getting better at.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/roles-and-expectations/","section":"How We Work 🚧","summary":"","title":"3. Roles \u0026 Expectations","type":"culture"},{"content":"Modeling is rarely the hardest part anymore. More often the bottleneck is framing the right problem and finding the data to answer it. Data-centric AI starts from a simple observation: progress happens one dataset at a time. This matters more in chemistry than in images or text, because experimental data is scarce, expensive and scattered across labs. A large part of our work is curating, generating and releasing datasets and benchmarks that let the field measure real progress.\n","externalUrl":null,"permalink":"/research/topics/datacentric-ai/","section":"Research","summary":"","title":"Data-Centric AI","type":"research"},{"content":"Perfumes, shampoos, foods and fuels are all mixtures, yet most machine learning in chemistry still studies one molecule at a time. Mixtures decide how drugs dissolve, how fuels ignite and how electrolytes move ions. The space of possible blends explodes with each added component, so exhaustive experiments are out of reach. We study how to represent mixtures well: which symmetries to build in, whether one representation can serve many tasks, and when physics helps.\n","externalUrl":null,"permalink":"/research/topics/mixtures/","section":"Research","summary":"","title":"Mixtures","type":"research"},{"content":" In person # I expect you to be in the lab in person at least three days a week. If you need more flexibility, just tell me. That\u0026rsquo;s a conversation we should have. Three days matters because you want to engage with other people, and other people have other ideas. One of the most valuable things you\u0026rsquo;ll get out of your training is the people you meet.\nOur shared in-person days are Wednesday and Thursday, so everyone overlaps. I also encourage you to set aside one day a week to catch up with people over coffee or lunch and learn what others are working on. Talking with people is a real part of doing research.\nMaker time and people time # Paul Graham\u0026rsquo;s essay on the maker\u0026rsquo;s schedule and the manager\u0026rsquo;s schedule argues that a single meeting can ruin a whole day of focused work. I mostly believe it. Part of your training is learning when to be around people and when to protect quiet time for focused work, and I try to do the same outside our shared days. When you do have meetings, try to stack them on the same day.\nHours and communication # Treat this like a job: work during the week, roughly 9 to 5. Tell people your communication preferences.\nDay-to-day communication happens on Zulip. Use public channels and topics when others might benefit from the answer, since a question you have is often one someone else has too. Zulip is asynchronous, and no one is expected to reply outside working hours.\nMeetings # We have a group meeting every week, and it\u0026rsquo;s the one recurring event to prioritize. We typically do 5-minute personal updates, and sometimes someone gives a longer talk (20 to 25 minutes plus discussion). It\u0026rsquo;s a good place to practice an upcoming conference talk or Research Committee Meeting (RCM) presentation. Once a quarter we take a meeting-free week for uninterrupted work, announced at the start of each term and planned around major deadlines.\nSprints and hackathons # I recommend them. Small, dedicated blocks of time working with other people are a really good way to learn a problem space quickly. Keep them to a single day, and don\u0026rsquo;t over-invest in how they\u0026rsquo;re judged.\nExternal collaborations # Align early on what success looks like and how data ownership works (a shared kickoff doc helps). Deadlines carry more weight with external collaborators, so if plans change or a timeline slips, let people know well in advance. Collaborations are also a great way to meet people in neighboring fields and travel to visit other labs.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/working-together/","section":"How We Work 🚧","summary":"","title":"4. Working Together","type":"culture"},{"content":"Machine learning can see and hear, but it has barely started to smell. Olfaction is our guide to the chemical world, and digitizing it means understanding how small molecules become percepts. In 2019 we sketched a map of odor space, then spent three years testing it: with human panels smelling new molecules, in the search for mosquito repellents, and against metabolism. Much of what we call flavor is retronasal smell, so food is part of the story too.\n","externalUrl":null,"permalink":"/research/topics/olfaction/","section":"Research","summary":"","title":"Olfaction \u0026 Sensory AI","type":"research"},{"content":" Problem first # We prioritize in this order: the problem, then the data, then the model. A clever model on a problem nobody cares about, or on data that can\u0026rsquo;t support it, won\u0026rsquo;t get far. When you describe a project, work through it in that order: the problem, the ML task, the data, the modeling, and how you\u0026rsquo;ll evaluate it.\nNew ideas # I\u0026rsquo;m open to research ideas. Sell me on them: why is it worth the time, and is it feasible? Anything can become a research paper. The work is finding an interesting way of telling the story. Common failure modes are getting too attached to one direction, not knowing the field well enough yet, and working in an environment that doesn\u0026rsquo;t support your interests. Talking to other researchers about their taste helps with all three.\nDe-risk your ideas # Learning to de-risk ideas is one of the most useful research skills you can build. The idea is \u0026ldquo;kill fast, learn fast\u0026rdquo;: test the core assumptions quickly, before investing months.\nOne exercise: list the likely bottlenecks, prioritize them, commit a fixed amount of time to exploring each, then decide whether to continue or punt. For example, if you want to fine-tune a large model, first time a single batch of your largest model on fake data. Does it fit in memory? Multiply by the number of batches and a guess for epochs. Is the time to train one model reasonable?\nBalancing multiple projects # Once you\u0026rsquo;ve ramped up, I expect you to juggle two or three projects at a time. Research has natural ebbs and flows, whether you\u0026rsquo;re waiting on compute, collaborators, or paper reviews, and having more than one project helps fill those gaps. Still, there is real value in focusing on one thing at a time, and we try to walk that tightrope together.\nDesign docs # Any time you\u0026rsquo;re about to spend more than a week on something, write a short design doc first covering what you\u0026rsquo;re building, why, the alternatives, and how you\u0026rsquo;ll know it worked. Writing it down helps clarify your plan and gives others a chance to spot issues early. Good introductions: Design docs at Google and Things I learned at Google: design docs.\nThe project spreadsheet # We keep a project spreadsheet [TBD] with ideas, timelines, and who\u0026rsquo;s working on what. Seeing active and paused ideas in one place makes it easier to talk to people and find connections across projects.\nPivoting is normal # Recycling, punting, and pivoting on projects is normal. Sometimes ideas don\u0026rsquo;t work, and what matters is diagnosing that as early as possible.\nSometimes deadlines get met by sprinting. In general it\u0026rsquo;s better to plan ahead, but sprints happen. If you\u0026rsquo;re going to do one, be focused about it so you can actually meet the deadline.\nReproducibility and data # Our work should be reproducible, with good data management. I\u0026rsquo;m a big fan of Cookiecutter Data Science, an opinionated and sensible way of organizing projects, and our project template builds on it. A few of its opinions we follow:\nRaw data is immutable. Never edit or overwrite it: code should take raw data to final results. Data mostly stays out of git. GitHub rejects files over 100 MB, so keep data on lab or cloud storage. Use notebooks for exploration and communication, and move reusable logic into source files. Keep modeling organized: configs and hyperparameters in JSON or YAML, predictions and training logs saved alongside. Build from the environment up, with an environment file (we use Mamba). Be transparent with your work and show it: put it in a GitHub repo and make commits. We usually won\u0026rsquo;t look at your commits in detail until we\u0026rsquo;re publishing something. What matters more is that you have a consistent way of keeping track of your work. When a project is ready to turn into a paper, see Writing \u0026amp; Publishing.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/projects-and-research-practices/","section":"How We Work 🚧","summary":"","title":"5. Research \u0026 Projects","type":"culture"},{"content":"Proteins and enzymes do the chemistry of life. Enzymes speed up reactions by many orders of magnitude and chain together into metabolic networks that build everything a cell needs. Many of nature\u0026rsquo;s most interesting molecules come from biosynthetic gene clusters, sets of genes that encode an entire assembly line. We model enzymes, their kinetics and the networks they form. We are lucky to sit just below the BioZone, a huge collaborative lab space for bioengineering.\n","externalUrl":null,"permalink":"/research/topics/proteins/","section":"Research","summary":"","title":"Proteins \u0026 Enzymes","type":"research"},{"content":"Good papers take a lot of iteration. Getting something out is real work, so we want to start early, share rough versions quickly, and iterate many times.\nStory and figures first # I really believe in trying to nail down the story and your contributions first. Sometimes that\u0026rsquo;s hard during open-ended discovery when you don\u0026rsquo;t know exactly where a project is going yet, and that\u0026rsquo;s normal. Still, a paper is built out of 4, 6, or 8 figures (see Aesthetics for how we think about figure arcs), and nailing down what those figures are is what actually tells the story.\nThere\u0026rsquo;s no single formula, and what works for you may change over time, though here\u0026rsquo;s a sequence I recommend:\nMock up the figures first. I\u0026rsquo;m a big fan of doodling on paper and taking screenshots. You can also use GenAI image tools to mock up a rough sense of how a visual layout might look (one of the few areas where I think GenAI is genuinely useful for images). Build a skeleton of the story beats: Introduction, Methods, Experiments and Results, Discussion, and the main points and contributions we want to cover in each. Write a 2 to 3 sentence pitch for your idea. Write bullet points inside each section for what you expect to cover. Fill in text, and repeat. The most common mistake I see on a first paper draft is waiting until everything feels done and perfect before showing anyone. Show drafts really early. It\u0026rsquo;s okay if it\u0026rsquo;s rough: iterating together is how it gets good.\nFor tools, I recommend Google Docs for drafting (fast, simple, collaborative) with Paperpile for references: copy a DOI and let the software figure out the rest. When we\u0026rsquo;re close to submission, we export with BibTeX to LaTeX on Overleaf.\nYou can also write with other people. Pair writing works much like pair programming, since talking through your thoughts with someone else puts you in a different mode. Communal writing sessions add a bit of social accountability: set a concrete goal (\u0026ldquo;finish half a page of the introduction\u0026rdquo;), write for a fixed block of time with breaks, share progress at the end, and talk through whatever is making the section hard.\nResources:\nStep-by-Step Guide to Undergraduate Writing How to write a first-class paper (Nature) Preparing manuscripts: scientific writing for publication Rubber duck debugging and pair programming Strengthening the case for pair programming Pair writing U of T Writing Centres, a great resource Feedback, internal reads, and user journeys # I tend to give writing feedback directly on the document or during our weekly 1:1s. If I\u0026rsquo;m busy or we need a longer pass, book a dedicated slot on my booking page [TBD] just to go over the draft together.\nIdeal lead times once it\u0026rsquo;s clear you\u0026rsquo;re aiming for a deadline:\nWorkshops and conferences need at least two weeks, with multiple iterations. Journal papers need at least a month to polish figures and writing. Before a paper goes out, everyone should do internal reads with labmates. Presenting your draft story in our weekly group meeting is also one of the best ways to practice and catch holes early.\nIf the project involves software, run a user journey before release: ask someone unfamiliar with the codebase to spend 30 to 60 minutes running the main workflow you expect most people to use, and have them write down quick impressions in a doc of every spot that felt rough. AI agents make code much more replicable now, though we\u0026rsquo;re building software for people first, so we still need to feel out how a newcomer actually experiences it.\nAuthorship # Authorship is a conversation to have early and often. It can be uncomfortable, so it\u0026rsquo;s much better to be honest up front and not have regrets later. Having explicit author contribution statements (like CRediT) is great practice.\nI encourage multiple first authors. Spread the wealth and be generous, while staying fair: we don\u0026rsquo;t add people just because they could be there. To be generous, you have to create those opportunities in the first place. Co-first author order can be decided loosely (even if it shouldn\u0026rsquo;t matter, the first position does). The last author is usually the PI, reflecting funding and supervision. In collaborations, that\u0026rsquo;s a discussion between the PIs.\nWhere we publish # Choosing a venue is contextual and depends on the work and the deadlines. Our archival venues are AI conferences (such as NeurIPS, ICML, and ICLR) and journals at the intersection of chemistry and machine learning (such as Nature Machine Intelligence, Chemical Science, Digital Discovery, and Journal of Chemical Information and Modeling).\nBecause our work often gets read by both ML reviewers and bench chemists, write as simply and plainly as you can. Most people enjoy clear, simple writing: even when the ideas are technical, distill why we\u0026rsquo;re tackling the problem into straightforward language, and make sure anyone looking at a figure can tell right away what claim you\u0026rsquo;re making.\nML workshops (at NeurIPS, ICLR, or ICML) are great milestones and stepping stones toward a full paper. Just keep in mind that workshops are usually non-archival unless they announce otherwise.\nPreprints and scooping # I recommend always preprinting. It\u0026rsquo;s better for visibility, citations, and open access, and it\u0026rsquo;s a good milestone. Post when you\u0026rsquo;re close to submission and confident. The only risk I see is a false sense of being done.\nScooping happens because good ideas occur to several people at once, and independent work also serves as replication. A nine- or ten-page paper only captures one angle of a problem anyway, so there\u0026rsquo;s usually room for multiple perspectives on the same idea.\nReviews, rebuttals, and rejections # For peer review and rebuttals, block that time off on your calendar ahead of time (for AI conferences, the rebuttal window is usually about a week). Guard your time during rebuttals: you can\u0026rsquo;t overdo everything, and you can\u0026rsquo;t make everybody happy.\nPapers get rejected all the time for different reasons. Sometimes you get a reviewer who doesn\u0026rsquo;t vibe with your work, and sometimes the work just isn\u0026rsquo;t mature enough yet. Treat reviews as a chance to get feedback and sharpen the next version.\nCode, data, and explainers # When a paper is preprinted and accepted, make sure the project has a clean, replicable GitHub repository and that the data is easy for others to grab and use.\nI also really encourage writing a short blog tutorial or interactive explainer alongside a paper, and sharing it on LinkedIn or other socials. Ask yourself: what is the story you didn\u0026rsquo;t get to tell because of the academic paper format? AI tools make building interactive pieces easier now, though you still have to think carefully about how to distill the ideas. It\u0026rsquo;s a great exercise in telling your story in a way that\u0026rsquo;s more personable and a little bit more free.\n","date":"13 November 2026","externalUrl":null,"permalink":"/culture/writing-and-publishing/","section":"How We Work 🚧","summary":"","title":"6. Writing \u0026 Publishing","type":"culture"},{"content":"Small molecules are the workhorses of chemistry: drugs, dyes, fragrances, pesticides and the building blocks of materials. Estimates put the number of drug-like molecules around 10⁶⁰, far more than anyone could make or test. We build models that learn useful representations of molecules, including their quantum and stereoelectronic character, and use them to search this space with intent. The goal is to propose molecules worth making and to explain why the model chose them.\n","externalUrl":null,"permalink":"/research/topics/small-molecules/","section":"Research","summary":"","title":"Small Molecules","type":"research"},{"content":" Tools # Zulip for chat [TBD] Google Docs and Google Sheets for writing and planning (Microsoft Teams is also available through U of T) Zoom for meetings, sometimes Teams Go-links for short links [TBD] Linux, VS Code, and Jupyter notebooks for coding Mamba for Python environments Setup and clusters # Our hpc_setup repo covers the first things to set up on your laptop, Python environments, and how to connect to Canadian compute clusters such as Balam, Killarney, Trillium, and others. New projects start from the project template.\nAgentic coding # Agentic coding tools are good, and I encourage you to try them. They will keep developing, and we\u0026rsquo;ll be using them. Our agent skills repo collects the skills we\u0026rsquo;ve written for them.\nAgentic workflows work well when you already know what you\u0026rsquo;re doing. The tricky part is that during your training, you\u0026rsquo;re still learning how to do things, so make sure you still understand how the underlying pieces work.\nBe careful not to accrue too much technical debt with agentic workflows. Debt is useful when it lets us move quickly, but you have to pay it back eventually (Martin Fowler explains it well), especially if the codebase grows faster than your understanding of it.\nGenerative AI # The main thing to ask yourself is whether you understand your work. Sometimes the only way to understand something is to build or write it yourself.\nIt\u0026rsquo;s easy right now to produce a lot of output without really understanding it. Everybody wants to be productive, but you are responsible for any code and writing you put your name on.\nFor code, use AI freely, as long as you read, understand, and test what it produces. For writing, AI is fine for polishing and brainstorming, while the ideas and arguments should come from you. Follow each venue\u0026rsquo;s disclosure rules, and never use AI to generate citations. For confidentiality, don\u0026rsquo;t paste unpublished data, collaborator or industry material, or manuscripts you\u0026rsquo;re reviewing into tools that may retain them. When you use AI for writing, think about who\u0026rsquo;s reading. If it\u0026rsquo;s a person, people tend to like shorter, more concise writing. If it\u0026rsquo;s a machine, length matters less.\nBe careful of the AI aesthetic, in writing and in visuals. It\u0026rsquo;s very bland, and readers and reviewers can judge it unfavorably. Passing your text through an AI-tropes checklist like tropes.fyi helps catch common habits (like em-dashes everywhere). See also The AI design aesthetic that\u0026rsquo;s taking over the internet in The New Yorker.\nU of T\u0026rsquo;s AI guidelines apply to all of us.\nHow I think about AI is close to how Terence Tao frames it for mathematics (paper, blog post) and how Ted Chiang writes about it: Why A.I. isn\u0026rsquo;t going to make art and his conversation with Ezra Klein (summary). Also worth reading: No, artificial intelligence is not conscious in The Atlantic.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/computing-and-infrastructure/","section":"How We Work 🚧","summary":"","title":"7. Computing \u0026 Infrastructure","type":"culture"},{"content":"Molecules are everywhere, so their effects show up across many systems and scales: solar cells, solutions, crystals, metal-organic frameworks and batteries. Materials add a layer that molecular models miss, since performance depends on composition, structure and processing history. We carry ideas from molecules and mixtures into this setting and pair them with automated platforms, so predictions become real candidates, such as new organic laser materials.\n","externalUrl":null,"permalink":"/research/topics/materials/","section":"Research","summary":"","title":"Materials","type":"research"},{"content":"Having a visual identity matters, both for communicating research and for building your career. The lab has a style I\u0026rsquo;ve curated, inspired by U of T\u0026rsquo;s brand colors plus some personal choices.\nDuring your training you\u0026rsquo;ll want to develop your own taste. Browse Google Fonts and pick what you like: a display font, a technical font, a body font. Pick colors you like. Still, it\u0026rsquo;s helpful to start from a sensible default, and the lab style is that default.\nFigures # Move quick, perfect later. A typical paper\u0026rsquo;s figures follow an arc:\nFigure 1 is your graphical abstract: tell the problem (and ideally the main takeaway) simply enough that an editor or an undergrad browsing papers can get it right away. Figures 2 to 3 are conceptual or data-centric: how you\u0026rsquo;re solving the problem and what the data looks like. Figures 4 to 8 are quantitative evidence. Sketch on a napkin or whiteboard first (see Writing \u0026amp; Publishing for how we mock up story beats and figure layouts early). There\u0026rsquo;s never enough time to make figures perfect from the start, so make a good-enough version with code (SVG or PNG), then fine-tune in Inkscape or GIMP.\nTitle plots around the purpose, not the tool (UMAP is a tool, so tell the reader why they\u0026rsquo;re looking at it). Label your axes, and add a small arrow showing which direction is better even when color already hints at it. Save heatmaps for when you genuinely have a lot of information (otherwise a bar plot or scatter plot is usually clearer), and don\u0026rsquo;t put a diverging colormap on a one-sided metric like AUROC.\nBe mindful of how color tells the story, and keep accessibility in mind. The Seaborn color palettes tutorial and ColorBrewer are good references.\nTypography # Role Font Weights Display and titles Spectral (serif) 400, 600 Body and UI Fira Sans (sans-serif) 300, 400, 500 Technical and data Fira Code (monospace) 400, 500 Picking a collection of colors # For categorical data, we use a custom 12-color set: a mix of ColorBrewer\u0026rsquo;s Dark2 (my favorite categorical palette) flavored with U of T\u0026rsquo;s accent colors, with gray anchored at the end for baselines. It comes in three flavors. Use Dark for lines, text, and markers on white, Normal for filled areas and bars, and Pastel for backgrounds, confidence bands, and highlights.\nU of T Institutional Colors U of T Primary (Navy) #002A5C U of T Secondary (Light Blue) #0072CE Color Dark Copy Python list Normal Copy Python list Pastel Copy Python list Teal #049D86 #60C2AC #9AD7C8 Pumpkin #D0670C #F79250 #FAB891 French Blue #5C78B6 #8CA0CB #B4C1DD Mexican Pink #E52A8F #EE88B3 #F5B2CC Avocado Green #75A30A #A6D854 #C3E792 Gold #D9B218 #FFD853 #FFE490 Caramel #A97420 #E7C395 #F0D8BA Sea Blue #0E7DA0 #78B4CD #A8CEDE Clay #B55344 #B75748 #D19287 Wisteria #7C3589 #B881C2 #D0ADD7 Duck Egg #0D8178 #B3EEE6 #CFF4EF Slate #666666 #B3B3B3 #CECECE For sequential scales, going from white to a single color works well. Pick one of the 12 colors (usually from Dark so the ramp has enough range) with a one-liner:\nsns.blend_palette([\u0026#34;#FFFFFF\u0026#34;, \u0026#34;#049D86\u0026#34;], as_cmap=True) Diverging palettes are great when there is a meaningful midpoint like zero separating negative and positive values, such as residuals or correlations. I\u0026rsquo;m partial to PiYG:\nsns.color_palette(\u0026#34;PiYG\u0026#34;, as_cmap=True) Templates # Slide and poster templates: [TBD]\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/aesthetics/","section":"How We Work 🚧","summary":"","title":"8. Aesthetics, Branding \u0026 Visuals","type":"culture"},{"content":"I think it\u0026rsquo;s good to be explicit about funding. Your position is an opportunity: you\u0026rsquo;re supported by many other people, by government and by taxes, to carry out a mission. That\u0026rsquo;s worth taking seriously and respecting.\nGraduate students # At U of T, a graduate student\u0026rsquo;s total funding package is roughly $31k to $40k per year, including tuition. It\u0026rsquo;s split between departmental base support (UTF), TAships, and the PI\u0026rsquo;s research grant (an RA stipend). Altogether, a graduate student costs the lab around $40k per year. Funding meets the ChemE departmental minimum.\nTAing for ChemE courses is part of the funding package (one term per year). Beyond that, you can TA extra if you want, just be careful with your time, especially during an MASc or PhD: it takes real time and energy, and that time and energy comes from somewhere else, though it\u0026rsquo;s your call.\nPostdocs # Postdocs are salaried employees with mandatory institutional benefits. Salaries start at $60k and can be topped up with grants.\nScholarships # I encourage applying for external scholarships. They look good on your CV and give the lab more flexibility with how we allocate funding. If you apply to the lab with your own funding, it makes it much easier for me to say yes, though it\u0026rsquo;s still not a guarantee.\nTravel # During a PhD, I aim for you to travel to at least two conferences (the same goes for postdocs). Traveling later in your career is usually better: you\u0026rsquo;ll have work to present, and you can make connections for your next step. Priority goes to presenting your own work, and it depends on the location.\nCosts and purchasing # The lab covers publication fees. Talk with me ahead of time when big expenses come up, such as flights or conference registration. Purchasing and reimbursement go through Concur. ","date":"2 November 2026","externalUrl":null,"permalink":"/culture/funding-and-operations/","section":"How We Work 🚧","summary":"","title":"9. Funding \u0026 Operations","type":"culture"},{"content":" Timelines and milestones # ChemE maintains official satisfactory-progress pages for both research degrees:\nMASc Satisfactory Progress (and general Master\u0026rsquo;s program info) PhD Satisfactory Progress, with dedicated timelines for PhD after an MASc and Direct-Entry PhD (plus PhD degree requirements) Along the way, the main formal checkpoints are your Reading Committee Meetings (RCM) and, for PhD students, the PhD Qualification Examination. Once you pass the qualifying exam, you really just have to meet with your RCM every 9 months (unless there is an issue) until you defend. We\u0026rsquo;ll also try to do an Individual Development Plan (IDP) to check in on your broader goals.\nMASc timeline (typically 2 years) # Stage ChemE checkpoints Lab focus Year 1 Safety (CHE2222H), Ethics (JDE1000H), Research Methods (CHE1102H), 1 to 2 courses, and your MASc Reading Committee Meeting (CHE3012Y) within 9 months Learn our stack, figure out how we work together, and scope a well-defined project you can succeed at Year 2 Any remaining courses, thesis, and MASc Departmental Oral Examination Finish and submit 1 led paper (a strong public artifact if you\u0026rsquo;re jumping somewhere else next) and wrap up your thesis PhD timeline (3 years post-MASc, or 4 to 5 years Direct-Entry) # Stage ChemE checkpoints Lab focus Year 1 Safety (CHE2222H), Ethics (JDE1000H), Research Methods (CHE1102H), 1 to 3 courses, first RCM (CHE3010Y) within 9 months, and PhD Qualification Exam (9 to 13 months post-MASc, or 13 to 16 months Direct-Entry) Figure out how we work together well, explore directions, take useful courses, and scope your first led project Year 2 Remaining courses, and RCM every 9 months after your qual (unless there\u0026rsquo;s an issue) Finish and submit your first led work, and start juggling a second project Year 3 RCM every 9 months, and achieve PhD candidacy (Year 3 post-MASc, or by Year 4 Direct-Entry) Lead your second, harder project (often connecting to experiments or collaborators) and present at conferences Years 4 to 5 RCM every 9 months, thesis, PhD Departmental Oral Exam, and SGS Final Oral Exam Work much more independently, lead your third, direction-setting work, explore internships or next steps, and staple your papers into a thesis Courses, community, and postdoc growth # A PhD is likely one of the last times in your life you\u0026rsquo;ll take courses. Use them to complement technical skills you want to build, and also think about other things you care about in life outside research. One of the best classes I\u0026rsquo;ve ever taken was an energy policy class at the Harvard Kennedy School: seeing how people outside technology and research thought about energy changed how I think about solving problems and how ideas actually get implemented in the world.\nFive years in a PhD (or a couple of years as a postdoc) is also one of the few times you\u0026rsquo;ll be surrounded by so many passionate people in one place. Lean on that. Meet people across groups, collaborate, or explore ideas you might want to embark on together, like starting a company.\nFor postdocs, a big part of the role is defining where you want to grow and making room for the experiences that help get you there, whether that\u0026rsquo;s lecturing experience, grant writing, or mentoring students.\nInternships # I think internships are great, especially closer to the end of your program when you have a strong foundation and are starting to think about your next step.\nOpportunities and sponsorships # There are many opportunities and sponsorships out there, and we track their deadlines in a shared spreadsheet [TBD]. They take time and energy, though they\u0026rsquo;re good for your career, and applying is a good way to put your work in perspective.\nLetters of recommendation # I tend to get two kinds of recommendation letter requests: from people I work with in research, and from students who took one of my classes.\nLab members and research collaborators. Please give me at least five business days\u0026rsquo; notice (one day only works if it\u0026rsquo;s urgent, and really rarely). Send me:\nYour CV When the deadline is What kind of narrative or points you think will be most useful for your application Don\u0026rsquo;t feel bad pinging me with a reminder, it really helps. Be selfish about these and make sure the letter gets in.\nStudents who took one of my classes. Please give me at least a month\u0026rsquo;s notice (I can sometimes be more flexible for people in ChemE or people I know well). I teach more than 150 students per year, so help me remember you clearly by sending:\nYour CV (having a photo on your CV or a LinkedIn link helps a lot, since I remember faces easily) Any useful links to remind me of your work Which class we took together, how much you attended (truthfully!), what kind of work you did, and what you liked about the class When the deadline is What kind of narrative you\u0026rsquo;re going for in your application Proposals # If you have a research proposal or idea, we want to iterate on it two or three times over about three weeks so you have time to think between drafts. The fastest route is booking a meeting [TBD] so we can go over it together.\nThesis # My personal opinion: theses are a relic of an older academic structure, and very few people read them. Don\u0026rsquo;t overspend time on yours. Staple your papers together and add an introduction and a conclusion.\nThe conclusion matters most. This is where you write to your past self: what did you want to hear when you started? It\u0026rsquo;s where you show you\u0026rsquo;ve grown into an expert on a topic and can explain clearly what you\u0026rsquo;ve learned. Be very careful with AI here especially, and spend time thinking about what you want to say in your own words.\nIf it\u0026rsquo;s done one week before the deadline, that\u0026rsquo;s fine, as long as the work is published. For AI use in your thesis, follow U of T\u0026rsquo;s AI guidelines.\nOutreach # I encourage outreach a lot. It\u0026rsquo;s a great way to remind yourself where our work sits within everything else. I\u0026rsquo;ve run STEM workshops for high school and undergraduate students, and I\u0026rsquo;d like to start Toronto science clubs in the future. Informal talks and partner programs are great too. If you have ideas, I\u0026rsquo;m really open. I\u0026rsquo;m always thinking about how we can do more.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/career-and-milestones/","section":"How We Work 🚧","summary":"","title":"10. Career \u0026 Milestones","type":"culture"},{"content":"Take vacation intentionally. That means don\u0026rsquo;t work, and don\u0026rsquo;t feel the need to answer messages.\nRole Vacation per year Notes Graduate students 15 days ChemE graduate students Employee postdocs (CUPE 3902 Unit 5) 3 weeks (15 days) Plus 5 sick days Trainee postdocs (external fellowships) 4 weeks (20 days) Personal leave covering vacation and sick days Everyone also gets statutory holidays and University closure days. Postdoc details are on the SGS postdoctoral office page.\nAnnounce your vacations ahead of time and check for deadline conflicts. That\u0026rsquo;s the biggest thing, especially if you\u0026rsquo;re working with a collaborator: make sure things are covered.\nBurnout # We\u0026rsquo;re in this for the long run. Like training for a marathon, it\u0026rsquo;s important to pace yourself. Small sprints are good, but you have to pay for them eventually.\nIf you\u0026rsquo;re feeling burnt out, I really recommend taking at least a day (but even more!). It can make a real difference. A day spent recovering is better than a day where you\u0026rsquo;re not working well.\nToronto has lots to offer for recharging. I personally love going to saunas (1, 2, 3), and the Central YMCA\u0026rsquo;s gym has one too. Exercise and meeting up with friends help too, and there are tons of running clubs (U of T Run Club, Running Rats). In general, just ask people what they do and you\u0026rsquo;ll get tons of answers and suggestions.\nIf you\u0026rsquo;re struggling, please reach out early. U of T\u0026rsquo;s Health \u0026amp; Wellness services are a good place to start. For a practical overview of how health coverage and medical care work here, see this guide to navigating healthcare at U of T written by Gary Tom.\nLeaves and accommodations # Life happens. Whether you need parental leave, medical leave, or accommodations, please communicate early so we can plan around it together. Keeping healthy comes first, and we\u0026rsquo;ll work with the policies in place so you can take care of yourself while still succeeding: see the SGS Leave of Absence \u0026amp; Parental Leave policy and U of T Accessibility Services.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/time-off-and-wellbeing/","section":"How We Work 🚧","summary":"","title":"11. Time Off \u0026 Wellbeing","type":"culture"},{"content":"A running grab bag of practical advice that doesn\u0026rsquo;t neatly fit anywhere else.\nPosters # A poster is a visual medium, so make the figures big and keep text short. Plan for 3 to 4 iterations with me before printing, and send drafts at least a few days ahead so there\u0026rsquo;s actually time to change things.\nA few things I almost always check before you print:\nPut your name near the top, and leave breathing room around the edges so the title isn\u0026rsquo;t hugging the border. Keep logos out of the tight corners and away from the title. Keep font sizes consistent (nothing random should be bigger than your title), and bold key statistics in larger type so people walking by can scan them. Group your visuals by idea (for example: inputs, labels, their interaction, and comparison to other datasets) instead of dropping in every plot you made. Match plot backgrounds to the section background, and check at full resolution that nothing is blurry. Standard sizes are usually 48x36 or 36x48 inches, though always check the conference board dimensions first. Resources and examples of aesthetically pleasing posters:\nPoster Gallery: a curated collection of strong academic posters (and for a recent favorite in the wild, see Nick Desnoyer\u0026rsquo;s OpenFlower poster). Fixing Academic Posters: The #BetterPoster Approach (Astrobites) and Zen Faulkes\u0026rsquo;s Better Posters blog. For printing around U of T, UTPoster on campus is the standard spot, and Staples is another option for same-day printing.\nTalks and slides # Putting real effort into a talk pays off more than people expect: a sharper talk leads to a better RCM presentation and a cleaner paper story.\nAim for at most ~15 slides, which keeps a steady pace of roughly one slide per minute. Write slide titles as full declarative sentences, and use the subtitle for the core takeaway (what someone should learn if they only caught that slide). Slides are a visual medium too, so lean on graphics. Corollary: if you put text on a slide, expect people to read it instead of listening to you. Stick to short phrases, single-line facts, or key-value pairs. Include compact citations right on the slide (Title, \u0026lsquo;YY, doi). Look at your audience. Part of a talk is performance, which means moving around a bit, varying the tone of your voice, and making it engaging. Two short reads I recommend:\nHow to Give an Academic Talk (Paul Edwards) The Message Box Workbook (COMPASS) Personal websites and headshots # Build a personal website early, even a very simple one. It gives you visibility when people look you up after a talk, poster, or paper. Free options include GitHub Pages (Lexi Jones\u0026rsquo;s walkthrough is an easy start), Google Sites, or Hugo with Blowfish (what our lab website uses).\nWhenever the department, an institute, or a conference has a headshot photographer, take the offer. One of my best photos came from accidentally coming into the office on a day they had a photographer, and I still use it all the time. Having a good photo also helps you visualize yourself in future roles.\nLocal conferences and volunteering # Submit talks and posters to local workshops and conferences (U of T events, Acceleration Consortium, Vector Institute, Canadian chemistry and engineering meetings). They\u0026rsquo;re low-bar to enter, great practice, and a good way to support the Canadian research ecosystem. If registration is tight or closed, check for student volunteer slots: volunteering usually gets you free registration and makes it easy to meet people.\nPractical ML references # Google Research\u0026rsquo;s Deep Learning Tuning Playbook: practical, opinionated defaults for choosing batch sizes, sweeping learning rates, and debugging training runs. ","date":"12 November 2026","externalUrl":null,"permalink":"/culture/miscellaneous-advice/","section":"How We Work 🚧","summary":"","title":"12. Miscellaneous Advice","type":"culture"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/tags/ai/","section":"Tags","summary":"","title":"Ai","type":"tags"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/alan-aspuru-guzik/","section":"Authors","summary":"","title":"Alan Aspuru-Guzik","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/anna-t.-thomas/","section":"Authors","summary":"","title":"Anna T. Thomas","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/","section":"Authors","summary":"","title":"Authors","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/benjamin-sanchez-lengeling/","section":"Authors","summary":"","title":"Benjamin Sanchez-Lengeling","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/caroline-cotto/","section":"Authors","summary":"","title":"Caroline Cotto","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/dominique-beaini/","section":"Authors","summary":"","title":"Dominique Beaini","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/ella-miray-rajaonson/","section":"Authors","summary":"","title":"Ella Miray Rajaonson","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/jungyoon-lee/","section":"Authors","summary":"","title":"Jungyoon Lee","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/kirill-neklyudov/","section":"Authors","summary":"","title":"Kirill Neklyudov","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/marta-skreta/","section":"Authors","summary":"","title":"Marta Skreta","type":"authors"},{"content":"Introduces SelectBench, a comprehensive benchmark assessing the selectivity of generative structure-based drug design (SBDD) models. Evaluates nine representative models across literature case studies, safety panels, and large-scale target pairs to quantify off-target discrimination.\n","date":"1 December 2026","externalUrl":"https://arxiv.org/abs/2610.24124","permalink":"/publications/rajaonson2026selectbench/","section":"Publications","summary":"","title":"On the Selectivity of Generative Models in Structure-Based Drug Design","type":"publications"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/sohum-patnaik/","section":"Authors","summary":"","title":"Sohum Patnaik","type":"authors"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/tags/","section":"Tags","summary":"","title":"Tags","type":"tags"},{"content":"Introduces TasteBench, a multimodal benchmark and competition for sensory prediction across 21K+ human evaluations on 215 plant-based foods and 15K flavor molecules, providing computational proxies to accelerate sustainable food discovery.\n","date":"1 December 2026","externalUrl":"https://openreview.net/forum?id=k3djmwrAXd","permalink":"/publications/thomas2026tastebench/","section":"Publications","summary":"","title":"TasteBench: Multimodal Benchmark for Sensory Prediction, from Molecules to Sustainable Foods","type":"publications"},{"content":"","date":"1 December 2026","externalUrl":null,"permalink":"/authors/william-chau/","section":"Authors","summary":"","title":"William Chau","type":"authors"},{"content":"This is the living handbook for the Chemical Cognition Lab: how we work, what we expect from each other, and why. It\u0026rsquo;s public on purpose so prospective members can see what the lab is like, current members have a shared reference, and we stay accountable to what we write down.\nIt will keep changing. If something here is unclear, outdated, or just wrong, post in #General on Zulip (or open a pull request in the website repo [TBD] if you\u0026rsquo;re in the lab).\nThinking of joining? Start with Values \u0026amp; Culture, Roles \u0026amp; Expectations, and Funding \u0026amp; Operations, then see Join the Lab for how to reach out.\nNew to the lab? Start with Getting Started.\n","date":"2 November 2026","externalUrl":null,"permalink":"/culture/","section":"How We Work 🚧","summary":"","title":"How We Work 🚧","type":"culture"},{"content":"Introduces IGLOO, a diagnostic framework that evaluates whether residue-level embedding gradients from protein language models faithfully capture functional importance against synthetic and experimental fitness landscapes.\n","date":"21 October 2026","externalUrl":"https://www.biorxiv.org/content/10.64898/2026.10.21.760763v1","permalink":"/publications/barghout2026igloo/","section":"Publications","summary":"","title":"Do protein representations encode variant function? Diagnosing residue-level protein attributions via embedding-space gradients","type":"publications"},{"content":"","date":"21 October 2026","externalUrl":null,"permalink":"/authors/jos%C3%A9-manuel-barraza-chavez/","section":"Authors","summary":"","title":"José Manuel Barraza-Chavez","type":"authors"},{"content":"","date":"21 October 2026","externalUrl":null,"permalink":"/authors/paola-driza/","section":"Authors","summary":"","title":"Paola Driza","type":"authors"},{"content":"","date":"21 October 2026","externalUrl":null,"permalink":"/authors/radhakrishnan-mahadevan/","section":"Authors","summary":"","title":"Radhakrishnan Mahadevan","type":"authors"},{"content":"","date":"21 October 2026","externalUrl":null,"permalink":"/authors/rana-a.-barghout/","section":"Authors","summary":"","title":"Rana A. Barghout","type":"authors"},{"content":"","date":"21 October 2026","externalUrl":null,"permalink":"/authors/shengqing-wu/","section":"Authors","summary":"","title":"Shengqing Wu","type":"authors"},{"content":"Couples deep learning turnover number predictions with stochastic simulated annealing to parameterize enzyme-constrained genome-scale models across 93 organisms, resolving long-standing biocatalytic data sparsity.\n","date":"6 October 2026","externalUrl":"https://doi.org/10.64898/2026.03.14.711833","permalink":"/publications/barghout2026kingems/","section":"Publications","summary":"","title":"kinGEMs: A scalable framework for resource-constrained models through stochastic tuning of deep learning-predicted kinetic parameters","type":"publications"},{"content":"","date":"6 October 2026","externalUrl":null,"permalink":"/authors/lya-chinas-serrano/","section":"Authors","summary":"","title":"Lya Chinas Serrano","type":"authors"},{"content":"Crowdsources predictive models across 26 international teams in the DREAM Olfactory Challenge, demonstrating that compact semantic representations of single odorants generalize to complex mixture perceptual spaces.\n","date":"1 August 2026","externalUrl":"https://doi.org/10.1073/pnas.2611057123","permalink":"/publications/satarifard2026semantic/","section":"Publications","summary":"","title":"A semantic-based community model for high-fidelity tuning of olfactory mixture distances","type":"publications"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/achilleas-ghinis/","section":"Authors","summary":"","title":"Achilleas Ghinis","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/aharon-ravia/","section":"Authors","summary":"","title":"Aharon Ravia","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/cici-xingyu-zheng/","section":"Authors","summary":"","title":"CiCi Xingyu Zheng","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/gaia-andreoletti/","section":"Authors","summary":"","title":"Gaia Andreoletti","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/jake-albrecht/","section":"Authors","summary":"","title":"Jake Albrecht","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/jiacheng-chen/","section":"Authors","summary":"","title":"Jiacheng Chen","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/jianing-tang/","section":"Authors","summary":"","title":"Jianing Tang","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/jianming-tang/","section":"Authors","summary":"","title":"Jianming Tang","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/jiazhen-he/","section":"Authors","summary":"","title":"Jiazhen He","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/joel-d.-mainland/","section":"Authors","summary":"","title":"Joel D. Mainland","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/laura-sisson/","section":"Authors","summary":"","title":"Laura Sisson","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/liangzhen-pan/","section":"Authors","summary":"","title":"Liangzhen Pan","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/matej-hladi%C5%A1/","section":"Authors","summary":"","title":"Matej Hladiš","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/maxence-lalis/","section":"Authors","summary":"","title":"Maxence Lalis","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/menglong-dong/","section":"Authors","summary":"","title":"Menglong Dong","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/min-zhang/","section":"Authors","summary":"","title":"Min Zhang","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/pablo-meyer/","section":"Authors","summary":"","title":"Pablo Meyer","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/pedro-il%C3%ADdio/","section":"Authors","summary":"","title":"Pedro Ilídio","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/ping-li/","section":"Authors","summary":"","title":"Ping Li","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/qian-shen/","section":"Authors","summary":"","title":"Qian Shen","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/richard-c.-gerkin/","section":"Authors","summary":"","title":"Richard C. Gerkin","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/robbe-dh%CC%81ondt/","section":"Authors","summary":"","title":"Robbe Dh́ondt","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/robert-pellegrino/","section":"Authors","summary":"","title":"Robert Pellegrino","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/siyuan-chen/","section":"Authors","summary":"","title":"Siyuan Chen","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/stephen-yang/","section":"Authors","summary":"","title":"Stephen Yang","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/stijn-vranckx/","section":"Authors","summary":"","title":"Stijn Vranckx","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/tiffany-yang/","section":"Authors","summary":"","title":"Tiffany Yang","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/vahid-satarifard/","section":"Authors","summary":"","title":"Vahid Satarifard","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/wenjie-yin/","section":"Authors","summary":"","title":"Wenjie Yin","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/xuebo-song/","section":"Authors","summary":"","title":"Xuebo Song","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/yang-liu/","section":"Authors","summary":"","title":"Yang Liu","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/yikun-han/","section":"Authors","summary":"","title":"Yikun Han","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/yuan-yuan/","section":"Authors","summary":"","title":"Yuan Yuan","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/yue-hu/","section":"Authors","summary":"","title":"Yue Hu","type":"authors"},{"content":"","date":"1 August 2026","externalUrl":null,"permalink":"/authors/zehua-wang/","section":"Authors","summary":"","title":"Zehua Wang","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/anna-thomas/","section":"Authors","summary":"","title":"Anna Thomas","type":"authors"},{"content":"Comprehensive review articulating an AI-driven framework for sustainable food formulation, spanning perception modeling, multi-omics fermentation engineering, and consumer acceptance.\n","date":"1 July 2026","externalUrl":"https://doi.org/10.1038/s43016-026-01380-7","permalink":"/publications/datta2026artificial/","section":"Publications","summary":"","title":"Artificial intelligence for food innovation","type":"publications"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/bianca-datta/","section":"Authors","summary":"","title":"Bianca Datta","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/dan-jurafsky/","section":"Authors","summary":"","title":"Dan Jurafsky","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/david-l.-kaplan/","section":"Authors","summary":"","title":"David L. Kaplan","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/ellen-kuhl/","section":"Authors","summary":"","title":"Ellen Kuhl","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/giorgia-del-missier/","section":"Authors","summary":"","title":"Giorgia Del Missier","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/ilias-tagkopoulos/","section":"Authors","summary":"","title":"Ilias Tagkopoulos","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/karim-pichara/","section":"Authors","summary":"","title":"Karim Pichara","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/kristina-gligori%C4%87/","section":"Authors","summary":"","title":"Kristina Gligorić","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/lisa-neidhardt/","section":"Authors","summary":"","title":"Lisa Neidhardt","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/markus-j.-buehler/","section":"Authors","summary":"","title":"Markus J. Buehler","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/matthew-watson/","section":"Authors","summary":"","title":"Matthew Watson","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/miek-schlangen/","section":"Authors","summary":"","title":"Miek Schlangen","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/tags/reviews/","section":"Tags","summary":"","title":"Reviews","type":"tags"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/rodrigo-ledesma-amaro/","section":"Authors","summary":"","title":"Rodrigo Ledesma-Amaro","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/skyler-r.-st.-pierre/","section":"Authors","summary":"","title":"Skyler R. St. Pierre","type":"authors"},{"content":"","date":"1 July 2026","externalUrl":null,"permalink":"/authors/yvonne-chow/","section":"Authors","summary":"","title":"Yvonne Chow","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/abhiram-chalamalasetty/","section":"Authors","summary":"","title":"Abhiram Chalamalasetty","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/adesh-rohan-mishra/","section":"Authors","summary":"","title":"Adesh Rohan Mishra","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/adrian-jinich/","section":"Authors","summary":"","title":"Adrian Jinich","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/eric-sivonxay/","section":"Authors","summary":"","title":"Eric Sivonxay","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/evan-walter-clark-spotte-smith/","section":"Authors","summary":"","title":"Evan Walter Clark Spotte-Smith","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/fleur-m.-ferguson/","section":"Authors","summary":"","title":"Fleur M. Ferguson","type":"authors"},{"content":"Develops differentiable heterogeneous graph neural network surrogates for kinetic Monte Carlo simulations of core-shell upconverting nanoparticles, achieving 6.5-fold experimental emission enhancement via gradient ascent.\n","date":"1 January 2026","externalUrl":"https://doi.org/10.1038/s43588-025-00917-3","permalink":"/publications/sivonxay2026gradient/","section":"Publications","summary":"","title":"Gradient-based optimization of complex nanoparticle heterostructures enabled by deep learning on heterogeneous graphs","type":"publications"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/lucas-attia/","section":"Authors","summary":"","title":"Lucas Attia","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/robert-bardhan/","section":"Authors","summary":"","title":"Robert Bardhan","type":"authors"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/samuel-m.-blau/","section":"Authors","summary":"","title":"Samuel M. Blau","type":"authors"},{"content":"Pairs a GINE molecular graph encoder with frozen ESM-C protein language model embeddings via cross-attention to predict covalent target engagement across pathogen proteomes with few-shot transfer.\n","date":"1 January 2026","externalUrl":"https://www.biorxiv.org/content/10.64898/2026.01.04.697588v1","permalink":"/publications/chalamalasetty2026tubercuprobe/","section":"Publications","summary":"","title":"TubercuProbe: A cross-attention graph-sequence model for cross-species chemoproteomic discovery in Mycobacterium tuberculosis","type":"publications"},{"content":"","date":"1 January 2026","externalUrl":null,"permalink":"/authors/xiaojing-xia/","section":"Authors","summary":"","title":"Xiaojing Xia","type":"authors"},{"content":"Establishes the first unified open-source benchmark suite spanning 11 diverse chemical mixture domains with standardized splitting protocols and deep learning baseline models.\n","date":"1 December 2025","externalUrl":"https://arxiv.org/abs/2506.12231","permalink":"/publications/rajaonson2025chemixhub/","section":"Publications","summary":"","title":"CheMixHub: Datasets and benchmarks for chemical mixture property prediction","type":"publications"},{"content":"","date":"1 December 2025","externalUrl":null,"permalink":"/authors/luis-martin-mej%C3%ADa-mendoza/","section":"Authors","summary":"","title":"Luis Martin Mejía Mendoza","type":"authors"},{"content":"","date":"1 December 2025","externalUrl":null,"permalink":"/authors/mahyar-rajabi-kochi/","section":"Authors","summary":"","title":"Mahyar Rajabi Kochi","type":"authors"},{"content":"","date":"1 December 2025","externalUrl":null,"permalink":"/authors/seyed-mohamad-moosavi/","section":"Authors","summary":"","title":"Seyed Mohamad Moosavi","type":"authors"},{"content":"A foundational pedagogical text and monograph introducing structural, geometric, and kinetic graph neural networks for molecules, macromolecular proteins, and complex biological pathways.\n","date":"1 September 2025","externalUrl":"https://doi.org/10.1021/acsinfocus.7e9017","permalink":"/publications/barrazachavez2025graph/","section":"Publications","summary":"","title":"Graph Data Modeling: Molecules, Proteins, \u0026 Chemical Processes","type":"publications"},{"content":"","date":"1 September 2025","externalUrl":null,"permalink":"/authors/ricardo-almada-monter/","section":"Authors","summary":"","title":"Ricardo Almada-Monter","type":"authors"},{"content":"","date":"1 August 2025","externalUrl":null,"permalink":"/authors/al%C3%A1n-aspuru-guzik/","section":"Authors","summary":"","title":"Alán Aspuru-Guzik","type":"authors"},{"content":"","date":"1 August 2025","externalUrl":null,"permalink":"/authors/brian-k.-lee/","section":"Authors","summary":"","title":"Brian K. Lee","type":"authors"},{"content":"","date":"1 August 2025","externalUrl":null,"permalink":"/authors/cher-tian-ser/","section":"Authors","summary":"","title":"Cher-Tian Ser","type":"authors"},{"content":"Introduces POMMix, incorporating inductive perceptual pooling operations on molecular graphs to predict the non-linear emergent sensory properties of multi-component odorant mixtures.\n","date":"1 August 2025","externalUrl":"https://doi.org/10.1088/2632-2153/adfffc","permalink":"/publications/tom2025olfactory/","section":"Publications","summary":"","title":"Does this smell the same? Learning representations of olfactory mixtures using inductive biases","type":"publications"},{"content":"","date":"1 August 2025","externalUrl":null,"permalink":"/authors/gary-tom/","section":"Authors","summary":"","title":"Gary Tom","type":"authors"},{"content":"","date":"1 August 2025","externalUrl":null,"permalink":"/authors/hyun-suk-park/","section":"Authors","summary":"","title":"Hyun Suk Park","type":"authors"},{"content":"Demonstrates that pairwise ranking surrogates systematically outperform point regression models in Bayesian optimization for molecular property design, mitigating miscalibration and heteroskedastic noise.\n","date":"1 August 2025","externalUrl":"https://doi.org/10.1063/5.0272663","permalink":"/publications/tom2025ranking/","section":"Publications","summary":"","title":"Ranking over regression for Bayesian optimization and molecule selection","type":"publications"},{"content":"","date":"1 August 2025","externalUrl":null,"permalink":"/authors/samantha-corapi/","section":"Authors","summary":"","title":"Samantha Corapi","type":"authors"},{"content":"","date":"1 August 2025","externalUrl":null,"permalink":"/authors/stanley-lo/","section":"Authors","summary":"","title":"Stanley Lo","type":"authors"},{"content":"Leverages graph neural representations from the Principal Odor Map (POM) to discover potent, novel repellent chemotypes validated experimentally against multiple mosquito disease vectors.\n","date":"1 July 2025","externalUrl":"https://doi.org/10.1093/chemse/bjaf021","permalink":"/publications/chemsenses2025repellents/","section":"Publications","summary":"","title":"A deep learning and digital archaeology approach for mosquito repellent discovery","type":"publications"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/alexander-b.-wiltschko/","section":"Authors","summary":"","title":"Alexander B. Wiltschko","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/carlos-ruiz/","section":"Authors","summary":"","title":"Carlos Ruiz","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/d.-michael-ando/","section":"Authors","summary":"","title":"D. Michael Ando","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/jacob-n.-sanders/","section":"Authors","summary":"","title":"Jacob N. Sanders","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/jeffrey-a.-riffell/","section":"Authors","summary":"","title":"Jeffrey A. Riffell","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/jennifer-n.-wei/","section":"Authors","summary":"","title":"Jennifer N. Wei","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/koen-j.-dechering/","section":"Authors","summary":"","title":"Koen J. Dechering","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/kurt-m.-groetsch/","section":"Authors","summary":"","title":"Kurt M. Groetsch","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/luuk-berning/","section":"Authors","summary":"","title":"Luuk Berning","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/marnix-vlot/","section":"Authors","summary":"","title":"Marnix Vlot","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/martijn-w.-vos/","section":"Authors","summary":"","title":"Martijn W. Vos","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/rob-w.-m.-henderson/","section":"Authors","summary":"","title":"Rob W. M. Henderson","type":"authors"},{"content":"","date":"1 July 2025","externalUrl":null,"permalink":"/authors/wesley-w.-qian/","section":"Authors","summary":"","title":"Wesley W. Qian","type":"authors"},{"content":"Introduces Stereoelectronics-Infused Molecular Graphs (SIMGs) and neural surrogate estimators that encode Natural Bond Orbital interactions into molecular graphs in seconds, boosting property prediction accuracy.\n","date":"1 January 2025","externalUrl":"https://doi.org/10.1038/s42256-025-01031-9","permalink":"/publications/reschuetzegger2025stereoelectronics/","section":"Publications","summary":"","title":"Advancing molecular machine learning representations with stereoelectronics-infused molecular graphs","type":"publications"},{"content":"","date":"1 January 2025","externalUrl":null,"permalink":"/authors/daniil-a.-boiko/","section":"Authors","summary":"","title":"Daniil A. Boiko","type":"authors"},{"content":"","date":"1 January 2025","externalUrl":null,"permalink":"/authors/gabe-gomes/","section":"Authors","summary":"","title":"Gabe Gomes","type":"authors"},{"content":"","date":"1 January 2025","externalUrl":null,"permalink":"/authors/thiago-resch%C3%BCtzegger/","section":"Authors","summary":"","title":"Thiago Reschützegger","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/agnieszka-wolos/","section":"Authors","summary":"","title":"Agnieszka Wolos","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/al%C3%A1n-vivanco/","section":"Authors","summary":"","title":"Alán Vivanco","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/bartosz-a.-grzybowski/","section":"Authors","summary":"","title":"Bartosz A. Grzybowski","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/benjam%C3%ADn-s%C3%A1nchez-lengeling/","section":"Authors","summary":"","title":"Benjamín Sánchez-Lengeling","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/carlota-bozal-ginesta/","section":"Authors","summary":"","title":"Carlota Bozal-Ginesta","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/chihaya-adachi/","section":"Authors","summary":"","title":"Chihaya Adachi","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/cyrille-lavigne/","section":"Authors","summary":"","title":"Cyrille Lavigne","type":"authors"},{"content":"A globally distributed, asynchronous self-driving laboratory discovers 21 state-of-the-art organic solid-state laser gain materials across multiple sites and automated robotics testbeds.\n","date":"1 May 2024","externalUrl":"https://doi.org/10.1126/science.adk9227","permalink":"/publications/striethkalthoff2024delocalized/","section":"Publications","summary":"","title":"Delocalized, asynchronous, closed-loop discovery of organic laser emitters","type":"publications"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/ekaterina-trushina/","section":"Authors","summary":"","title":"Ekaterina Trushina","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/felix-strieth-kalthoff/","section":"Authors","summary":"","title":"Felix Strieth-Kalthoff","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/han-hao/","section":"Authors","summary":"","title":"Han Hao","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/jenya-vestfrid/","section":"Authors","summary":"","title":"Jenya Vestfrid","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/joshua-derasp/","section":"Authors","summary":"","title":"Joshua Derasp","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/junliang-liu/","section":"Authors","summary":"","title":"Junliang Liu","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/kazuhiro-hotta/","section":"Authors","summary":"","title":"Kazuhiro Hotta","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/leticia-bodo/","section":"Authors","summary":"","title":"Leticia Bodo","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/mariana-medina-s%C3%A1nchez/","section":"Authors","summary":"","title":"Mariana Medina-Sánchez","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/martin-d.-burke/","section":"Authors","summary":"","title":"Martin D. Burke","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/martin-seifrid/","section":"Authors","summary":"","title":"Martin Seifrid","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/masashi-mamada/","section":"Authors","summary":"","title":"Masashi Mamada","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/mason-guy/","section":"Authors","summary":"","title":"Mason Guy","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/mohammad-haddadnia/","section":"Authors","summary":"","title":"Mohammad Haddadnia","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/n.-m.-anoop-krishnan/","section":"Authors","summary":"","title":"N. M. Anoop Krishnan","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/nicholas-h.-angello/","section":"Authors","summary":"","title":"Nicholas H. Angello","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/oscar-guti%C3%A9rrez-espinoza/","section":"Authors","summary":"","title":"Oscar Gutiérrez-Espinoza","type":"authors"},{"content":"Synthesizes key outcomes from the AI4Mat workshop at NeurIPS 2023, charting trajectories for self-driving laboratories, simulation-to-materials workflows, and language models for physical science.\n","date":"1 May 2024","externalUrl":"https://doi.org/10.1039/d4dd90010c","permalink":"/publications/miret2024perspective/","section":"Publications","summary":"","title":"Perspective on AI for accelerated materials design at the AI4Mat-2023 workshop at NeurIPS 2023","type":"publications"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/rafal-roszak/","section":"Authors","summary":"","title":"Rafal Roszak","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/riley-j.-hickman/","section":"Authors","summary":"","title":"Riley J. Hickman","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/robert-pollice/","section":"Authors","summary":"","title":"Robert Pollice","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/santiago-miret/","section":"Authors","summary":"","title":"Santiago Miret","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/shangyu-li/","section":"Authors","summary":"","title":"Shangyu Li","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/th%C3%A9ophile-gaudin/","section":"Authors","summary":"","title":"Théophile Gaudin","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/tony-c.-wu/","section":"Authors","summary":"","title":"Tony C. Wu","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/tuul-tsagaantsooj/","section":"Authors","summary":"","title":"Tuul Tsagaantsooj","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/vandana-rathore/","section":"Authors","summary":"","title":"Vandana Rathore","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/vandana-srikumar/","section":"Authors","summary":"","title":"Vandana Srikumar","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/vineeth-venugopal/","section":"Authors","summary":"","title":"Vineeth Venugopal","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/wesley-wang/","section":"Authors","summary":"","title":"Wesley Wang","type":"authors"},{"content":"","date":"1 May 2024","externalUrl":null,"permalink":"/authors/xun-tang/","section":"Authors","summary":"","title":"Xun Tang","type":"authors"},{"content":"Abigail Djossou is a postdoctoral researcher in the Chemical Cognition Lab. She holds a PhD in computer science from Université de Sherbrooke, where she worked on eukaryotic gene prediction with hidden semi-Markov models. She is working on MecCog, a community challenge assessing how well agentic AI recovers APOE4-related mechanisms of Alzheimer\u0026rsquo;s disease. Her interests include benchmarking, consensus metrics, and machine learning for biological sequences.\n","externalUrl":null,"permalink":"/people/abigail-djossou/","section":"People","summary":"","title":"Abigail Djossou","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/aziz-abdur-rakib/","section":"People","summary":"","title":"Aziz Abdur Rakib","type":"people"},{"content":"Visit my website: https://www.benmsanchez.com/\n","externalUrl":null,"permalink":"/people/benjamin-sanchez/","section":"People","summary":"","title":"Benjamin Sánchez-Lengeling","type":"people"},{"content":"","externalUrl":null,"permalink":"/categories/","section":"Categories","summary":"","title":"Categories","type":"categories"},{"content":"","externalUrl":null,"permalink":"/","section":"Chemical Cognition Lab","summary":"","title":"Chemical Cognition Lab","type":"page"},{"content":"","externalUrl":null,"permalink":"/people/dumebi-nasa-okolie/","section":"People","summary":"","title":"Dumebi Nasa-Okolie","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/edward-lombo/","section":"People","summary":"","title":"Edward Lombo","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/elizabeth-reynolds/","section":"People","summary":"","title":"Elizabeth Reynolds","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/ella-rajaonson/","section":"People","summary":"","title":"Ella M. Rajaonson","type":"people"},{"content":"Jenn Kolwich is a chemist and microbiologist by training. They completed their PhD at Queen\u0026rsquo;s University, where they developed direct mass spectrometry sampling workflows to study the chemical ecology of living microbes in situ. Now, they draw on this dual expertise to identify data needs at the interface of microbial chemical ecology and AI, developing benchmark datasets and standards for predictive modelling of marine microbes under climate stress. Their research interests include metabolomics, data-centric AI, and microbial ecology.\n","externalUrl":null,"permalink":"/people/jenn-kolwich/","section":"People","summary":"","title":"Jenn Kolwich","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/jessica-anirisaihan/","section":"People","summary":"","title":"Jessica Anirisaihan","type":"people"},{"content":" We are actively recruiting grad students (MASc and PhD) and postdocs, as well as undergrad and master's students for research projects. Before reaching out, please read our How We Work lab handbook to get a sense of our culture, expectations, and how we operate. Who we\u0026rsquo;re looking for # We want people who care about what we\u0026rsquo;re trying to do here. In problem spaces and skills, our work mixes:\nScience: Chemistry, biochemistry, or problems that touch people\u0026rsquo;s lives. Small molecules, proteins, DNA, and sometimes no molecules at all. AI: Thinking about data in a mathy/statistical way. Engineering mindset: Solving problems and writing good, useful code. Be good at one or two and want to get better at the rest. Just as much, we look for people who share our values: high-quality work, balance, and generous collaboration.\nNon-traditional paths are welcome # My path wasn\u0026rsquo;t a straight line. I grew up in Guanajuato, Mexico, went to public school, and started university studying math because I wanted to make video games. I played in two electroacoustic bands, worked as a barista for a year, and was the oldest PhD student in my cohort. Along the way I modeled placentas, did quantum chemistry, spent five years at Google teaching computers to predict smell, and then started this lab.\nDon\u0026rsquo;t hide the things that got you here: a different field, time in industry, a career change, a gap, or a school nobody has heard of. It\u0026rsquo;s all part of your story, and I\u0026rsquo;d like to hear it.\nAdmissions and funding # Grad student applicants (MASc and PhD) have to apply formally through ChemE admissions. Writing to me helps: if we connect and I think you\u0026rsquo;re a good candidate, you\u0026rsquo;re more likely to be selected, though I\u0026rsquo;ve also selected students from the pool who didn\u0026rsquo;t write to me first.\nFor undergrads, postdocs, and other paths, every situation is different, so look into the options and write to me. Having your own funding helps, though what I really care about is that you\u0026rsquo;ll be a good fit for the group and in an environment that lets you grow.\nHow to reach out # Email me at ben.sanchez@utoronto.ca with [Prospective] in the subject line, including:\nWhich paper, project, or idea caught your eye, and why. Which part of the lab handbook you find interesting. Why you want to work on these problems. Any external funding (scholarships or fellowships). Your CV, plus links to code or projects you\u0026rsquo;re proud of. Write it yourself # I want to hear how you think. No AI-written emails.\nWhat happens next # If your email catches my eye, I\u0026rsquo;ll follow up by email or set up a small chat. If you\u0026rsquo;re in the Toronto area, you\u0026rsquo;re also always welcome to talk to students in the lab if they have time.\nSadly, email sometimes gets overwhelming, so it is hard to respond always. If you haven\u0026rsquo;t heard back in two weeks, please send a nudge.\n","externalUrl":null,"permalink":"/contact/","section":"Chemical Cognition Lab","summary":"","title":"Join the Chemical Cognition Lab","type":"page"},{"content":"","externalUrl":null,"permalink":"/people/jose-manuel-barraza/","section":"People","summary":"","title":"Jose Manuel Barraza","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/kevin-zhu/","section":"People","summary":"","title":"Kevin Zhu","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/paola-driza/","section":"People","summary":"","title":"Paola Driza","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/","section":"People","summary":"","title":"People","type":"people"},{"content":"","externalUrl":null,"permalink":"/publications/","section":"Publications","summary":"","title":"Publications","type":"publications"},{"content":"","externalUrl":null,"permalink":"/people/rana-barghout/","section":"People","summary":"","title":"Rana A. Barghout","type":"people"},{"content":"","externalUrl":null,"permalink":"/research/","section":"Research","summary":"","title":"Research","type":"research"},{"content":"","externalUrl":null,"permalink":"/people/rosalie-wang/","section":"People","summary":"","title":"Rosalie Wang","type":"people"},{"content":"","externalUrl":null,"permalink":"/series/","section":"Series","summary":"","title":"Series","type":"series"},{"content":"","externalUrl":null,"permalink":"/people/shengqing-wu/","section":"People","summary":"","title":"Shengqing Wu","type":"people"},{"content":"","externalUrl":null,"permalink":"/research/topics/","section":"Research","summary":"","title":"Topics","type":"research"},{"content":"","externalUrl":null,"permalink":"/people/uyen-hua/","section":"People","summary":"","title":"Uyen Hua","type":"people"},{"content":"","externalUrl":null,"permalink":"/people/william-chau/","section":"People","summary":"","title":"William Chau","type":"people"}]