Tommi Jaakkola
Thomas M. Siebel Distinguished Professor
Develops machine learning methods and theory for generating semi-structured data such as molecules and proteins, including the BoltzGen protein design model.
Tommi Jaakkola's lab at MIT works on learning theory & optimization. 65 papers since 2023, 55 at top venues, mostly ICLR and ICML. 51 alumni. Applications close .
At a glance
- papers, 2023–2026
- 17/ year
- at T1 venues
- 85%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 51
Papers by yearfrom the lab homepage, 2023–2026
- 9 papers2026 · T1 6 · collecting
- 17 papers2025 · T1 16
- 21 papers2024 · T1 18
- 18 papers2023 · T1 15
T1 venuesOtherVenues: ICLR, ICML, NeurIPS, CVPR, TMLR
Research focuspaper titles that use each topic's words
- Learning theory & optimization10 papers
- protein design2 papers
- generating1 paper
- molecules0 papers
- proteins0 papers
Paper list
From the lab homepage.
20269 papers
- Current Opinion in Structural BiologyAI-based methods for simulating, sampling, and predicting protein ensembles (opens in a new tab)
- ICML (T1)Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasoner (opens in a new tab)
- ICML (T1)Diamond Maps: Efficient Reward Alignment via Stochastic Flow Maps (opens in a new tab)
- ICML (T1)Divide-and-Denoise: A Game-Theoretic Method for Fairly Composing Diffusion Models (opens in a new tab)
- CVPR (T1)Flow Map Distillation Without Data (opens in a new tab)
- ICLR (T1)GLASS Flows: Efficient Inference for Reward Alignment of Flow and Diffusion Models (opens in a new tab)
- COLMHow Well Do Agentic Skills Work in the Wild: Benchmarking LLM Skill Usage in Realistic Settings (opens in a new tab)
- COLMOnline Reasoning Calibration: Test-Time Training Enables Generalizable Conformal LLM Reasoning (opens in a new tab)
- ICLR (T1)SPG: Sandwiched Policy Gradient for Masked Diffusion Language Models (opens in a new tab)
202517 papers
- ICLR (T1)An Information Criterion for Controlled Disentanglement of Multimodal Data (opens in a new tab)
- Nature Methods (T1)Atom-level enzyme active site scaffolding using RFdiffusion2 (opens in a new tab)
- ICLR (T1)Composing Unbalanced Flows for Flexible Docking and Relaxation (opens in a new tab)
- ICLR (T1)Data Distillation for extrapolative protein design through exact preference optimization (opens in a new tab)
- ICLR (T1)Fictitious Synthetic Data Can Improve LLM Factuality via Prerequisite Learning (opens in a new tab)
- ICLR (T1)Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein Design (opens in a new tab)
- ICLR (T1)Generator Matching: Generative modeling with arbitrary Markov processes (opens in a new tab)
- ICML (T1)Identifying biological perturbation targets through causal differential networks (opens in a new tab)
- CVPR (T1)Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps (opens in a new tab)
- ICML (T1)LEAPS: A discrete neural sampler via locally equivariant networks (opens in a new tab)
- NeurIPS (T1)Learning Diffusion Models with Flexible Representation Guidance (opens in a new tab)
- NeurIPS (T1)Next Semantic Scale Prediction via Hierarchical Diffusion Language Models (opens in a new tab)
202421 papers
- NeurIPS (T1)A Recipe for Charge Density Prediction (opens in a new tab)
- ICML (T1)AlphaFold Meets Flow Matching for Generating Protein Ensembles (opens in a new tab)
- ICLR (T1)Conformal Language Modeling (opens in a new tab)
- CVPR (T1)Correcting Diffusion Generation through Resampling (opens in a new tab)
- ICLR (T1)Deep Confident Steps to New Pockets: Strategies for Docking Generalization (opens in a new tab)
- WIREs Computational Molecular ScienceDiffusion models in protein structure and docking (opens in a new tab)
- ICML (T1)Dirichlet Flow Matching with Applications to DNA Sequence Design (opens in a new tab)
- ICML (T1)DisCo-Diff: Enhancing Continuous Diffusion Models with Discrete Latents (opens in a new tab)
- ICML (T1)Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design (opens in a new tab)
- NeurIPS (T1)Generative Modeling of Molecular Dynamics Trajectories (opens in a new tab)
- Nature Reviews Methods PrimersGraph Neural Networks (opens in a new tab)
- NeurIPS (T1)Hamiltonian Score Matching and Generative Flows (opens in a new tab)
202318 papers
- Science (T1)Autonomous, multiproperty-driven molecular discovery: From predictions to measurements and back (opens in a new tab)
- NeurIPS (T1)Compositional Sculpting of Iterative Generative Processes (opens in a new tab)
- Nature (T1)De novo design of protein structure and function with RFdiffusion (opens in a new tab)
- Nature Chemical BiologyDeep learning-guided discovery of an antibiotic targeting Acinetobacter baumannii (opens in a new tab)
- Machine Learning for Drug DiscoveryDiffDock-PP: Rigid Protein-Protein Docking with Diffusion Models (opens in a new tab)
- ICLR (T1)DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking (opens in a new tab)
- ICLR (T1)Diffusion Probabilistic Modeling of Protein Backbones in 3D for the motif-scaffolding problem (opens in a new tab)
- ICLR (T1)Efficiently Controlling Multiple Risks with Pareto Testing (opens in a new tab)
- Machine Learning for Drug Discovery WorkshopEigenFold: Generative Protein Structure Prediction with Diffusion Models (opens in a new tab)
- TMLR (T1)Forces are not Enough: Benchmark and Critical Evaluation for Machine Learning Force Fields with Molecular Simulations (opens in a new tab)
- NeurIPS (T1)Hierarchical Planning with Foundation Models (opens in a new tab)
- ICLR (T1)Is Conditional Generative Modeling all you need for Decision Making? (opens in a new tab)
Where alumni went51 listed
Apply to MITFall 2027
MIT EECS Computer Science PhD
- Round
- Single round
- Opens
- —
- GRE
- See program page
- English
- TOEFL/IELTS
- Fee
- USD 90
Confirmed on the program page.
Similar labsnearest by research summary
- USUC Berkeley
Amanda Coston
- Learning theory & optimization
- USUT Austin
Qiang Liu
- Learning theory & optimization
- CNZhejiang
Lei Cheng
- Learning theory & optimization
- CNZhejiang
Zemin Liu
- Graph learning
- USNYU
Siddharth Garg
- Trustworthy & safe ML
- CNTsinghua
Mingsheng Long
- Knowledge, QA & retrieval
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
MITcloses