Roger Grosse
Associate Professor
Studies neural network optimization, Bayesian learning and training data attribution for AI alignment.
Roger Grosse's lab at Toronto works on learning theory & optimization, probabilistic & causal inference and interpretability, evaluation & safety of LMs. 12 papers since 2023, 12 at top venues, mostly NeurIPS and ICML. 12 alumni, most often to Anthropic and Google. Applications close .
At a glance
- papers, 2023–2026
- 3/ year
- at T1 venues
- 100%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 12
Papers by yearfrom the lab homepage, 2023–2026
- 0 papers2026 · T1 0 · collecting
- 4 papers2025 · T1 4
- 5 papers2024 · T1 5
- 3 papers2023 · T1 3
T1 venuesOtherVenues: NeurIPS, ICML, ICLR
Research focuspaper titles that use each topic's words
- attribution3 papers
- Probabilistic & causal inference2 papers
- Learning theory & optimization1 paper
- Interpretability, evaluation & safety of LMs0 papers
- network optimization0 papers
Paper list
From the lab homepage.
20254 papers
- NeurIPS (T1)Better Training Data Attribution via Better Inverse Hessian-Vector Products (opens in a new tab)
- NeurIPS (T1)Distributional Training Data Attribution (opens in a new tab)
- NeurIPS (T1)Reducing the Probability of Bad Outputs in Language Models Using Probabilistic Inference
- NeurIPS (T1)What is Your Data Worth to GPT? LLM-Scale Data Valuation with Influence Functions (opens in a new tab)
20245 papers
- NeurIPS (T1)Connecting the Dots: LLMs can Infer and Verbalize Latent Structure from Disparate Training Data (opens in a new tab)
- ICML (T1)Measuring Stochastic Data Complexity with Boltzmann Influence Functions (opens in a new tab)
- ICML (T1)Probabilistic Inference in Language Models via Twisted Sequential Monte Carlo (opens in a new tab)
- ICLR (T1)REFACTOR: Learning to Extract Theorems from Proofs (opens in a new tab)
- NeurIPS (T1)Training Data Attribution via Approximate Unrolled Differentiation (opens in a new tab)
Where alumni went12 listed, 82% to industry
Industry 9Academia 2
Most often to
- Anthropic
- NVIDIA
Also to
- MIT
Apply to TorontoFall 2027
Computer Science MSc/PhD
- Round
- Single round (Dec 1)
- Opens
- —
- GRE
- optional
- English
- TOEFL/IELTS if applicable
- Fee
- CAD 135
Confirmed on the program page.
Similar labsnearest by research summary
- GBImperial
Roberto Bondesan
- Learning theory & optimization
- USUSC
Vatsal Sharan
- Learning theory & optimization
- CAToronto
J. Christopher Beck
- Learning theory & optimization
- USCMU
Rayid Ghani
- Trustworthy & safe ML
- USUMD
William Regli
- Interpretability, evaluation & safety of LMs
- USGeorgia Tech
Jesse Thomason
- Multimodal & grounded language
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.
Torontocloses