Vardan Papyan
Assistant Professor
Studies the theory of deep learning, including neural collapse.
Vardan Papyan's lab at Toronto works on learning theory & optimization. 11 papers since 2023, 11 at top venues, mostly ICLR and ICML. 9 alumni, most often to Boson AI and Bluekey AI. Applications close .
At a glance
- papers, 2023–2026
- 3/ year
- at T1 venues
- 100%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 9
Papers by yearfrom DBLP, 2023–2026
- 0 papers2026 · T1 0 · collecting
- 5 papers2025 · T1 5
- 5 papers2024 · T1 5
- 1 papers2023 · T1 1
T1 venuesOtherVenues: ICLR, ICML, NeurIPS
Research focuspaper titles that use each topic's words
- neural2 papers
- Learning theory & optimization0 papers
- mathematics0 papers
- deep learning0 papers
- theory0 papers
Paper list
From DBLP.
20255 papers
- NeurIPS (T1)Attention Sinks: A 'Catch, Tag, Release' Mechanism for Embeddings (opens in a new tab)
- ICLR (T1)Breach By A Thousand Leaks: Unsafe Information Leakage in 'Safe' AI Responses (opens in a new tab)
- ICLR (T1)OATS: Outlier-Aware Pruning Through Sparse and Low Rank Decomposition (opens in a new tab)
- ICML (T1)On the Importance of Gaussianizing Representations (opens in a new tab)
- ICLR (T1)Transformer Block Coupling and its Correlation with Generalization in LLMs (opens in a new tab)
20245 papers
- NeurIPS (T1)Linguistic Collapse: Neural Collapse in (Large) Language Models (opens in a new tab)
- ICML (T1)Out of the Ordinary: Spectrally Adapting Regression for Covariate Shift (opens in a new tab)
- ICML (T1)Position: Fundamental Limitations of LLM Censorship Necessitate New Approaches (opens in a new tab)
- ICLR (T1)Pushing Boundaries: Mixup's Influence on Neural Collapse (opens in a new tab)
- ICML (T1)Sparsest Models Elude Pruning: An Exposé of Pruning's Current Capabilities (opens in a new tab)
20231 paper
Where alumni went9 listed, 83% to industry
Industry 5Academia 1
Most often to
- Boson AI
- Bluekey AI
- Cohere
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
- SGNTU
Yang Liu
- Trustworthy & safe ML
- USStanford
John Duchi
- Learning theory & optimization
- SGNTU
Dacheng Tao
- Learning theory & optimization
- CNSJTU
Jia Wang
- Learning theory & optimization
- USUIUC
Varun Chandrasekaran
- Trustworthy & safe ML
- CNTsinghua
Hang Su
- Trustworthy & safe ML
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- DBLP
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
Torontocloses