Ashia Wilson
Associate Professor
Develops foundations for efficient and safe machine learning, including optimization and sampling methods, AI safety, and valid evaluation of AI systems.
Ashia Wilson's lab at MIT works on learning theory & optimization, trustworthy & safe ML and interpretability, evaluation & safety of LMs. 23 papers since 2023, 17 at top venues, mostly ICML and NeurIPS. 7 PhD students. 5 alumni. Applications close .
At a glance
- papers, 2023–2026
- 6/ year
- at T1 venues
- 74%
- PhD students
- 7
- MS students
- 0
- alumni listed
- 5
Papers by yearfrom the lab homepage, 2023–2026
- 12 papers2026 · T1 9 · collecting
- 6 papers2025 · T1 4
- 4 papers2024 · T1 3
- 1 papers2023 · T1 1
T1 venuesOtherVenues: ICML, NeurIPS, FAccT, COLT, TMLR
Research focuspaper titles that use each topic's words
- Learning theory & optimization7 papers
- Interpretability, evaluation & safety of LMs2 papers
- Trustworthy & safe ML0 papers
- sampling methods0 papers
- develops foundations0 papers
Paper list
From the lab homepage.
202612 papers
- AISTATS (T1)A Consequentialist Critique of Binary Classification Evaluation: Theory, Practice, and Tools
- COLT (T1)Accelerated Convex Optimization via Hamiltonian Dynamics with Deterministic Integration Time
- PNAS (T1)Efficient and Accurate Steering of Large Language Models through Attention-Guided Feature Learning
- FAccTImpacts of Semivalue-Based Data Curation on Fair Machine Learning Outcomes
- TMLR (T1)Influence Estimation in Statistical Models Using the Fisher Information Matrix
- CHI (T1)Interaction Context Often Increases Sycophancy in LLMs
- ICML (T1)Near-Optimal Private Linear Regression via Iterative Hessian Mixing
- AISTATS OPTIMAL WORKSHOPPandora's Regret: A Proper Scoring Rule for Evaluating Sequential Search
- ICML (T1)Position: AI Evaluations Should be Grounded on a Theory of Capability
- NEURIPS BIALIGN WORKSHOPPosition: Alignment has a Fantasia Problem
- TMLR (T1)Unstable Unlearning: The Hidden Risk of Concept Resurgence in Diffusion Models
- NeurIPS (T1)When Are Semivalue-Based Decisions Identifiable? Robust Data Selection under Utility Ambiguity
20256 papers
- ICLR (T1)Adaptive Backtracking Line Search
- NeurIPS (T1)Aligning Evaluation with Clinical Priorities: Calibration, Label Shift, and Error Costs
- FAccTAllocation Multiplicity: Evaluating the Promises of the Rashomon Set
- ALTHigh-accuracy sampling from constrained spaces with the Metropolis-adjusted Preconditioned Langevin Algorithm
- NeurIPS (T1)Homogeneous Algorithms Can Reduce Competition in Personalized Pricing
- NeurIPS (T1)The Gaussian Mixing Mechanism: Renyi Differential Privacy via Gaussian Sketches
20244 papers
- FAccTAlgorithmic Pluralism: A Structural Approach to Equal Opportunity
- COLT (T1)Fast Sampling from Constrained Spaces Using the Metropolis-Adjusted Mirror Langevin Algorithm
- ICML (T1)Mean-field Underdamped Langevin Dynamics and its Spacetime Discretization
- ICML (T1)Scarce Resource Allocations That Rely on Machine Learning Should Be Randomized
20231 paper
- ICML (T1)Accelerated Stochastic Optimization Methods under Quasar-convexity
Where alumni went5 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
- USNYU
Allen Liu
- Learning theory & optimization
- CNZhejiang
Yueting Zhuang
- Knowledge, QA & retrieval
- DETUM
Zeynep Akata
- Trustworthy & safe ML
- USMichigan
Atul Prakash
- Trustworthy & safe ML
- CNTsinghua
Kaifeng Lyu
- Learning theory & optimization
- 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.
MITcloses