Publications
Selected recent publications and working papers. For a full list, see Google Scholar.
Alignment has a Fantasia Problem
Nathanael Jo, Zoe De Simone, Mitchell Gordon, and Ashia Wilson. arXiv
Pandora's Regret: A Proper Scoring Rule for Evaluating Sequential Search
Gerardo Flores, Yash Deshpande, Jannis R. Brea, and Ashia C. Wilson. arXiv
Creo: From One-Shot Image Generation to Progressive, Co-Creative Ideation
Zoe De Simone, Angie Boggust, Fredo Durand, Arvind Satyanarayan, and Ashia C. Wilson. arXiv
Evaluation without Generation: Non-Generative Assessment of Harmful Model Specialization with Applications to CSAM
Vinith Suriyakumar, Ayush Sekhari, Lena Stempfle, Robertson Wang, Michael Simpson, Rebecca Portnoff, Marzyeh Ghassemi, and Ashia C. Wilson. arXiv
Accelerated Convex Optimization via Hamiltonian Dynamics with Deterministic Integration Time
Xiuyuan Wang, Vishwak Srinivasan, Qiang Fu, Siddharth Mitra, Andre Wibisono, and Ashia C. Wilson.
A Consequentialist Critique of Binary Classification Evaluation: Theory, Practice, and Tools
Gerardo Flores, Abby Schiff, Alyssa Smith, Julia Fukuyama, and Ashia C. Wilson. arXiv
Interaction Context Often Increases Sycophancy in LLMs
Shomik Jain, Charlotte Park, Matheus Vianna, Dana Calacci, and Ashia C. Wilson. arXiv
Position: AI Evaluations Should be Grounded on a Theory of Capability
Nathan Jo and Ashia C. Wilson. arXiv
Near-Optimal Private Linear Regression via Iterative Hessian Mixing
Omri Lev, Vishwak Srinivasan, Moshe Shenfeld, Katrina Ligett, and Ashia C. Wilson. arXiv
Impacts of Semivalue-Based Data Curation on Fair Machine Learning Outcomes
Hannah Diehl, Justin Steil, and Ashia C. Wilson. paper
Aligning Evaluation with Clinical Priorities: Calibration, Label Shift, and Error Costs
Gerardo Flores, Alyssa Smith, Julia Fukuyama, and Ashia C. Wilson. arXiv
The Gaussian Mixing Mechanism: Renyi Differential Privacy via Gaussian Sketches
Omri Lev, Vishwak Srinivasan, Moshe Shenfeld, Katrina Ligett, Ayush Sekhari, and Ashia C. Wilson. arXiv
Homogeneous Algorithms Can Reduce Competition in Personalized Pricing
Nathan Jo, Kathleen Creel, Manish Raghavan, and Ashia C. Wilson. arXiv
Adaptive Backtracking Line Search
Laurent Lessard, Joao Cavalcanti, and Ashia C. Wilson. arXiv
High-accuracy sampling from constrained spaces with the Metropolis-adjusted Preconditioned Langevin Algorithm
Vishwak Srinivasan, Andre Wibisono, and Ashia C. Wilson. arXiv
Allocation Multiplicity: Evaluating the Promises of the Rashomon Set
Shomik Jain, Margaret Wang, Kathleen Creel, and Ashia C. Wilson. arXiv
Algorithmic Pluralism: A Structural Approach To Equal Opportunity
Shomik Jain, Vinith Suriyakumar, Kathleen Creel, and Ashia C. Wilson. arXiv
Scarce Resource Allocations That Rely On Machine Learning Should Be Randomized
Shomik Jain, Kathleen Creel, and Ashia C. Wilson. arXiv
Unstable Unlearning: The Hidden Risk of Concept Resurgence in Diffusion Models
Vinith Suriyakumar, Rohan Alur, Ayush Sekhari, Manish Raghavan, and Ashia C. Wilson. arXiv
LLM Output Homogenization is Task Dependent
Shomik Jain, Jack Lanchantin, Maximilian Nickel, Karen Ullrich, Jamelle Watson-Daniels, and Ashia C. Wilson. arXiv
UCD: Unlearning in LLMs via Contrastive Decoding
Vinith M. Suriyakumar, Ayush Sekhari, and Ashia C. Wilson. arXiv
Adaptive kernel selection for Stein Variational Gradient Descent
Moritz Melcher, Simon Weissmann, Jakob Zech, and Ashia C. Wilson. arXiv