Poster Session I (12:00 PM - 1:30 PM)

Policy evaluation in reinforcement learning: The impact of temporal dependence and multi-step lookahead Presenter: Yaqi Duan Indiscriminate Poisoning Attacks on Unsupervised Contrastive Learning Presenter: Hao He Adversarial Policies Beat Superhuman Go AIs Presenter: Tony Wang On counterfactual inference with unobserved confounding Presenter: Abhin Shah Dementia Detection using Text and Speech from Neuropsychological Exams Presenter: Nauman Dawalatabad Clinical relevance of generative counterfactual explanations Presenter: Swami Sankaranarayanan Root-cause analysis and repair of failures in autonomous decision-making Presenter: Charles Dawson ObSynth: An Interactive Synthesis System for Generating Object Models from Natural Language Specifications Presenter: Alex Gu A Usability Study of Nomon: a Flexible Interface for Motor-Impaired Users Presenter: Nicholas Bonaker Distilling Model Failures as Directions in Latent Space Presenter: Saachi Jain Reproducibility in Optimization: Theoretical Framework and Limits Presenter: Kwangjun Ahn Straightening Out the Straight-Through Estimator: Overcoming Optimization Challenges in Vector Quantized Networks Presenter: Minyoung Huh Redeeming intrinsic rewards via constrained optimization Presenter: Zhang-Wei Hong Characterizing Intrinsic Compositionality in Transformers with Tree Projections Presenter: Pratyusha Sharma Automatic Gradient Descent: Deep Learning without Hyperparameters Presenter: Jeremy Bernstein Quasimetric Reinforcement Learning Presenter: Tongzhou Wang

---

Poster Session II (4:00 PM - 5:30 PM)

TRAK: Attributing Model Behavior at Scale Presenter: Sung Min (Sam) Park Multi-fidelity time-optimal trajectory generation Presenter: Gilhyun Ryou Learning Representations for Control: A Case Study in Linear Systems Presenter: Yi Tian Language Entailment as Robust Self-learners Presenter: Hongyin Luo SimPer: Simple Self-Supervised Learning of Periodic Targets Presenter: Yuzhe Yang Datamodels: Predicting Predictions with Training Data Presenter: Andrew Ilyas The Exact Sample Complexity Gain from Invariances for Kernel Regression on Manifolds Presenter: Behrooz Tahmasebi Is CLIP Fooled by Optical Illusions? Presenter: Jerry Ngo On Bridging The Gap Between Mean Field And Finite Width In Deep Random Neural Networks With Batch Normalization Presenter: Hadi Daneshmand Learning Surrogates for Diverse Vehicle Emission Models Presenter: Edgar Ramirez Sanchez Erasing Concepts from Diffusion Models Presenter: Joanna Materzynska Powderworld: A Platform for Understanding Generalization via Rich Task Distributions Presenter: Kevin Frans Estimation of Models with Limited Data by Leveraging Shared Structure Presenter: Maryann Rui Gaussian processes at the Helm(holtz): A more fluid model for ocean currents Presenter: Renato Berlinghieri Judging Facts, Judging Norms: Training Machine Learning Models to Judge Humans Requires a Modified Approach to Labeling Data Presenter: Aparna Balagopalan Is Conditional Generative Modeling all you need for Decision Making? Presenter: Anurag Ajay UniverSeg: Universal Medical Image Segmentation Presenter: Victor Butoi Temporal Logic Imitation: Learning Plan-Satisficing Motion Policies from Demonstrations Presenter: Felix Yanwei Wang Diversity and Expertise in Binary Votes Aggregation Presenter: Manon Revel