Machine Learning

Balancing Efficiency and Expressiveness: Subgraph GNNs with Walk-Based Centrality

Gradient Boosting Reinforcement Learning

IT^3: Idempotent Test-Time Training

Policy Gradient via Tree Expansion

RL-Initialized Genetic Algorithms for Faster Vehicle Routing

Foldable SuperNets: Scalable Neural Merging with Different Initialization and Tasks

Add-it: Training-Free Object Insertion in Images With Pretrained Diffusion Models

Analyzing Large Language Models by Learning on Token Distribution Sequences

ComfyGen: Prompt-Adaptive Workflows for Text-to-Image Generation

From Glucose Patterns to Health Outcomes (Workshop): A Generalizable Foundation Model for Continuous Glucose Monitor Data Analysis

Workshop version of the CGM foundation model paper (later published in Nature, 2025).