10 papers
The Seriality Gap in Video Diffusion Models
Jorge Diaz Chao, Konpat Preechakul, Yuxi Liu +1
The paper investigates why video diffusion models struggle with tasks that require sequential causal reasoning, such as multi‑ball collisions, and identifies a "seriality gap" wher…
The Serial Scaling Hypothesis
Yuxi Liu, Konpat Preechakul, Kananart Kuwaranancharoen +1
While machine learning has advanced through massive parallelization, we identify a critical blind spot: some problems are fundamentally sequential. These "inherently serial" proble…
Predicting kernel regression learning curves from only raw data statistics
Dhruva Karkada, Joseph Turnbull, Yuxi Liu +1
We study kernel regression with common rotation-invariant kernels on real datasets including CIFAR-5m, SVHN, and ImageNet. We give a theoretical framework that predicts learning cu…
ClimaOoD: Improving Anomaly Segmentation via Physically Realistic Synthetic Data
Yuxing Liu, Zheng Li, Huanhuan Liang +3
Anomaly segmentation seeks to detect and localize unknown or out-of-distribution (OoD) objects that fall outside predefined semantic classes a capability essential for safe autonom…
RM-RL: Role-Model Reinforcement Learning for Precise Robot Manipulation
Xiangyu Chen, Chuhao Zhou, Yuxi Liu +1
Precise robot manipulation is critical for fine-grained applications such as chemical and biological experiments, where even small errors (e.g., reagent spillage) can invalidate an…
Trace-Focused Diffusion Policy for Multi-Modal Action Disambiguation in Long-Horizon Robotic Manipulation
Yuxuan Hu, Xiangyu Chen, Chuhao Zhou +4
Generative model-based policies have shown strong performance in imitation-based robotic manipulation by learning action distributions from demonstrations. However, in long-horizon…