Bo Chen

CSRv2: Unlocking Ultra-Sparse Embeddings

Scaling Attention via Feature Sparsity

GOAL: Balance Multimodal Learning with Gradient Orthogonalization and Adaptive Leveraging

Turning Uncertainty into Control: Bi-level Training with Editable Bayesian Layers

Route Experts by Sequence, Not by Token

Channel Matters: Estimating Channel Influence for Multivariate Time Series

Beyond Matryoshka: Revisiting Sparse Coding for Adaptive Representation

OmiAD: One-Step Adaptive Masked Diffusion Model for Multi-class Anomaly Detection via Adversarial Distillation

Enhancing Uncertainty Estimation and Interpretability with Bayesian Non-negative Decision Layer

Advancing Graph Generation through Beta Diffusion

Channel-wise Influence: Estimating Data Influence for Multivariate Time Series