Sinno Jialin Pan

Online Reusable Resource Allocation with Adversarial Requests

HarmoMoE: Unifying Domain-Specialized Experts into a Mixture-of-Experts Model under Privacy Constraints

Deep Coupling Learning for Solving PDEs

Robust Equation Structure Learning with Adaptive Refinement

Beyond Speedup - Utilizing KV Cache for Sampling and Reasoning

Global and Local Topology-Aware Graph Generation via Dual Conditioning Diffusion

Adversarial Unlearning of Poisoned Features for Backdoor Defense in Federated Learning

Diverse Text Decoding via Iterative Reweighting

Bridging Successor Measure and Online Policy Learning with Flow Matching-Based Representations

Neural Operator-based Curriculum Learning for Physics-Informed Neural Networks

Sparse-Compression Diffusion Models

FuseGPT: Prune-and-Fuse Knowledge Redistribution for Efficient Transformers

Analytical Restructuring of Feed-Forward Networks for Accelerated LLM Inference

GLID$^2$E: A Gradient-Free Lightweight Fine-tune Approach for Discrete Biological Sequence Design

Probabilistic Neural Pruning via Sparsity Evolutionary Fokker-Planck-Kolmogorov Equation

MetaDefense: Defending Fine-tuning based Jailbreak Attack Before and During Generation

Backdoor Attacks on Dense Retrieval via Public and Unintentional Triggers

Beyond Self-Interest: How Group Strategies Reshape Content Creation in Recommendation Platforms?

Parameter and Memory Efficient Pretraining via Low-rank Riemannian Optimization

KVTuner: Sensitivity-Aware Layer-Wise Mixed-Precision KV Cache Quantization for Efficient and Nearly Lossless LLM Inference