Jie Zhou

Hidden Decoding achieves unprecedented performance improvements in large language models by scaling computation along the sequence length without modifying the Transformer architecture.

UltraX achieves the highest average performance across datasets while using fewer training tokens, redefining efficiency in data refinement for LLMs.

IBRSteG enables the undetectable embedding of secret 3D scenes with superior capacity and security, all while eliminating the need for scene-specific optimization.

Threshold-sensitive KV cache pruning is out; ReFreeKV's adaptive approach achieves robust memory efficiency without predefined limits.

RAVA reveals that incorporating geometric evidence can drastically enhance viewpoint alignment in image generation, outperforming traditional methods that rely solely on semantic embeddings.

UoU revolutionizes fingerprint recognition by enabling a universal model that reuses representations across diverse applications, challenging the limitations of traditional task-specific approaches.

FlowRAG transforms how we approach multi-hop reasoning by leveraging a quad-level graph structure that enhances both semantic recall and explicit reasoning paths.

R2RDreamer achieves spatial generalization improvements in manipulation tasks by leveraging 3D-aware data augmentation without the pitfalls of complex scene setups or sim-to-real gaps.

Agents-K1 transforms how we extract and reason about scientific knowledge, achieving superior performance in multi-hop reasoning tasks compared to existing methods.

Forget full attention: a hybrid sparse-linear attention model, MiniCPM-SALA, achieves 3.5x faster inference and supports 1M context length on a single GPU, all while maintaining comparable performance.