Yang Ren

Closing nearly 90% of performance gaps in low-bit quantized models with a compact repair codec could redefine efficiency in language model deployment.

UniHEAR boosts retrieval performance in KB-VQA by 6.7 points, overcoming the limitations of single-source retrieval and blind reranking.

SNNs can achieve higher accuracy and lower latency by learning the optimal spiking resolution for each layer, rather than relying on predefined burst structures.