Peking University (Organization): news timeline & CVEs

China's AI industry closes ranks as Deepseek ships open-source software for Huawei's Ascend chips

DeepSeek open-sourced TileLang and chip libraries for Huawei's Ascend AI processors.

DeepSeek and Huawei released open-source programming tools for Huawei Ascend chips, including computation and inter-chip data-movement libraries built around TileLang, which DeepSeek says is easier to program than Nvidia CUDA while still using the hardware fully. The companies optimized a supernode of 128 Ascend 950 chips. TileLang began at Peking University, DeepSeek has used it for about a year, and it is now the company's main tool for AGI work. The partnership targets China's software gap with CUDA as Huawei says new processors will be widely used for training next year and plans fewer overseas chip sales because domestic demand exceeds supply.

The DecoderSep 30, 14:37 UTC · 1d agoAI industry

DeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imagination

DeCAL, a contact-aware dexterous vision-language-action model with visuo-tactile fusion, reports 71% average task success.

DeCAL is a physically-grounded dexterous vision-language-action (VLA) model built on a Mixture-of-Transformers architecture with specialized experts for understanding, imagination, and action generation. It introduces Adaptive Visuo-Tactile Fusion with contact-aware gating and Visuo-Tactile Latent Co-Imagination to jointly model visual and tactile dynamics. It reports state-of-the-art results with a 71% average success rate and 83.4% progress success rate, plus generalization to unseen scenarios.

arXiv cs.AI / cs.LG / cs.CLSep 8, 17:47 UTC · 23d agoAI research1

Peking University

Timeline2 articles

Appears with

China's AI industry closes ranks as Deepseek ships open-source software for Huawei's Ascend chips

DeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imagination