AI agents can play StarCraft II now? AVA, proposed in this paper from @CamelAIOrg , can execute complex tactical maneuvers without explicit training, achieving comparable performance to traditional MARL methods that require substantial training iterations. A win for VLM and RL!
@jiqizhixin
-
Sharing of a Scientific Article from arxiv on March 24, 2025
By
–
Paper: https://arxiv.org/pdf/2503.16743
-
SuperARC Benchmark Shows AI Far from AGI Achievement
By
–
According to the new benchmark, SuperARC, AI is still far from achieving AGI or ASI.
-
Why Apple Hasn’t Built Its Own LLM Yet
By
–
Google has Gemini. Meta released Llama. Amazon built Nova. Alibaba open-sourced Qwen. Tencent created Hunyuan.
So the real question is—why hasn’t Apple built its own LLM yet? -
Andrej Karpathy Launches New Blog Sharing AI Research Notes
By
–
Andrej Karpathy has launched a new blog to share his "random" notes. Add it to your RSS list—definitely worth following.
Link: -
Unitree G1 Robot Demonstrates Kip-Up and Tai Chi Kung Fu
By
–
Unitree just dropped a new video showcasing the G1 robot pulling off a kip-up and performing Tai Chi Kung Fu. Impressive moves! 🤖🔥 It even took a kick from a human and kept going! 🤣 https://t.co/Xn2vxVlW2M
— 机器之心 JIQIZHIXIN (@jiqizhixin) 21 mars 2025Unitree just dropped a new video showcasing the G1 robot pulling off a kip-up and performing Tai Chi Kung Fu. Impressive moves! It even took a kick from a human and kept going!
-
New Scientific Publication on arXiv from March 21, 2025
By
–
Paper: https://arxiv.org/pdf/2503.14554
-
Asynchronous RL Agents Learn Faster, Achieve Higher Returns
By
–
A not-so-surprising finding: with asynchronous RL, agents learn faster and achieve significantly higher returns. Still, it's nice to see it empirically verified.
-
Stanford’s Gaussian Atlas: 3D Scene Representation Breakthrough
By
–
Project Page: https://
cs.stanford.edu/~xtiange/proje
cts/gaussianatlas/
…
Paper: https://
arxiv.org/abs/2503.15877 -
GaussianVerse: Large-Scale 3D Gaussian Dataset Released
By
–
They also built GaussianVerse, a large-scale dataset with 205K high-quality 3D Gaussian fittings of various 3D objects.