NEW paper from NVIDIA. They discuss robot programming that compounds experience instead of throwing it away. Traditional robot programming forces you to orchestrate perception, contact dynamics, diverse configurations, and constant execution failures by hand. Most learned approaches then bury what they learned in opaque weights. ASPIRE runs a code-as-policy loop that autonomously writes and refines control programs. A closed-loop execution engine exposes fine-grained multimodal traces, so the system diagnoses its own failures, synthesizes repairs, and validates them. Validated fixes distill into a reusable skill library, and evolutionary search explores diverse task sequences beyond single-trajectory tuning. ASPIRE gains up to 77 percent on LIBERO-Pro under perturbation, 72 percent on Robosuite bimanual handover, and 32 percent on BEHAVIOR-1K. On LIBERO-Pro Long it hits 31 percent zero-shot versus 4 percent for prior methods, with early sim-to-real transfer across embodiments. Paper: arxiv.org/abs/2607.00272 Learn to build effective AI agents in our academy: academy.dair.ai/

ENPIRE -> ASPIRE, our 2nd work in the series for Physical AutoResearch. We are building the components for robot self-improvement, one /skill at a time.

Thanks @DrJimFan ! Please check our project website for the details: research.nvidia.com/labs/gea…

/𝚜𝚔𝚒𝚕𝚕 for Robotics! Excited to share our project ASPIRE, Instead of solving each task from scratch, ASPIRE lets robots continuously accumulate reusable skills that transfer across tasks, embodiments, and sim-to-real settings. Check @guanzhi_wang 's post for the details!

ML ENERGY x NVIDIA blog post is live! Power is a binding constraint for AI factories, and energy efficiency converts directly to revenue. This collaborative blog post covers our works on providing energy observability and improving performance per watt: - Understanding and optimizing inference energy consumption with the ML ENERGY Benchmark & Leaderboard - Optimizing training energy consumption with Perseus & Kareus Our works are open-source on GitHub, and we look forward to more collaborations with NVIDIA! developer.nvidia.com/blog/ma… @NVIDIAAI @NVIDIAAIInfra

Great work!

Energy and power are first-class resources in scaling AI compute. ML.ENERGY builds open-source infrastructure for measuring, understanding, and optimizing the energy use of ML workloads. Start here: - ml.energy - github.com/ml-energy

Reinstate Food Trucks on the University of Michigan's North Campus - Sign the Petition! chng.it/DPN6dNZpbN via @Change

How to leverage the white-box info (i.e. source code) for fuzzing compilers? Check out our work “WhiteFox 🦊: White-Box Compiler Fuzzing Empowered by Large Language Models” at OOPSLA 2024! w/ @yinlin_deng, @lry89757, JIayi Yao, @JiaweiLiu_, @Reyhaneh, and @LingmingZhang (1/N)

We are excited to announce our lab's papers at #ICML2024! 🧠✨ Come and discuss our latest research from LLM evaluation to efficient LLM serving & inference! See you there! 1️⃣ Poster: MuxServe: Flexible Spatial-Temporal Multiplexing for Multiple LLM Serving 📍 Location & Time: poster session 1 — Hall C 4-9 #816, 11:30 AM on Tuesday July 23 📜 TL;DR: MuxServe Boosts multiple LLM serving throughput by up to 1.8x through flexible spatial-temporal multiplexing. 2️⃣ Poster: Break the Sequential Dependency of LLM Inference Using Lookahead Decoding 📍 Location & Time: poster session 2 — Hall C 4-9 #411, 1:30 PM on Tuesday July 23 📜 TL;DR: An exact and parallel decoding algorithm that accelerates LLM decoding without needing auxiliary models or data stores. 3️⃣ Poster: Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference 📍 Location & Time: poster session 3 — Hall C 4-9 #709, 11:30 AM on Wednesday July 24 📜 TL;DR: Chatbot Arena is an open platform for evaluating LLMs based on human preferences through crowdsourced pairwise comparisons, and it’s becoming a widely cited leaderboard for its robust and credible evaluation methods. 4️⃣ Poster: CLLMs: Consistency Large Language Models 📍 Location & Time: poster session 4 — Hall C 4-9 #604, 1:30 PM on Wednesday July 24 📜 TL;DR: We introduce a new family of LLMs optimized for fast Jacobi decoding, achieving a 2.4x to 3.4x improvement in generation speed across multiple benchmarks without compromising quality. 5️⃣ Poster: Online Speculative Decoding 📍 Location & Time: poster session 5 — Hall C 4-9 #605, 11:30 AM on Thursday July 25 📜 TL;DR: OSD improves the efficiency of large language model inference by continuously updating the draft models with user query data, resulting in a significant reduction in latency and an increase in token acceptance rates. 6️⃣ Poster: InferCept: Efficient Intercept Support for Augmented Large Language Model Inference 📍 Location & Time: poster session 5 — Hall C 4-9 #709, 11:30 AM on Thursday July 25 📜 TL;DR: InferCept is the first inference framework for augmented LLMs, efficiently serving LLMs that can query tools, ML models, and virtual environments.

Great work under the leadership of @jiangfeiduan and @haozhangml !

Zhou, we're so proud of you 💚🥹 #ChineseGP

Still optimizing throughput for LLM Serving? Think again: Goodput might be a better choice! Splitting prefill from decode to different GPUs yields - up to 4.48x goodput - up to 10.2x stricter latency criteria Blog: hao-ai-lab.github.io/blogs/d… Paper: arxiv.org/abs/2401.09670

Holy moly Starknet Provision just came my way! 😎 I am eligible for STRK! Getting ready to claim 👨‍💻👩‍💻 You might be eligible, too! provisions.starknet.io #Starknet

P19 ➡️ P9 What a shift put in by @ZhouGuanyu24 today. 👏 #QatarGP #F1

🤣

:wq R.I.P Bram Moolenaar groups.google.com/g/vim_anno…

🧐 discuss.python.org/t/a-steer…

#LLVM LLVM的Debug手段真的太原始了😇，逐层对比开发，看十几万行的Debug Log，开发效率太低了😭

可能是因为我就是垃圾😫