2026年7月30日 技术热点总结

项目简介：AxisRL is an agentic RL post-training framework built on SGLang rollout, Megatron training, and real-world agent workflows.

技术栈：Python

核心介绍：English · 简体中文 XYZ AI Lab · Technical Blog AxisRL is an agentic RL post-training framework built on SGLang rollout, Megatron training, and real-world agent workflows. AxisRL connects high-throughput rollout, large-scale training, weight synchronization, data movement, resource scheduling, and reproducible debugging inside one coherent framework. SGLang and Megatron remain the core serving and training engines; AxisRL handles the …

项目简介：Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

技术栈：JavaScript、agent-skills、claude-code、claude-code-plugin、claude-code-skill、claude-skills、cursor-rules、ponytail、prompt-engineering

核心介绍：Ponytail He says nothing. He writes one line. It works. Install Before / after How it works Numbers Commands FAQ You know him. Long ponytail. Oval glasses. Been at the company longer than the version control. You show him fifty lines; he says nothing and replaces them with one. Ponytail puts him inside your AI agent. Skills p…

项目简介：Bringing real-time voice to life.

技术栈：JavaScript

核心介绍：Persona A realtime character presence for desktop voice experiences. Persona is a cross-platform desktop character that gives voice conversations an expressive visual identity alongside your work. Linux requires pw-dump and pw-record on PATH. Windows process-loopback requires Windows 10 build 20348 or newer. macOS asks once for System Audio Recording permission. Each listener is scoped to the supported application’s playback process. Persona

项目数据：⭐ 482 Stars，🍴 43 Forks

项目简介：AxisAgentic: An Extensible Runtime and Trajectory-Collection Framework for Long-Horizon Agents.

技术栈：Python

核心介绍：English · 简体中文 XYZ AI Lab · Technical Report

项目数据：⭐ 474 Stars，🍴 72 Forks

Hyperspectral imaging (HSI) is useful for material discrimination, but operational mine screening also depends on how many false alarms must be inspected before targets are found. This paper studies PFM-1 landmine detection in unmanned aerial vehicle (UAV) visible and near-infrared (VNIR) HSI using spectral angle mapper (SAM), matched filter (MF), adaptive coherence estimator (ACE), and constrained energy minimization (CEM). We compare a ground-measured SVC signature, a fully informed in-scene core-pixel signature, and a simulated human-in-the-loop signature bootstrap. Besides receiver oper…

In RLHF pipelines, reward scoring blocks policy updates. Slow scoring bottlenecks the entire loop, since no update runs until every rollout gets a score. And yet most setups just default to PyTorch eager mode or torch.compile, no one checks if that's actually fastest. Scoring itself is small. Rollout generation eats far more of a typical RLHF step. But scoring and generation fight over the same CPU and GPU resources, so a faster scoring engine doesn't shrink step time on its own. It mainly frees up capacity generation can use instead. We built a native C++ inference engine on ONNX Runtime. …

A language model p_θ(y mid x) trained on reasoning tasks learns to solve problems via multiple distinct strategies, yet these strategies are implicit and entangled within the model's response distribution. We study the problem of decomposing the response distribution of a given pretrained language model into a structured, strategy-conditioned representation. Specifically, we learn a latent-variable factorization p_θ(y mid x) leadsto (r_ϕ(z mid x), g_ϕ(y mid x,z)), where a router r maps each input to a distribution over latent strategies z and a generator g produces the response conditioned …

RL for code correctness is now established: have the model generate a program, run it against hidden test cases, and reward solutions that pass. Extending this to code optimization seems straightforward: just add execution time to the reward. But in practice, once timing drives the reward, small problems in measurement noise, reward sparsity, or GRPO instability overwhelm the signal and make RL fail: generated solutions are barely faster, and more of them can fail. We make execution time learnable through three stages: (1) how code is tested, by building DMC-Optim with large optimization te…

Biomedical image analysis spans diverse modalities and tasks, yet real-world deployment is hindered by severe distribution shifts across scanners, protocols, and patient populations. High-performing models consequently require repeated domain-specific fine-tuning, which is a costly cycle that becomes impractical when labels are scarce or privacy constraints limit data sharing. We propose OPERA (Offline Policy-guided Expert Routing and Adaptation), a multi-agent ensemble framework that addresses this deployment bottleneck by treating expert weight assignment as an offline policy learning pro…

Modern coding agents are usually evaluated by whether they eventually produce a correct patch, but patch generation depends on an earlier context-acquisition stage: finding the repository files needed for the task. We introduce Agent Retrieval Bench, a file-level benchmark for this upstream retrieval problem. Samples are built from real coding-workflow signals and evaluated against frozen base-commit repositories, with relevance defined by what an agent needs next rather than direct query-file semantic similarity. The benchmark covers four positive-retrieval tasks: code2test, comment2contex…