VOL. 030
8 research signalsDAILY EDITION / SAVED SNAPSHOT
A dated Agentic RL Daily snapshot with 8 verified primary-source signals across papers, official releases, deployment evidence, and safety or alignment findings.
EDITOR'S VIEW
FULL EDITION
Archived / 2026-08-16
01
Heterogeneity-Aware Belief Synchronization for Semantic Communication in AI-Native 6G Networks
A verified papers signal preserved from the original daily snapshot. The primary source is linked for full context, while the archive keeps the original publication date and source attribution intact.
02
Capability Sheaves for Compositional Agent-Harness Repair: Controlled Quotients and a Real-Repository Stress Test
Agent harnesses combine retrieval, routing, state, provenance, and verification, but locally successful components may disagree on shared state.
03
Vero: Can AI Agents Build Formally Verified Software Repositories?
AI agents are increasingly used for programming, but do not provide any guarantee on the correctness of generated code.
04
MARC v1: An Open-Source Multi-Agent Framework for Clinical AI Reasoning and Coordination
We present Multi-Agent Reasoning and Coordination (MARC), an open-source framework that replaces monolithic LLM prompting with deterministic multi-agent orchestration for clinical reasoning.
05
AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Design
Transforming multimodal sources into condensed and structured media outputs can be fundamentally conceptualized as a long-horizon agentic process centered on a model-harness system.
06
UniTexture: Cross-Task Universal Adversarial Textures for Vision-Language-Action Models
Vision-Language-Action (VLA) models have emerged as generalist robotic policies capable of following diverse language instructions and performing a wide range of manipulation tasks.
07
Teach the Magnitude, Not the Direction: Verifier-Bounded Credit Assignment for Multi-Turn Multi-step LLM Agents
Reinforcement learning with verifiable rewards (RLVR) offers a verifier-bounded performance ceiling for training multi-turn tool-use agents, yet its trajectory-level credit assignment conflates heterogeneous per-turn outcomes into a single reward signal.
08
verl-project/verl v0.9.0
官方来源补充信号：# v0.9.0 ## Highlights ### Training #### Megatron - DeepSeek-V4 GRPO end-to-end with Megatron-Bridge actor/ref, vLLM rollout and FP8/MXFP4 weight transfer (#6473), plus a contiguous context-parallel layout (#7221) and CP fixes that make long-context DeepSeek-V4 runnable (#7297). - [Megatron Lite (`mlite`)](https://verl.readthedocs.io/en/latest/advance/megatron_lite_backend.html) backend for DeepSeek-V4, GLM-5 and Kimi-K2.5/K2.6, with 256-GPU GRPO launchers (#6791, #7091). - Muon optimizer support via Megatron-Core `TensorParallelMuon`, with AdamW fallback for non-2D params and an opt-in `muon_