arXiv Machine Learning September 2, 2026 By Zinco J, Xunjie Zhu, Shen Huang, Zhenyi Wang, Pengjun Xie, Jieping Ye

MemoryWalker: Stop Training Agents on Contexts They Never Saw

Read the original on arXiv Machine Learning →

MemoryWalker addresses the conditioning problem that arises when training agents with compressed context during rollout. It introduces two exact, gradient‑equivalent corrections—LogitTree, a segmented K‑forward traversal, and a packed 4D attention mask—alongside SDCC, a self‑distillation method that reduces the train‑deployment gap by minimizing KL divergence at each eviction. Experiments on seven web‑search benchmarks show that SDCC significantly lowers logit drift and boosts rollout rewards compared to naive training.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv AI

Jul 24

LinearARD: Linear-Memory Attention Distillation for RoPE Restoration

arXiv:2604. 00004v2 Announce Type: replace-cross Abstract: The extension of context windows in Large Language Models is typically facilitated by scaling positional encodings followed by lightweight Continual Pre-Training (CPT).

By Ning Yang, Hengyu Zhong, Wentao Wang, Baoliang Tian, Haijun Zhang, Jun Wang

llms rag efficiency benchmarks

arXiv AI

Aug 19

LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents

LEGO-RL is a framework that connects native coding-agent harnesses with scalable policy‑gradient training without altering the harnesses’ internal flow. It achieves faithful optimization through in‑process LLM proxying, reliable execution via sandbox orchestration, and observable training with automated validation and a Live UI. Experiments show LEGO‑RL improves the Qwen3.5‑35B‑A3B model’s performance on three native harnesses while preserving high rollout‑training probability correlation.

By Yiming Du, Yuxin Jiang, Tao Yuan, Jianbo Dai, Shaowei Wang, Jierun Chen, Chaofan Tao, Xianzhi Yu, Lifeng Shang, Kam-Fai Wong, Xiaohui Li, Haoli Bai

llms agents reinforcement-learning safety

arXiv AI

Jul 15

Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models

arXiv:2607. 12463v1 Announce Type: new Abstract: Coding agents must integrate external tool returns into ongoing reasoning - a capability that standard left-to-right pretraining on code exposes only in its forward direction.

By Yubo Wang, Jiarong Liang, Yuxuan Zhang, Xuye Liu, Cong Wei, Yuyu Zhang, Ping Nie, Wenhu Chen

agents benchmarks safety

arXiv Machine Learning

Aug 10

AsyncWebRL: Efficient Asynchronous Reinforcement Learning for Multi-Step Visual Web Agents

arXiv:2606. 05597v3 Announce Type: replace Abstract: Training vision-language web agents with multi-step RL is compute-intensive, with two dominant forms of inefficiency: idle GPUs in synchronous RL, and trajectories that use more steps and tokens than necessary.

By Hao Bai, Rui Yang, Chenlu Ye, Spencer Whitehead, Aviral Kumar, Tong Zhang

agents reinforcement-learning efficiency multimodal

arXiv Machine Learning

Jul 20

Looped Latent Attention: Cross-Loop KV Compression for Looped Transformers

arXiv:2607. 15456v1 Announce Type: new Abstract: Looped, weight-tied Transformers reduce parameters by reusing a block, but decoding still stores a separate K/V cache for every recurrence step.

By James O' Neill, Fergal Reid

llms efficiency

arXiv AI

Aug 11

WDL-OPD: Weak-Driven On-Policy Distillation via Mixture-Constrained Co-Training

arXiv:2608. 09447v1 Announce Type: cross Abstract: On-policy distillation (OPD) aligns a student with a teacher on trajectories sampled from the student itself, reducing the train-test state mismatch of offline distillation.

By Zehao Chen, Gongxun Li, Tianxiang Ai, Yifei Li, Zixuan Huang, Wang Zhou, Tao Huang, Fuzhen Zhuang, Xianglong Liu, Jianxin Li, Deqing Wang, Yikun Ban

efficiency

One email a morning, machine-written

Related stories

LinearARD: Linear-Memory Attention Distillation for RoPE Restoration

LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents

Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models

AsyncWebRL: Efficient Asynchronous Reinforcement Learning for Multi-Step Visual Web Agents

Looped Latent Attention: Cross-Loop KV Compression for Looped Transformers

WDL-OPD: Weak-Driven On-Policy Distillation via Mixture-Constrained Co-Training