agent-training

Tag

Cards List
#agent-training

@huggingface: Training Agents 4: From reward functions to environments.

X AI KOLs Timeline ↗ · 2026-09-10 Cached

Hugging Face shares a replay of a broadcast discussing advanced techniques for training AI agents, emphasizing the shift from reward functions to environment-based approaches.

0 favorites 0 likes
#agent-training

AgentBrew: Offline Tool-Use Agent Learning from Raw Real-World Trajectories

arXiv cs.AI ↗ · 2026-09-10 Cached

AgentBrew introduces an offline training framework for tool-use agents that learns from raw interaction trajectories without task verifiers, using retrospective task inference and PMI-based credit assignment to improve performance on real-world applications like GitHub and Notion.

0 favorites 0 likes
#agent-training

@juliafedorin: She left Waterloo without graduating. Now she's a Member of Technical Staff at @OpenAI. @BrookeaJoseph joined a 12-pers…

X AI KOLs Following ↗ · 2026-09-08 Cached

An interview with Brookea Joseph, a Member of Technical Staff at OpenAI, detailing her career from a startup intern to conducting computer use research for AI agents, which she views as crucial for advancing toward AGI.

0 favorites 0 likes
#agent-training

Introducing Solaris (17 minute read)

TLDR AI ↗ · 2026-09-01 Cached

Runway introduces Solaris, the first AI model in their Interface World Models family, which generates real-time interactive interfaces by synthesizing frames directly, eliminating the need for code and enabling dynamic design and agent training.

0 favorites 0 likes
#agent-training

Agent Lightning v1.0

Hacker News Top ↗ · 2026-08-24 Cached

Agent Lightning v1.0 is a lightweight agentic reinforcement learning framework by Microsoft, refactored for training AI agents with real harnesses and achieving substantial benchmark improvements, such as a 14.6 percentage point gain on SWE-bench.

0 favorites 0 likes
#agent-training

Write, Execute, Refine: From Skill Followers to Skill Optimizers via Reinforcement Learning from Execution Feedback

arXiv cs.CL ↗ · 2026-08-19 Cached

The paper introduces WER, a multi-phase framework that trains a Skill Optimizer using reinforcement learning from execution feedback to improve tool-using agents, achieving significant performance gains on benchmarks like BFCL v4 and τ2-bench.

0 favorites 0 likes
#agent-training

@DeFiMinty: Can AI systems keep generating harder training tasks for AI agents? New research from Tencent suggests they can. Recurs…

X AI KOLs Following ↗ · 2026-08-08 Cached

New research from Tencent introduces Recursive Synthetic Terminal Tasks (RST), a method that progressively generates harder, verifiable training tasks for AI agents. Starting from 639 tasks, it produced 37,484 verified tasks across 15 rounds, and reinforcement learning with these tasks improved Qwen3.5-27B from 22.7% to 32.0% on Terminal-Bench Hard.

0 favorites 0 likes
#agent-training

When Privileged Guidance Misaligns: State-Matched Routing and Contextualized Self-Distillation for Multi-Turn Agents

arXiv cs.AI ↗ · 2026-08-07 Cached

This paper introduces State-Matched Routing and Contextualized Self-Distillation (SMRC-SD), a method that selectively applies privileged trajectory guidance only when the student agent's current state matches the reference, improving multi-turn agent performance on ALFWorld and WebShop benchmarks.

0 favorites 0 likes
#agent-training

Beyond Simply Environment Scaling: Designing Effective Environment Distributions for Multimodal Agent Learning

Hugging Face Daily Papers ↗ · 2026-08-06 Cached

The paper argues that simply scaling multimodal environments does not always improve agent training, and proposes Ability-aware Environment Selection (AES) and Hierarchical Difficulty Curriculum (HDC) to better structure environment distributions along diversity and difficulty dimensions.

0 favorites 0 likes
#agent-training

CalibForge: Adversarial Solver Calibration for Scaling Learnable Terminal Tasks

Hugging Face Daily Papers ↗ · 2026-08-06 Cached

CalibForge is an autonomous terminal-task synthesis system that uses adversarial solver calibration to create learnable tasks for training terminal agents. It constructs 5,431 calibrated tasks and improves agent performance on Terminal-Bench2.0, SWE-bench Pro, and Doc2Repo.

0 favorites 0 likes
#agent-training

Deep Research Pretraining via Predictive Navigation

arXiv cs.CL ↗ · 2026-08-04 Cached

Introduces Deep Research Pretraining (DRP), an offline framework that generates search-open-write trajectories from citation and hyperlink evidence structures. Qwen3-14B models pretrained on 1B tokens with DRP outperform matched no-DRP baselines on deep research benchmarks, even with less supervised fine-tuning data.

0 favorites 0 likes
#agent-training

SKILL-KD: Contrastive Skill Distillation for LLM Agents

Hugging Face Daily Papers ↗ · 2026-08-04 Cached

SKILL-KD is a contrastive skill distillation framework that improves LLM agents by distilling actionable discrepancies between teacher and student trajectories into textual skill patches, with drift-aware consolidation to iteratively refine skills.

0 favorites 0 likes
#agent-training

@MSFTResearch: Computer-use AI agents struggle with multi-step workflows like email and customer support. Echoverse trains agents in r…

X AI KOLs Timeline ↗ · 2026-07-30 Cached

Microsoft Research introduces Echoverse, a set of deep, evolving environments for training computer-use agents. A 9B model trained on these environments nearly doubles its baseline score, coming within 14 points of GPT-5.4, demonstrating that high-fidelity simulation and co-evolution of model, world, and verifier significantly improve agent performance on multi-step workflows.

0 favorites 0 likes
#agent-training

From Outcomes to Actions: Leveraging Hindsight for Long-Horizon Language Agent Training

arXiv cs.LG ↗ · 2026-07-21 Cached

Introduces Hindsight Policy Optimization (HPO), a novel policy gradient method that uses an intent space and Wasserstein distance to reduce variance in long-horizon language agent training, showing improved stability over GRPO and PPO.

0 favorites 0 likes
#agent-training

NexForge: Scaling Agent Capabilities through Requirement-Driven Task Synthesis for LLMs

Hugging Face Daily Papers ↗ · 2026-07-21 Cached

NexForge is a requirement-driven framework that synthesizes diverse, executable agent tasks and expert trajectories for LLM post-training, outperforming prior methods and achieving state-of-the-art open-source agent performance on Terminal-Bench.

0 favorites 0 likes
#agent-training

Show HN: I RL-trained an agent that trains models with RL (for –$1.3k)

Hacker News Top ↗ · 2026-07-14 Cached

A developer built a pipeline where an RL-trained AI agent creates and submits RL training jobs for small models, rewarding the agent for better performance. The project is fully open-sourced and demonstrates transfer to held-out tasks.

0 favorites 0 likes
#agent-training

TurnOPD: Making On-Policy Distillation Turn-Aware for Efficient Long-Horizon Agent Training

arXiv cs.AI ↗ · 2026-07-08 Cached

TurnOPD introduces turn-level budgeting for on-policy distillation of long-horizon agents, addressing inefficiencies in vanilla OPD by adaptive rollout-depth and progressive turn-normalized loss budgeting, achieving better accuracy under equal training budgets.

0 favorites 0 likes
#agent-training

@qingke_ai: https://x.com/qingke_ai/status/2071281892964659384

X AI KOLs Timeline ↗ · 2026-06-28 Cached

该文章由ROLL团队分享了在终端环境中进行Agentic RL训练时的实践经验,包括环境管理器设计、异步训练管线以及多种模式切换,并对比了RLVR与Agentic RL的本质区别。

0 favorites 0 likes
#agent-training

The Verification Horizon: No Silver Bullet for Coding Agent Rewards

arXiv cs.AI ↗ · 2026-06-26 Cached

该论文指出,对于当前的编码智能体,验证解决方案比生成解决方案更为困难,且任何固定的奖励函数都无法随着能力增长而持续有效。作者通过四种奖励构建的实验表明,针对性的验证设计可以抑制奖励黑客行为并提升任务完成质量。

0 favorites 0 likes
#agent-training

@gabepereyra: Harvey partnered with @appliedcompute to train a legal agent. We optimized each part of the agent stack, including the …

X AI KOLs Following ↗ · 2026-06-22 Cached

Harvey partnered with Applied Compute to train a legal agent, optimizing the agent stack and post-training the GLM-5.1 model using reward signals from their Legal Agent Benchmark.

0 favorites 0 likes
← Previous
Next →
← Back to home

Submit Feedback