ISE: An Execution-Grounded Recipe for Multi-Turn OS-Agent Trajectories

• URL: http://arxiv.org/abs/2606.11520v2
• Date: Sat, 13 Jun 2026 09:55:32 GMT
• Title: ISE: An Execution-Grounded Recipe for Multi-Turn OS-Agent Trajectories
• Abstract summary: Training capable OS agents requires data that simultaneously captures structured user intents, multi-turn task delegation, and grounded tool execution--properties absent from existing datasets.<n>We propose ISE (Intent -> Simulate -> Execute), a three-stage synthesis paradigm that addresses these gaps jointly.
• Score: 18.739245125968175
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Training capable OS agents requires data that simultaneously captures structured user intents, multi-turn task delegation, and grounded tool execution--properties absent from existing datasets. We propose ISE (Intent -> Simulate -> Execute), a three-stage synthesis paradigm that addresses these gaps jointly. Stage 1 constructs roughly 50000 structured intents via a 4D framework (Persona x Domain x Task x Complexity); after deduplication the pool contains 43956 unique intents and attains a Vendi Score of 61.57 over the entire pool on mpnet-base-v2 embeddings (cosine kernel, q=1). Stage 2 drives multi-turn user-agent interaction through a role-locked user simulator that grounds each user turn in actual execution outcomes, producing 23132 complete trajectories averaging 8.12 user turns and 68.24 total dialogue turns. Stage 3 runs every tool call inside a live, isolated OS workspace, generating authentic failure-recovery dynamics instead of simulated responses. Fine-tuning on ISETrace improves ClawEval pass@1 from 19.3 to 37.7 using Qwen3-8B on agent tool-use tasks with a standard protocol. This result outperforms zero-shot GPT-4o and the larger Qwen3-32B base model which is four times bigger. An ablation on Stage 2 proves multi-turn simulation brings a large portion of the performance gain. We release all source code and dataset at https://github.com/Valiere01/ISE-Trace.

Related papers

• PersonaForge: Realistic Multi-Turn User Simulation for Agentic Systems [53.65117408512214]
  We introduce textbfPersonaForge, a user simulation framework for synthesizing realistic multi-turn user--agent interactions.<n>Using PersonaForge, we construct a 6.3K-record training dataset and textbfPersonaForge-Bench, a manually annotated 138-task benchmark.<n>Experiments show that PersonaForge training improves the composite score by +4.1%, with gains across all four dimensions.
  arXiv  Detail & Related papers  (2026-08-28T14:33:19Z)
• UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations [68.29650478987561]
  UI-Mate is a foundation GUI agent that integrates an environment-grounded training stack with in-context demonstration learning.<n>It sets a new open-weight state of the art on general computer-use benchmarks, scoring 77.0% on OSWorld and 66.2% on WindowsAgentArena.
  arXiv  Detail & Related papers  (2026-08-16T20:59:59Z)
• ProCUA-SFT Technical Report [80.97543110323542]
  Training computer-use agents (CUAs) requires large-scale, diverse trajectory data collected in full desktop environments.<n>We present ProCUA-SFT, a dataset of 3.1M step-level SFT samples distilled from 93K synthetic trajectories.<n>Fine-tuning UI-TARS 7B on ProCUA-SFT for one epoch yields 45.0% on OSWorld -- an 18.7 percentage-point improvement over the base model.
  arXiv  Detail & Related papers  (2026-06-15T22:04:11Z)
• EEVEE: Towards Test-time Prompt Learning in the Real World for Self-Improving Agents [64.96332056338923]
  EEVEE is the first multi-dataset test-time prompt learning framework for LLM agents.<n>It enables test-time prompt learning under real-world task streams.<n>EEVEE improves average multi-benchmark scores by 10.38 and 24.32 points over Qwen3-4B-Instruct and DeepSeek-V3.2.
  arXiv  Detail & Related papers  (2026-06-09T17:57:16Z)
• Synthesize and Reward -- Reinforcement Learning for Multi-Step Tool Use in Live Environments [12.645050883623982]
  We present PROVE (Programmatic Rewards On Verified Environments), a framework with three contributions.<n>A library of 20 stateful MCP servers exposes 343 tools, enabling live-execution RL training with session-scoped state isolation.<n>A state-machine data synthesis pipeline generates multi-turn tool-call trajectories grounded in live-sampled server state, so generated queries reference entities that actually exist.<n>On BFCL Multi-Turn, tau2-bench, and T-Eval, PROVE yields improvements of up to +10.2, +6.8, and +6.5 points respectively.
  arXiv  Detail & Related papers  (2026-06-02T16:52:31Z)
• MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research [39.68394140712043]
  MobileGym is a lightweight, fully controllable environment for everyday mobile use.<n>A layered state model and a declarative task-definition framework keep state programmability and task creation practical at scale.<n>A single programmatic judging mechanism delivers both deterministic evaluation verdicts and dense rewards.
  arXiv  Detail & Related papers  (2026-05-25T17:59:49Z)
• EngiAI: A Multi-Agent Framework and Benchmark Suite for LLM-Driven Engineering Design [0.4499833362998488]
  We introduce a benchmark suite with three evaluation dimensions.<n>We present EngiAI, a Multi-Agent System (MAS) implementation built on LangGraph.
  arXiv  Detail & Related papers  (2026-05-19T12:12:09Z)
• Three Roles, One Model: Role Orchestration at Inference Time to Close the Performance Gap Between Small and Large Agents [0.4666493857924357]
  We study whether inference-time scaffolding alone, without any additional training compute, can improve the performance of a small model in complex multi-step environments.<n>We evaluate Qwen3-8B on the AppWorld benchmark under both full-precision and 4-bit quantized configurations.<n>On full-precision inference, our scaffolded 8B model surpasses DeepSeek-Coder 33B Instruct (7.1%) from the original AppWorld evaluation.
  arXiv  Detail & Related papers  (2026-04-13T13:40:33Z)
• APEX-EM: Non-Parametric Online Learning for Autonomous Agents via Structured Procedural-Episodic Experience Replay [7.370176470430802]
  LLM-based autonomous agents lack persistent procedural memory.<n>We present APEX-EM, a non-parametric online learning framework that accumulates, retrieves, and reuses structured procedural plans.
  arXiv  Detail & Related papers  (2026-03-31T00:24:56Z)
• One Battle After Another: Probing LLMs' Limits on Multi-Turn Instruction Following with a Benchmark Evolving Framework [51.50565654314582]
  Large language models can follow users' instructions throughout a dialogue spanning multiple topics.<n>Existing benchmarks are often limited to a fixed number of turns, making them susceptible to saturation and failing to account for the user's interactive experience.<n>We propose a framework for assessing multi-turn instruction-following ability.
  arXiv  Detail & Related papers  (2025-11-05T14:39:59Z)
• NaviAgent: Bilevel Planning on Tool Navigation Graph for Large-Scale Orchestration [13.925896302382043]
  Large language models (LLMs) have recently demonstrated the ability to act as function call agents by invoking external tools.<n>We propose NaviAgent, which decouples task planning from tool execution through graph-based modeling of the tool ecosystem.<n> Experiments show that NaviAgent achieves the best task success rates across models and tasks, and integrating TWMN further boosts performance by up to 17 points on complex tasks.
  arXiv  Detail & Related papers  (2025-06-24T10:39:07Z)
• RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation [52.2244588424002]
  We present RoboTwin 2.0, a scalable framework for automated, large-scale generation of diverse and realistic data.<n>At its core is RoboTwin-OD, an object library of 731 instances across 147 categories with semantic and manipulation-relevant annotations.<n>To improve sim-to-real transfer, RoboTwin 2.0 applies structured domain randomization along five axes.
  arXiv  Detail & Related papers  (2025-06-22T16:26:53Z)
• APIGen-MT: Agentic Pipeline for Multi-Turn Data Generation via Simulated Agent-Human Interplay [86.01901238059261]
  APIGen-MT is a framework that generates verifiable and diverse multi-turn agent data.<n>We train a family of models -- the xLAM-2-fc-r series with sizes ranging from 1B to 70B parameters.<n>Our models outperform frontier models such as GPT-4o and Claude 3.5 on $tau$-bench and BFCL benchmarks.
  arXiv  Detail & Related papers  (2025-04-04T17:13:57Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.