PIVOT: Bridging Planning and Execution in LLM Agents via Trajectory Refinement

• URL: http://arxiv.org/abs/2605.11225v1
• Date: Mon, 11 May 2026 20:43:20 GMT
• Title: PIVOT: Bridging Planning and Execution in LLM Agents via Trajectory Refinement
• Abstract summary: Large language model (LLM)-based agents generate seemingly coherent plans that fail upon execution due to infeasible actions, constraint violations, and compounding errors over extended horizons.<n>PIVOT addresses this plan-execution misalignment through a self-supervised framework that treats trajectories as optimizable objects iteratively refined via environment interaction.
• Score: 22.285969725581044
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language model (LLM)-based agents frequently generate seemingly coherent plans that fail upon execution due to infeasible actions, constraint violations, and compounding errors over extended horizons. PIVOT (Plan-Inspect-eVOlve Trajectories) addresses this plan-execution misalignment through a self-supervised framework that treats trajectories as optimizable objects iteratively refined via environment interaction. The framework comprises four stages: PLAN generates candidate trajectories; INSPECT executes them and computes structured losses with textual gradients encoding plan-execution discrepancies; EVOLVE applies these signals to produce improved trajectories; and VERIFY performs a final global check against task constraints. A monotonic acceptance process ensures a non-decreasing solution quality. Empirical evaluations on DeepPlanning and GAIA demonstrate state-of-the-art performance: with human-in-the-loop (HITL) feedback, PIVOT establishes a strong upper bound up to 94% relative improvement in constraint satisfaction, while its fully autonomous variant retains substantial gains, showing that the core trajectory-refinement mechanism remains effective without external supervision. At the same time, PIVOT remains computationally efficient, requiring up to 3x to 5x fewer tokens than competing refinement methods. These findings establish that (self- or human-supervised) feedback-based trajectory optimization is a principled methodology for mitigating plan-execution gaps in autonomous agent systems.

Related papers

• Grasp-Then-Plan with Failure Attribution: A Closed Two-Stage Framework for Precise and Generalizable Robotic Manipulation [12.60091812522059]
  In robotic manipulation, the tight coupling between grasping and motion planning often obscures the true source of failure.<n>We propose GTP-FA, a task-oriented two-stage grasp-then-plan framework that generates grasp candidates and performs downstream motion planning conditioned on the selected grasp.<n>We show that GTP-FA improves the corresponding base learners across RL, IL, diffusion-policy, and VLA-based settings, achieving substantially higher overall task success rates.
  arXiv  Detail & Related papers  (2026-06-02T09:29:03Z)
• IterInject: Indirect Prompt Injection Against LLM Agents via Feedback-Guided Iterative Optimization [33.58630045185762]
  Existing attacks rely on static payloads that cannot adapt to agent-specific defenses.<n>We introduce oursys, a feedback-guided iterative framework that closes the loop between injection, diagnosis, and diagnosis.<n>A step generates new disguise seeds from failure patterns, enabling the strategy space to self-evolve.
  arXiv  Detail & Related papers  (2026-05-23T17:00:06Z)
• AvAtar: Learning to Align via Active Optimal Transport [70.846866626571]
  We propose a principled active alignment framework for optimal transport alignment called AvAtar.<n>We quantify the informativeness of a candidate by measuring its gradient-based impact on the global alignment result.<n>By encoding the global alignment result via effective utility functions, AvAtar is applicable to general alignment problems under the OT framework.
  arXiv  Detail & Related papers  (2026-05-23T04:40:21Z)
• TIER: Trajectory-Invariant Execution Rewards for Multi-Step Tool Composition [62.56752617853322]
  Outcome-based rewards provide only sparse feedback, while trajectory-supervised rewards depend on annotated reference solutions.<n>We propose TIER: Trajectory-Invariant Execution Rewards, a reward framework that derives supervision directly from function schemas and runtime execution.
  arXiv  Detail & Related papers  (2026-05-16T03:47:26Z)
• AnchorRefine: Synergy-Manipulation Based on Trajectory Anchor and Residual Refinement for Vision-Language-Action Models [60.04879435087352]
  Vision-language-action (VLA) policies generate actions within a single unified space.<n>We propose AnchorRefine, a hierarchical framework that factorizes VLA action modeling into trajectory anchor and residual refinement.<n> Experiments on LIBERO, CALVIN, and real-robot tasks demonstrate that AnchorRefine consistently improves both regression-based and diffusion-based VLA backbones.
  arXiv  Detail & Related papers  (2026-04-20T04:25:24Z)
• HAD: Combining Hierarchical Diffusion with Metric-Decoupled RL for End-to-End Driving [51.268878540511054]
  We propose HAD, an end-to-end planning framework with a Hierarchical Diffusion Policy.<n>We show that HAD achieves new state-of-the-art performance on both NAVSIM and HUGSIM.
  arXiv  Detail & Related papers  (2026-04-04T04:12:47Z)
• dTRPO: Trajectory Reduction in Policy Optimization of Diffusion Large Language Models [71.26391195427878]
  Diffusion Large Language Models (dLLMs) introduce a new paradigm for language generation.<n>In this work, we aim to improve the policy optimization for dLLMs by reducing the cost of the trajectory probability calculation.<n>We evaluate dTRPO on 7B dLLMs across instruction-following and reasoning benchmarks.
  arXiv  Detail & Related papers  (2026-03-19T11:55:52Z)
• Anticipatory Planning for Multimodal AI Agents [77.62643381558613]
  We introduce TraceR1, a two-stage reinforcement learning framework that explicitly trains anticipatory reasoning.<n>TraceR1 is evaluated across seven benchmarks, covering online computer-use, offline computer-use benchmarks, and multimodal tool-use reasoning tasks.<n>Results show that anticipatory trajectory reasoning is a key principle for building multimodal agents that can reason, plan, and act effectively in complex real-world environments.
  arXiv  Detail & Related papers  (2026-03-17T16:55:11Z)
• Let's Reward Step-by-Step: Step-Aware Contrastive Alignment for Vision-Language Navigation in Continuous Environments [38.97818584066075]
  Vision-Language Navigation in Continuous Environments (VLN-CE) requires agents to learn complex reasoning from long-horizon human interactions.<n>Current training paradigms struggle to balance generalization capability, error recovery and training stability.<n>We introduce Step-Aware Contrastive Alignment (SACA), a framework designed to extract dense supervision from imperfect trajectories.
  arXiv  Detail & Related papers  (2026-03-10T14:45:50Z)
• Act, Think or Abstain: Complexity-Aware Adaptive Inference for Vision-Language-Action Models [7.802379200026965]
  We propose an adaptive framework that dynamically routes VLA execution based on the complexity of the perceived state.<n>Our approach transforms the VLA's vision-language backbone into an active detection tool by projecting latent embeddings into an ensemble of parametric and non-parametric estimators.
  arXiv  Detail & Related papers  (2026-03-05T13:14:41Z)
• Self-Correcting VLA: Online Action Refinement via Sparse World Imagination [55.982504915794514]
  We propose Self-Correcting VLA (SC-VLA), which achieve self-improvement by intrinsically guiding action refinement through sparse imagination.<n>SC-VLA achieve state-of-the-art performance, yielding the highest task throughput with 16% fewer steps and a 9% higher success rate than the best-performing baselines.
  arXiv  Detail & Related papers  (2026-02-25T06:58:06Z)
• Reshaping Action Error Distributions for Reliable Vision-Language-Action Models [69.38615670891038]
  In robotic manipulation, vision-language-action (VLA) models have emerged as a promising paradigm for learning generalizable and scalable robot policies.<n>We focus on continuous-action VLA models and move beyond conventional MSE-based regression by reshaping action error distributions during training.<n>We evaluate our approaches across standard, few-shot, and noisy settings on multiple representative VLA architectures.
  arXiv  Detail & Related papers  (2026-02-04T05:37:09Z)
• Plan, Verify and Fill: A Structured Parallel Decoding Approach for Diffusion Language Models [26.195809569100774]
  Plan-Verify-Fill (PVF) is a training-free paradigm that grounds planning via quantitative validation.<n>PVF actively constructs a hierarchical skeleton by prioritizing high-leverage semantic anchors.<n>It reduces the Number of Evaluation Functions (NFE) by up to 65% compared to confidence-based parallel decoding.
  arXiv  Detail & Related papers  (2026-01-18T03:53:01Z)
• Plan Verification for LLM-Based Embodied Task Completion Agents [10.439882851477162]
  Large language model (LLM) based task plans and corresponding human demonstrations for embodied AI may be noisy.<n>We propose an iterative verification framework in which a Judge LLM critiques action sequences and a Planner LLM applies the revisions.
  arXiv  Detail & Related papers  (2025-09-02T19:06:56Z)
• Self-Corrective Task Planning by Inverse Prompting with Large Language Models [9.283971287618261]
  We introduce InversePrompt, a novel self-corrective task planning approach.<n>Our method incorporates reasoning steps to provide clear, interpretable feedback.<n>Results on benchmark datasets show an average 16.3% higher success rate over existing LLM-based task planning methods.
  arXiv  Detail & Related papers  (2025-03-10T13:35:51Z)

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.