Aligning Progress and Feasibility: A Neuro-Symbolic Dual Memory Framework for Long-Horizon LLM Agents

• URL: http://arxiv.org/abs/2604.02734v1
• Date: Fri, 03 Apr 2026 05:11:19 GMT
• Title: Aligning Progress and Feasibility: A Neuro-Symbolic Dual Memory Framework for Long-Horizon LLM Agents
• Abstract summary: Large language models (LLMs) have demonstrated strong potential in long-horizon decision-making tasks.<n>We propose a Neuro-Symbolic Dual Memory Framework that explicitly decouples semantic progress guidance from logical feasibility verification.
• Score: 20.162094925591354
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) have demonstrated strong potential in long-horizon decision-making tasks, such as embodied manipulation and web interaction. However, agents frequently struggle with endless trial-and-error loops or deviate from the main objective in complex environments. We attribute these failures to two fundamental errors: global Progress Drift and local Feasibility Violation. Existing methods typically attempt to address both issues simultaneously using a single paradigm. However, these two challenges are fundamentally distinct: the former relies on fuzzy semantic planning, while the latter demands strict logical constraints and state validation. The inherent limitations of such a single-paradigm approach pose a fundamental challenge for existing models in handling long-horizon tasks. Motivated by this insight, we propose a Neuro-Symbolic Dual Memory Framework that explicitly decouples semantic progress guidance from logical feasibility verification. Specifically, during the inference phase, the framework invokes both memory mechanisms synchronously: on one hand, a neural-network-based Progress Memory extracts semantic blueprints from successful trajectories to guide global task advancement; on the other hand, a symbolic-logic-based Feasibility Memory utilizes executable Python verification functions synthesized from failed transitions to perform strict logical validation. Experiments demonstrate that this method significantly outperforms existing competitive baselines on ALFWorld, WebShop, and TextCraft, while drastically reducing the invalid action rate and average trajectory length.

Related papers

• EchoFuzz: Empowering Smart Contract Fuzzing with Large Language Models [12.59838258252938]
  We propose EchoFuzz, an LLM-guided fuzzing framework introducing Vulnerable Call Function Sequences (VFCS)<n>We develop a chain-guided LLM approach, that combines static analysis with logical understanding to generate contract-specific VFCS candidates.<n>Experiments show EchoFuzz outperforms state-of-the-art methods, achieving 29% higher branch coverage and detecting 62% more vulnerabilities.
  arXiv  Detail & Related papers  (2026-09-13T12:31:54Z)
• Dual Process Motion Planning [53.570848178455776]
  We introduce a dual-process architecture that combines the strengths of robust reasoning and learning.<n>Our framework integrates state-of-the-art symbolic solvers as a System-2'' component with experience-driven System-1'' modules.
  arXiv  Detail & Related papers  (2026-09-01T13:52:12Z)
• MIRTH: Mutual-Information Reasoning with Temporal Hubs for Vision-Language-Action Agents [15.061825959360192]
  VLA models have emerged as a powerful paradigm for transferring semantic knowledge from web-scale data to physical robotic control.<n>MIRTH is a unified framework designed to address these challenges.<n>MIRTH augments a pretrained VLA backbone with three key innovations: (1) dual-scale temporal memory hubs that compress long-term scene evolution and short-term motion trends into compact embeddings; (2) latent reasoning tokens optimized via a mutual-information objective carving out a semantic plan space to align multimodal context with action trajectories; and (3) a parallel action decoding scheme that replaces autoregressive generation with vector-wise prediction to maximize control throughput
  arXiv  Detail & Related papers  (2026-06-30T05:57:13Z)
• eMoT: evolving Memory-of-Thought via Symbolic Anchoring and Memory Corrosion [34.722617170864964]
  We propose eMoT (evolving Memory-of-Thought), a unified framework that stabilizes multi-step reasoning.<n>eMoT consists of three interconnected modules: (i) a memory corrosion mechanism that reinforces high-utility reasoning structures; (ii) a symbolic anchoring engine that utilizes Python for deterministic computation, much like a human uses a calculator; and (iii) a consistency-driven refinement process that aligns neural inference with symbolic outcomes.<n>On the traditional task Game of 24, eMoT achieves 100% accuracy, surpassing the baseline by up to 17.6%.
  arXiv  Detail & Related papers  (2026-06-01T10:41:12Z)
• How LLMs Fail and Generalize in RTL Coding for Hardware Design? [56.361436215029045]
  We introduce a new error taxonomy grounded in problem solvability, inspired by cognitive theory.<n>Our taxonomy categorizes failures into syntactic, semantic, solvable functional, and unsolvable functional types.
  arXiv  Detail & Related papers  (2026-04-26T14:34:49Z)
• DexWorldModel: Causal Latent World Modeling towards Automated Learning of Embodied Tasks [54.32016216994156]
  We introduce the Causal Latent World Model (CLWM), which employs DINOv3 features as generative targets to disentangle interaction semantics from visual noise.<n>To overcome memory scaling, CLWM features a Dual-State Test-Time Training (TTT) Memory that guarantees a strict $mathcalO(1)$ footprint for long-horizon tasks.<n>We present EmbodiChain, an online framework that establishes the Efficiency Law by injecting an infinite flow of physics-grounded trajectories during training.
  arXiv  Detail & Related papers  (2026-04-13T03:19:36Z)
• Training LLMs for Multi-Step Tool Orchestration with Constrained Data Synthesis and Graduated Rewards [76.49428173793386]
  LLMs must invoke multiple dependent APIs in the correct order while propagating intermediate outputs.<n>Training models to handle such faces two obstacles: existing environments focus on simple per-turn function calls with simulated data, and binary rewards.<n>We present a framework addressing both challenges. First, we construct a reinforcement learning environment backed by a large-scale cache of real API responses, enabling a data synthesis pipeline that samples valid multi-step orchestration traces.<n>Second, we propose a graduated reward design that decomposes correctness into atomic validity and orchestration.
  arXiv  Detail & Related papers  (2026-03-25T18:31:39Z)
• BiManiBench: A Hierarchical Benchmark for Evaluating Bimanual Coordination of Multimodal Large Language Models [30.38546017617527]
  We introduce BiManiBench, a benchmark evaluating Multi Large Language Models (MLLMs) across three tiers.<n>Our framework isolates unique bimanual challenges, such as arm reachability and kinematic constraints.<n>Analysis of over 30 state-of-the-art models reveals that despite high-level reasoning proficiency, MLLMs struggle with dual-arm spatial grounding and control.
  arXiv  Detail & Related papers  (2026-02-09T08:47:14Z)
• OMG-Agent: Toward Robust Missing Modality Generation with Decoupled Coarse-to-Fine Agentic Workflows [9.617220633655716]
  We present textbfunderlineOmni-textbfunderlineModality textbfunderlineGeneration Agent (textbfOMG-Agent)
  arXiv  Detail & Related papers  (2026-02-04T02:25:40Z)
• StreamVLA: Breaking the Reason-Act Cycle via Completion-State Gating [8.031311182554981]
  Long-horizon robotic manipulation requires bridging the gap between high-level planning and low-level control.<n>Current Vision-Language-Action models often entangle these processes, performing redundant multimodal reasoning at every timestep.<n>We present StreamVLA, a dual-system architecture that unifies textual task decomposition, visual goal imagination, and continuous action generation within a single parameter-efficient backbone.
  arXiv  Detail & Related papers  (2026-02-01T08:51:17Z)
• CoG: Controllable Graph Reasoning via Relational Blueprints and Failure-Aware Refinement over Knowledge Graphs [53.199517625701475]
  CoG is a training-free framework inspired by Dual-Process Theory that mimics the interplay between intuition and deliberation.<n>CoG significantly outperforms state-of-the-art approaches in both accuracy and efficiency.
  arXiv  Detail & Related papers  (2026-01-16T07:27:40Z)
• Resolving Indirect Calls in Binary Code via Cross-Reference Augmented Graph Neural Networks [14.474710100574866]
  We introduce CupidCall, a novel approach for resolving indirect calls using graph neural networks.<n>We leverage advanced compiler-level type analysis to generate high-quality callsite-callee training pairs.<n>We further design a graph neural model that leverages augmented CFGs and relational graph convolutions for accurate target prediction.
  arXiv  Detail & Related papers  (2025-07-24T20:54:41Z)
• Disentangling Memory and Reasoning Ability in Large Language Models [97.26827060106581]
  We propose a new inference paradigm that decomposes the complex inference process into two distinct and clear actions.<n>Our experiment results show that this decomposition improves model performance and enhances the interpretability of the inference process.
  arXiv  Detail & Related papers  (2024-11-20T17:55:38Z)
• SHERL: Synthesizing High Accuracy and Efficient Memory for Resource-Limited Transfer Learning [63.93193829913252]
  We propose an innovative METL strategy called SHERL for resource-limited scenarios. In the early route, intermediate outputs are consolidated via an anti-redundancy operation. In the late route, utilizing minimal late pre-trained layers could alleviate the peak demand on memory overhead.
  arXiv  Detail & Related papers  (2024-07-10T10:22:35Z)

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.