Textual Equilibrium Propagation for Deep Compound AI Systems

• URL: http://arxiv.org/abs/2601.21064v2
• Date: Mon, 02 Feb 2026 21:45:50 GMT
• Title: Textual Equilibrium Propagation for Deep Compound AI Systems
• Abstract summary: Large language models (LLMs) are increasingly deployed as part of compound AI systems.<n>Recent approaches that propagate textual feedback globally (e.g., TextGrad) make it feasible to optimize such pipelines, but we find that performance degrades as system depth grows.<n>In particular, long-horizon agentic exhibit two depth-scaling failure modes: 1) exploding textual gradient, where textual feedback grows exponentially with depth, leading to prohibitively long message and amplifies evaluation biases; and 2) vanishing textual gradient, where limited long-context ability causes models overemphasize partial feedback and compression of lengthy feedback causes downstream messages to lose specificity gradually
• Score: 63.266386654809544
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) are increasingly deployed as part of compound AI systems that coordinate multiple modules (e.g., retrievers, tools, verifiers) over long-horizon workflows. Recent approaches that propagate textual feedback globally (e.g., TextGrad) make it feasible to optimize such pipelines, but we find that performance degrades as system depth grows. In particular, long-horizon agentic workflows exhibit two depth-scaling failure modes: 1) exploding textual gradient, where textual feedback grows exponentially with depth, leading to prohibitively long message and amplifies evaluation biases; and 2) vanishing textual gradient, where limited long-context ability causes models overemphasize partial feedback and compression of lengthy feedback causes downstream messages to lose specificity gradually as they propagate many hops upstream. To mitigate these issues, we introduce Textual Equilibrium Propagation (TEP), a local learning principle inspired by Equilibrium Propagation in energy-based models. TEP includes two phases: 1) a free phase where a local LLM critics iteratively refine prompts until reaching equilibrium (no further improvements are suggested); and 2) a nudged phase which applies proximal prompt edits with bounded modification intensity, using task-level objectives that propagate via forward signaling rather than backward feedback chains. This design supports local prompt optimization followed by controlled adaptation toward global goals without the computational burden and signal degradation of global textual backpropagation. Across long-horizon QA benchmarks and multi-agent tool-use dataset, TEP consistently improves accuracy and efficiency over global propagation methods such as TextGrad. The gains grows with depth, while preserving the practicality of black-box LLM components in deep compound AI system.

Related papers

• Do Language Models Converge to Themselves? Recursive Self-Refinement as Textual Relaxation [0.6509233267425589]
  We generate 10-step refinement trajectories for 50 ICML 2025 abstracts under both default-temperature and deterministic decoding.<n>We analyze normalized edit distance, exact and approximate fixed points, word-count stability, exponential relaxation, and external LLM-as-a-judge evaluation.
  arXiv  Detail & Related papers  (2026-06-26T13:44:53Z)
• Group-Graph Policy Optimization for Long-Horizon Agentic Reinforcement Learning [85.02566758103008]
  Group-Graph Policy Optimization (G2PO) is a novel group-based Reinforcement Learning algorithm tailored for multi-turn agentic tasks.<n>G2PO substantially outperforms state-of-the-art prompt-based and RL baselines, achieving remarkable success rate improvements of up to 22.2% over GRPO.
  arXiv  Detail & Related papers  (2026-06-22T08:12:47Z)
• Taming the Thinker: Conditional Entropy Shaping for Adaptive LLM Reasoning [63.36243250685778]
  Conditional Entropy Shaping (CES) is a framework that dynamically controls token-level response entropy.<n>CES consistently improves average accuracy while reducing response length relative to DAPO.
  arXiv  Detail & Related papers  (2026-05-19T04:41:51Z)
• PRISM: Prior Rectification and Uncertainty-Aware Structure Modeling for Diffusion-Based Text Image Super-Resolution [51.96078493242164]
  PRISM is a single-step diffusion-based Text-SR framework.<n> PRISM achieves state-of-the-art performance with millisecond-level inference.
  arXiv  Detail & Related papers  (2026-05-13T05:31:06Z)
• Optimizing Multi-Agent Weather Captioning via Text Gradient Descent: A Training-Free Approach with Consensus-Aware Gradient Fusion [0.0]
  We introduce WeatherTGD, a training-free multi-agent framework that reinterprets collaborative caption refinement through the lens of Text Gradient Descent (TGD)<n>Our system deploys three specialized LLM agents including a Statistical Analyst, a Physics Interpreter, and a Meteorology Expert that generate domain-specific textual gradients from weather time series observations.<n> Experiments on real-world meteorological datasets demonstrate that WeatherTGD achieves significant improvements in both LLM-based evaluation and human expert evaluation.
  arXiv  Detail & Related papers  (2026-03-23T07:55:45Z)
• TextResNet: Decoupling and Routing Optimization Signals in Compound AI Systems via Deep Residual Tuning [48.34766553598802]
  TextResNet is a framework that reformulates the optimization process to achieve precise signal routing via four key innovations.<n>Our results show that TextResNet achieves superior performance compared to Text, but also exhibits remarkable stability for agentic tasks in compound AI systems.
  arXiv  Detail & Related papers  (2026-02-09T06:29:32Z)
• RAPO++: Cross-Stage Prompt Optimization for Text-to-Video Generation via Data Alignment and Test-Time Scaling [59.088798018184235]
  textbfRAPO++ is a cross-stage prompt optimization framework.<n>It unifies training-data-aligned refinement, test-time iterative scaling, and large language model fine-tuning.<n> RAPO++ achieves significant gains in semantic alignment, compositional reasoning, temporal stability, and physical plausibility.
  arXiv  Detail & Related papers  (2025-10-23T04:45:09Z)
• LUMA: Low-Dimension Unified Motion Alignment with Dual-Path Anchoring for Text-to-Motion Diffusion Model [18.564067196226436]
  We propose a text-to-motion diffusion model that incorporates dual-path anchoring to enhance semantic alignment.<n>LUMA achieves state-of-the-art performance, with FID scores of 0.035 and 0.123, respectively.
  arXiv  Detail & Related papers  (2025-09-29T17:58:28Z)
• TopoSizing: An LLM-aided Framework of Topology-based Understanding and Sizing for AMS Circuits [7.615431299673158]
  Traditional black-box optimization achieves sampling efficiency but lacks circuit understanding.<n>We propose TopoSizing, an end-to-end framework that performs robust circuit understanding directly from raw netlists.
  arXiv  Detail & Related papers  (2025-09-17T16:52:46Z)
• Rethinking Prompt Optimization: Reinforcement, Diversification, and Migration in Blackbox LLMs [10.434732630519377]
  We propose a novel Automatic Prompt Optimization (APO) framework centered on enhancing the feedback mechanism.<n>To mitigate the noise inherent in LLM-generated feedback, we introduce a technique called feedback diversification.<n>Our approach consistently outperforms strong baselines, achieving significant accuracy improvements, faster convergence, and lower computational costs.
  arXiv  Detail & Related papers  (2025-07-14T00:20:14Z)
• Multimodal LLM-Guided Semantic Correction in Text-to-Image Diffusion [52.315729095824906]
  MLLM Semantic-Corrected Ping-Pong-Ahead Diffusion (PPAD) is a novel framework that introduces a Multimodal Large Language Model (MLLM) as a semantic observer during inference.<n>It performs real-time analysis on intermediate generations, identifies latent semantic inconsistencies, and translates feedback into controllable signals that actively guide the remaining denoising steps.<n>Extensive experiments demonstrate PPAD's significant improvements.
  arXiv  Detail & Related papers  (2025-05-26T14:42:35Z)
• Exploiting Edited Large Language Models as General Scientific Optimizers [10.489095729605152]
  Large language models (LLMs) have been widely adopted in mathematical optimization in scientific scenarios.<n>We propose a conceptually simple and general bi-level optimization method, namely textbfGeneral textbfScientific textbfOptimizers (GSO)
  arXiv  Detail & Related papers  (2025-03-08T18:01:11Z)
• Learning Inverse Laplacian Pyramid for Progressive Depth Completion [18.977393635158048]
  LP-Net is an innovative framework that implements a multi-scale, progressive prediction paradigm based on Laplacian Pyramid decomposition.<n>At the time of submission, LP-Net ranks 1st among all peer-reviewed methods on the official KITTI leaderboard.
  arXiv  Detail & Related papers  (2025-02-11T06:21:42Z)
• Low-Light Image Enhancement via Generative Perceptual Priors [75.01646333310073]
  We introduce a novel textbfLLIE framework with the guidance of vision-language models (VLMs)<n>We first propose a pipeline that guides VLMs to assess multiple visual attributes of the LL image and quantify the assessment to output the global and local perceptual priors.<n>To incorporate these generative perceptual priors to benefit LLIE, we introduce a transformer-based backbone in the diffusion process, and develop a new layer normalization (textittextbfLPP-Attn) guided by global and local perceptual priors.
  arXiv  Detail & Related papers  (2024-12-30T12:51:52Z)
• In-context Demonstration Matters: On Prompt Optimization for Pseudo-Supervision Refinement [71.60563181678323]
  Large language models (LLMs) have achieved great success across diverse tasks, and fine-tuning is sometimes needed to further enhance generation quality.<n>To handle these challenges, a direct solution is to generate high-confidence'' data from unsupervised downstream tasks.<n>We propose a novel approach, pseudo-supervised demonstrations aligned prompt optimization (PAPO) algorithm, which jointly refines both the prompt and the overall pseudo-supervision.
  arXiv  Detail & Related papers  (2024-10-04T03:39:28Z)
• Entropy-Regularized Token-Level Policy Optimization for Language Agent Reinforcement [67.1393112206885]
  Large Language Models (LLMs) have shown promise as intelligent agents in interactive decision-making tasks. We introduce Entropy-Regularized Token-level Policy Optimization (ETPO), an entropy-augmented RL method tailored for optimizing LLMs at the token level. We assess the effectiveness of ETPO within a simulated environment that models data science code generation as a series of multi-step interactive tasks.
  arXiv  Detail & Related papers  (2024-02-09T07:45:26Z)

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.