CANTANTE: Optimizing Agentic Systems via Contrastive Credit Attribution

• URL: http://arxiv.org/abs/2605.13295v1
• Date: Wed, 13 May 2026 10:09:10 GMT
• Title: CANTANTE: Optimizing Agentic Systems via Contrastive Credit Attribution
• Abstract summary: We introduce CANTANTE, a framework that decomposes system-level rewards into per-agent update signals.<n>We evaluate it against GEPA and MIPROv2 on programming, mathematical reasoning, and multi-hop question answering.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: LLM-based multi-agent systems have demonstrated strong performance across complex real-world tasks, such as software engineering, predictive modeling, and retrieval-augmented generation. Yet automating their configuration remains a structural challenge, as scores are available only at the system level, whereas the parameters governing agent behavior are local. We argue that optimizing these systems is fundamentally a credit-assignment problem. We therefore introduce CANTANTE, a framework that decomposes system-level rewards into per-agent update signals by contrasting rollouts of multiple joint configurations on the same query. We instantiate it for prompt optimization, treating agent prompts as learnable system parameters. We evaluate CANTANTE against GEPA and MIPROv2 on programming (MBPP), mathematical reasoning (GSM8K), and multi-hop question answering (HotpotQA). Across these benchmarks, CANTANTE achieves the best average rank among all evaluated optimizers and consistently outperforms unoptimized prompts. It improves over the strongest baseline by +18.9 percentage points on MBPP and +12.5 percentage points on GSM8K, while incurring a lower inference cost. It remains within one standard deviation of the strongest baseline on HotpotQA. Crucially, our credit correlation analysis confirms that the attributer produces meaningful per-agent signals rather than echoing the global system score.

Related papers

• PACE: A Proxy for Agentic Capability Evaluation [60.3743414796937]
  We introduce PACE, a framework that constructs proxy benchmarks by selecting instances from existing non-agentic evaluations.<n>Experiments across 14 models, 4 agentic benchmarks, and 19 non-agentic benchmarks show that PACE-Bench predicts agentic scores with leave-one-out cross-validation (LOOCV) mean absolute error (MAE) under 4%, Spearman correlation above 0.80, and pairwise model-ranking accuracy around 85%, all at much less than 1% of the full agentic evaluation cost.
  arXiv  Detail & Related papers  (2026-07-02T10:59:03Z)
• Agent-Orchestrated Adaptive RAG: A Comparative Study on Structured and Multi-Hop Retrieval [0.0]
  Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by grounding their responses in external knowledge.<n>This paper presents an Agent-Orchestrated Adaptive RAG framework that introduces dynamic query decomposition, iterative retrieval, and a bounded self-reflective evaluation loop.
  arXiv  Detail & Related papers  (2026-06-04T03:38:46Z)
• Agent-GWO: Collaborative Agents for Dynamic Prompt Optimization in Large Language Models [69.55139736609367]
  Agent-GWO is a dynamic prompt optimization framework for complex reasoning.<n>We show that Agent-GWO consistently improves accuracy and stability over existing prompt optimization methods.
  arXiv  Detail & Related papers  (2026-04-14T07:35:37Z)
• Agent Q-Mix: Selecting the Right Action for LLM Multi-Agent Systems through Reinforcement Learning [68.85408801740228]
  We propose textbfAgent Q-Mix, a reinforcement learning framework that reformulates topology selection as a cooperative Multi-Agent Reinforcement Learning problem.<n>Our method learns decentralized communication decisions using Qmix value factorization, where each agent selects from a set of communication actions that jointly induce a round-wise communication graph.<n>Agent Q-Mix achieves the highest average accuracy compared to existing methods while demonstrating superior token efficiency and robustness against agent failure.
  arXiv  Detail & Related papers  (2026-04-01T00:38:24Z)
• Auto-Prompting with Retrieval Guidance for Frame Detection in Logistics [0.0]
  Prompt engineering plays a critical role in adapting large language models (LLMs) to complex reasoning and labeling tasks without the need for extensive fine-tuning.<n>We propose a novel prompt optimization pipeline for frame detection in logistics texts, combining retrieval-augmented generation (RAG), few-shot prompting, chain-of-thought (CoT) reasoning, and automatic CoT synthesis (Auto-CoT) to generate highly effective task-specific prompts.<n>Our framework is evaluated on a real-world logistics text annotation task, where reasoning accuracy and labeling efficiency are critical.
  arXiv  Detail & Related papers  (2025-12-22T10:29:51Z)
• Stochastic Self-Organization in Multi-Agent Systems [28.70691568233268]
  Multi-agent systems (MAS) based on Large Language Models (LLMs) have the potential to solve tasks that are beyond the reach of any single LLM.<n>We introduce a response-conditioned framework that adapts communication on-the-fly.
  arXiv  Detail & Related papers  (2025-10-01T09:08:04Z)
• Maestro: Joint Graph & Config Optimization for Reliable AI Agents [53.71882250666667]
  Maestro is a holistic-agnostic framework for LLM agents that jointly searches over graphs and configurations to maximize agent quality.<n>On the IFBench and HotpotQA benchmarks, Maestro consistently surpasses leading prompts--MIPROv2, GEPA, and GEPA+--by an average of 12%--4.9%, and 4.86%, respectively.
  arXiv  Detail & Related papers  (2025-09-04T20:00:37Z)
• Optimizing the Interface Between Knowledge Graphs and LLMs for Complex Reasoning [0.0]
  Large Language Models (LLMs) with Knowledge Graphs (KGs)<n>Cognane is a modular framework for end-to-end KG construction and retrieval.<n>We optimize parameters related to chunking, graph construction, retrieval, and prompting.
  arXiv  Detail & Related papers  (2025-05-30T11:27:59Z)
• FlowReasoner: Reinforcing Query-Level Meta-Agents [63.602173107171076]
  This paper proposes a query-level meta-agent named FlowReasoner to automate the design of query-level multi-agent systems.<n>Our core idea is to incentivize a reasoning-based meta-agent via external execution feedback.
  arXiv  Detail & Related papers  (2025-04-21T17:35:42Z)
• On the Role of Feedback in Test-Time Scaling of Agentic AI Workflows [71.92083784393418]
  Agentic AI (systems that autonomously plan and act) are becoming widespread, yet their task success rate on complex tasks remains low.<n>Inference-time alignment relies on three components: sampling, evaluation, and feedback.<n>We introduce Iterative Agent Decoding (IAD), a procedure that repeatedly inserts feedback extracted from different forms of critiques.
  arXiv  Detail & Related papers  (2025-04-02T17:40:47Z)
• LLaMA-Berry: Pairwise Optimization for O1-like Olympiad-Level Mathematical Reasoning [56.273799410256075]
  The framework combines Monte Carlo Tree Search (MCTS) with iterative Self-Refine to optimize the reasoning path. The framework has been tested on general and advanced benchmarks, showing superior performance in terms of search efficiency and problem-solving capability.
  arXiv  Detail & Related papers  (2024-10-03T18:12:29Z)

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.