Retry, Switch, or Abstain? Learning Strategy-Aware Tool-Use Policies via Controlled Error Injection

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Chaoran Chen, Vy Nguyen, Ziji Zhang, Abhinav Gullapalli, Ziyi Wang, Yuxuan Lu, Dakuo Wang, Jing Huang, Zhou Yu, Jin Lai
• Venue: arXiv: cs.AI
• Published: 2026-08-12
• Collected: 2026-08-13
• URL: http://arxiv.org/abs/2608.11977v1
• PDF: https://arxiv.org/pdf/2608.11977v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

Researchers present BENCH2ROBUST, a framework for converting reliable tool-use benchmarks into stochastic environments with controlled solvability. This allows study of agent's response to tool failures, which can be either transient, persistent, or silent. The research focuses on two interventions: Bayesian Tool Memory (BTM) and curriculum-controlled reinforcement learning. Results show that both interventions improve robustness in various models, with combining them providing the best results.

What the Document Says

• The paper presents a framework named BENCH2ROBUST for converting failure-free tool-use benchmarks into controlled stochastic environments.
• These environments allow episodes requiring retrying, switching, or stopping after all viable paths are exhausted.
• Researchers study two interventions: structured runtime recovery context through Bayesian Tool Memory (BTM), and curriculum-controlled reinforcement learning.
• The experiments were conducted across 7 models from 4 families and two multi-turn benchmark families.
• Tool failures produce a robustness gap in most of the models.
• Bayesian Tool Memory (BTM) improves robustness by up to 16.8 percentage points without retraining.

Key Concepts Mentioned

• tool-using LLM agents
• robust recovery
• controlled error injection
• Bayesian Tool Memory (BTM)
• curriculum-controlled reinforcement learning

Claims Made by the Document

• BENCH2ROBUST converts failure-free tool-use benchmarks into controlled stochastic environments.
• Tool failures produce a robustness gap in most of the models.
• Bayesian Tool Memory (BTM) improves robustness by up to 16.8 percentage points without retraining.

Evidence Used

• The paper describes BENCH2ROBUST framework for this purpose.
• The results show the robustness gap and improvement due to BTM.

Methods / Approach

• BENCH2ROBUST framework
• Bayesian Tool Memory (BTM)
• curriculum-controlled reinforcement learning

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Tool-using LLM agents are commonly trained and evaluated in environments where tool calls succeed reliably, yet deployed tools can fail transiently, persistently, or silently. Robust recovery therefore requires more than repeated retries: an agent may need to retry the same path, switch to an alternative, or recognize that no viable path remains. We present BENCH2ROBUST, a framework that converts failure-free tool-use benchmarks into controlled stochastic environments with scenario-controlled solvability, where episodes explicitly require retrying, switching, or stopping after available paths are exhausted. We use BENCH2ROBUST to study two complementary interventions: structured runtime recovery context through Bayesian Tool Memory (BTM), and curriculum-controlled reinforcement learning. Across 7 models from 4 families and two multi-turn benchmark families, tool failures produce a near-universal robustness gap. On held-out Retail tasks, BTM improves robustness by up to 16.8 percentage points without retraining, while RL learns complementary recovery behavior that remains beneficial without inference-time BTM. Combining the two reaches 40.8-45.5% under injection while preserving failure-free performance. These results suggest that robust tool use benefits from combining environment-specific recovery knowledge with learned recovery behavior.