Inference-Time Scaling of Verification: Self-Evolving Deep Research Agents via Test-Time Rubric-Guided Verification (2026.findings-acl)
Copied to clipboard
Yuxuan Wan, Tianqing Fang, Zaitang LI, Yintong Huo, Wenxuan Wang, Haitao Mi, Dong Yu, Michael R. Lyu
| Challenge: | Recent advances in Deep Research Agents (DRAs) are transforming automated knowledge discovery and problem-solving. |
| Approach: | They propose an inference-time scaling of verification wherein an agent self-improves at test time by evaluating its generated answers. |
| Outcome: | The proposed model outperforms vanilla agent-as-judge and LLM judge baselines by 12%–48% in meta-evaluation F1 score. |
Similar Papers
Agentic Rubrics as Contextual Verifiers for SWE Agents (2026.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have rapidly advanced on coding tasks, enabling increasingly capable software engineering agents for real-time code editing and bug fixing. |
| Approach: | They propose to use a rubric checklist to create a context-grounded rubric for SWE agents. |
| Outcome: | The proposed rubrics achieve a score of 54.2% on Qwen3-Coder-30B-A3B and 40.6% on Qween3-332B . |
Step-level Verifier-guided Hybrid Test-Time Scaling for Large Language Models (2025.emnlp-main)
Copied to clipboard
Kaiyan Chang, Yonghao Shi, Chenglong Wang, Hang Zhou, Chi Hu, Xiaoqian Liu, Yingfeng Luo, Yuan Ge, Tong Xiao, JingBo Zhu
| Challenge: | Recent training-based TTS methods, such as continued reinforcement learning, have surged in popularity, while training-free TTS approaches are gradually fading from prominence. |
| Approach: | They propose a fine-grained sequential scaling method guided by process verification that integrates training-free TTS methods with other classical parallel scaling methods at the step level. |
| Outcome: | Experiments on five instruction-tuned large language models (LLMs) show that training-free TTS methods can extend reasoning performance boundaries. |
AgentV-RL: Scaling Reward Modeling with Agentic Verifier (2026.findings-acl)
Copied to clipboard
Jiazheng Zhang, Ziche Fu, Zhiheng Xi, Wenqing Jing, Mingxu Chai, Wei He, Guoqiang Zhang, Chenghao Fan, Chenxin An, Wenxiang Chen, Zhicheng Liu, Haojie Pan, Dingwei Zhu, Tao Gui, Qi Zhang, Xuanjing Huang
| Challenge: | Existing approaches to improve LLM reasoning are limited in complex domains and lack external grounding makes verifiers unreliable on computation-intensive tasks. |
| Approach: | They propose a framework that transforms reward modeling into a multi-turn, tool-augmented deliberative process. |
| Outcome: | The proposed framework surpasses state-of-the-art ORMs by 25.2% under parallel and sequential TTS. |
S2R: Teaching LLMs to Self-verify and Self-correct via Reinforcement Learning (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to incentivize LLMs’ deep thinking abilities require large-scale data or significant training efforts. |
| Approach: | They introduce an efficient framework that enhances LLM reasoning by teaching models to self-verify and self-correct during inference. |
| Outcome: | The proposed framework outperforms models trained on long-CoT distilled data with 3.1k initialization samples and achieves an accuracy improvement of 51.0% to 81.6%. |
DeepFact: Co-Evolving Benchmarks and Agents for Deep Research Factuality (2026.acl-long)
Copied to clipboard
Yukun Huang, Leonardo F. R. Ribeiro, Momchil Hardalov, Bhuwan Dhingra, Markus Dreyer, Venkatesh Saligrama
| Challenge: | Existing fact-checkers usually target general-domain atomic claims . citation-grounded fact- checking ignores claims without explicit citations . |
| Approach: | They propose to use a benchmark to test whether claim-level factuality is transferable . they instantiate **Audit-then-Score** as a versioned DRR factualism benchmark . |
| Outcome: | The proposed benchmark outperforms the best prior deep-research and traditional fact-checkers by 14.3 and 24.9 points. |
DR-Arena: an Automated Evaluation Framework for Deep Research Agents (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for evaluating deep research capabilities rely on static datasets. |
| Approach: | They propose a fully automated evaluation framework that pushes DR agents to their capability limits through dynamic investigation. |
| Outcome: | DR-Arena achieves a Spearman correlation of 0.94 with the LMSYS Search Arena leaderboard. |
DeepResearcher: Scaling Deep Research via Reinforcement Learning in Real-world Environments (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) with web search capabilities show significant potential for deep research. |
| Approach: | They introduce a framework for end-to-end training of LLM-based deep research agents . they implement a specialized multi-agent architecture where browsing agents extract relevant information from various webpage structures. |
| Outcome: | The proposed framework improves on open-domain research tasks by 28.9 points over prompt engineering and 7.2 points over RAG-based RL agents. |
TTVS: Boosting Self-Exploring Reinforcement Learning via Test-time Variational Synthesis (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing test-time methods are limited in specialized or novel domains where supervision is prohibitively expensive or unavailable. |
| Approach: | They propose a framework that augments training stream from unlabeled test queries. |
| Outcome: | Extensive experiments show TTVS outperforms state-of-the-art RL-based techniques on unlabeled test-time data. |
CURE: Critique-Driven Unified Reinforcement Learning for Test-Time Self-Improvement (2026.acl-long)
Copied to clipboard
| Challenge: | Existing critique-guided methods fail to equip models with the autonomous improvement capabilities required for test-time scaling. |
| Approach: | They propose a framework that jointly optimizes a single policy for standard solving, critiquing, and guided re-exploration. |
| Outcome: | The proposed framework maintains competitive single-turn performance and unlocks effective inference-time scaling. |
Check Your Work: Structured Checklist Feedback for Improving Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Recent advances in Large Language Models have been driven by verifiable feedback in deterministic domains like mathematics and code. |
| Approach: | They propose to decompose granular, prompt-specific checklists into a scalar reward and use them to transform them into skalar rewards. |
| Outcome: | The proposed approach yields an 11.8% win-rate improvement on AlpacaEval 2.0 using Qwen3-8B, outperforming holistic reward models and existing checklist baselines. |