# Optimizing Decomposition for Optimal Claim Verification

arXiv: https://arxiv.org/abs/2503.15354 · HTML: https://mlsanity.com/paper/2503.15354/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 33 of its 52 references that are in the map.
- First posted on arXiv 19 Mar 2025.

## Builds on (most cited first, 30 of 33)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation](https://mlsanity.com/paper/1406.1078/) (2014)
- [Asynchronous Methods for Deep Reinforcement Learning](https://mlsanity.com/paper/1602.01783/) (2016)
- [High-Dimensional Continuous Control Using Generalized Advantage Estimation](https://mlsanity.com/paper/1506.02438/) (2015)
- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [FEVER: a large-scale dataset for Fact Extraction and VERification](https://mlsanity.com/paper/1803.05355/) (2018)
- [FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation](https://mlsanity.com/paper/2305.14251/) (2023)
- [Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks](https://mlsanity.com/paper/2204.07705/) (2022)
- [A Survey on Automated Fact-Checking](https://mlsanity.com/paper/2108.11896/) (2021)
- [Dynamic Prompt Learning via Policy Gradient for Semi-structured Mathematical Reasoning](https://mlsanity.com/paper/2209.14610/) (2022)
- [MiniCheck: Efficient Fact-Checking of LLMs on Grounding Documents](https://mlsanity.com/paper/2404.10774/) (2024)
- [FacTool: Factuality Detection in Generative AI -- A Tool Augmented Framework for Multi-Task and Multi-Domain Scenarios](https://mlsanity.com/paper/2307.13528/) (2023)
- [Get Your Vitamin C! Robust Fact Verification with Contrastive Evidence](https://mlsanity.com/paper/2103.08541/) (2021)
- [Active Example Selection for In-Context Learning](https://mlsanity.com/paper/2211.04486/) (2022)
- [Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO](https://mlsanity.com/paper/2005.12729/) (2020)
- [Diverse Demonstrations Improve In-context Compositional Generalization](https://mlsanity.com/paper/2212.06800/) (2022)
- [A Review on Bilevel Optimization: From Classical to Evolutionary Approaches and Applications](https://mlsanity.com/paper/1705.06270/) (2017)
- [WiCE: Real-World Entailment for Claims in Wikipedia](https://mlsanity.com/paper/2303.01432/) (2023)
- [VERISCORE: Evaluating the factuality of verifiable claims in long-form text generation](https://mlsanity.com/paper/2406.19276/) (2024)
- [Complex Claim Verification with Evidence Retrieved in the Wild](https://mlsanity.com/paper/2305.11859/) (2023)
- [Factcheck-Bench: Fine-Grained Evaluation Benchmark for Automatic Fact-checkers](https://mlsanity.com/paper/2311.09000/) (2023)
- [A Closer Look at Claim Decomposition](https://mlsanity.com/paper/2403.11903/) (2024)
- [Decomposition Dilemmas: Does Claim Decomposition Boost or Burden Fact-Checking Performance?](https://mlsanity.com/paper/2411.02400/) (2024)
- [OpenFactCheck: A Unified Framework for Factuality Evaluation of LLMs](https://mlsanity.com/paper/2408.11832/) (2024)
- [In-context Learning Generalizes, But Not Always Robustly: The Case of Syntax](https://mlsanity.com/paper/2311.07811/) (2023)
- [Atomic Inference for NLI with Generated Facts as Atoms](https://mlsanity.com/paper/2305.13214/) (2023)

## Built on by (most cited first, 10 of 10)

- [Coordinating Search-Informed Reasoning and Reasoning-Guided Search in Claim Verification](https://mlsanity.com/paper/2506.07528/) (2025)
- [Extractive Fact Decomposition for Interpretable Natural Language Inference in one Forward Pass](https://mlsanity.com/paper/2509.18901/) (2025)
- [Distill and Align Decomposition for Enhanced Claim Verification](https://mlsanity.com/paper/2602.21857/) (2026)
- [Scaling Participation in Modular AI Systems](https://mlsanity.com/paper/2606.07812/) (2026)
- [Evidence Absence Is Not Evidence Insufficiency: Diagnosing NEI Construction Artifacts in Fact Verification](https://mlsanity.com/paper/2605.26663/) (2026)
- [CREDENCE: Claim Reduction for Decomposition & Enhanced Credibility -- Semantic Metrics and Convergence Analysis](https://mlsanity.com/paper/2606.19819/) (2026)
- [When Verification Fails: How Compositionally Infeasible Claims Escape Rejection](https://mlsanity.com/paper/2604.10990/) (2026)
- [The Warrant Gap: Claim-Conditioned Re-scoring for Fact-Checking](https://mlsanity.com/paper/2606.24627/) (2026)
- [ElementCheck: Complexity-Aware Long-Form Text Factuality Evaluation via Sentence Elements](https://mlsanity.com/paper/2608.26118/) (2026)
- [MedSNIP: Building and Benchmarking Snippet-Level Granularity for Medical Fact Verification](https://mlsanity.com/paper/2609.12884/) (2026)

## Abstract (arXiv)

Current research on the textitDecompose-Then-Verify paradigm for evaluating the factuality of long-form text typically treats decomposition and verification in isolation, overlooking their interactions and potential misalignment. We find that existing decomposition policies, typically hand-crafted demonstrations, do not align well with downstream verifiers in terms of atomicity -- a novel metric quantifying information density -- leading to suboptimal verification results. We formulate finding the optimal decomposition policy for optimal verification as a bilevel optimization problem. To approximate a solution for this strongly NP-hard problem, we propose dynamic decomposition, a reinforcement learning framework that leverages verifier feedback to learn a policy for dynamically decomposing claims to verifier-preferred atomicity. Experimental results show that dynamic decomposition outperforms existing decomposition policies, improving verification confidence by 0.07 and accuracy by 0.12 (on a 0-1 scale) on average across varying verifiers, datasets, and atomcities of input claims.