From Fragments to Facts: A Curriculum-Driven DPO Approach for Generating Hindi News Veracity Explanations
On the map: LLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Abstract (arXiv)
In an era of rampant misinformation, generating reliable news explanations is vital, especially for under-represented languages like Hindi. Lacking robust automated tools, Hindi faces challenges in scaling misinformation detection. To bridge this gap, we propose a novel framework integrating Direct Preference Optimization (DPO) with curriculum learning to align machine-generated explanations with human reasoning. Fact-checked explanations from credible sources serve as preferred responses, while LLM outputs highlight system limitations and serve as non-preferred responses. To refine task-specific alignment, we introduce two key parameters -- Actuality and Finesse -- into the DPO loss function, enhancing explanation quality and consistency. Experiments with LLMs (Mistral, Llama, Gemma) and PLMs (mBART, mT5) confirm the framework's effectiveness in generating coherent, contextually relevant explanations. This scalable approach combats misinformation and extends automated explanation generation to low-resource languages.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 39 of its 76 references that are in the map. First posted on arXiv 7 Jul 2025.
Builds on
39 of 76 references.
Closest ancestors · 5
Same direction, newest first
- 2024BPO: Towards Balanced Preference Optimization between Knowledge Breadth and Depth in Alignment16 Nov
- 2024Self-Augmented Preference Optimization: Off-Policy Paradigms for Language Model Alignment31 May
- 2024Filtered Direct Preference Optimization22 Apr
- 2024Token-level Direct Preference Optimization18 Apr
- 2024Curry-DPO: Enhancing Alignment using Curriculum Learning & Ranked Preferences12 Mar
Other references · 34
- 2023GPT-4 Technical Reportcited by 18,396
- 2024The Llama 3 Herd of Modelscited by 14,715
- 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
- 2019BERTScore: Evaluating Text Generation with BERTcited by 5,512
- 2015Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learningcited by 4,448
- 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,224
- 2023Mistral 7Bcited by 3,164
- 2020mT5: A massively multilingual pre-trained text-to-text transformercited by 1,818
- 2023FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generationcited by 1,282
- 2023Diffusion Model Alignment Using Direct Preference Optimizationcited by 790
- 2024A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applicationscited by 431
- 2020Multilingual Translation with Extensible Multilingual Pretraining and Finetuningcited by 287
22 more
- 2024Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMscited by 273
- 2023Large Language Models for Generative Information Extraction: A Surveycited by 185
- 2024Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantificationcited by 162
- 2022End-to-End Multimodal Fact-Checking and Explanation Generation: A Challenging Dataset and Modelscited by 107
- 2020Explainable Automated Fact-Checking: A Surveycited by 83
- 2024On Softmax Direct Preference Optimization for Recommendationcited by 64
- 2020TRIE: End-to-End Text Reading and Information Extraction for Document Understandingcited by 61
- 2024Curriculum Direct Preference Optimization for Diffusion and Consistency Modelscited by 36
- 2024Evidence-Driven Retrieval Augmented Response Generation for Online Misinformationcited by 32
- 2020Hostility Detection Dataset in Hindicited by 31
- 2024LongGenBench: Long-context Generation Benchmarkcited by 28
- 2024Denoising Task Difficulty-based Curriculum for Training Diffusion Modelscited by 27
- 2023Benchmarking the Generation of Fact Checking Explanationscited by 25
- 2013Polyglot: Distributed Word Representations for Multilingual NLPcited by 19
- 2021Explainable Fact-checking through Question Answeringcited by 16
- 2023DecipherPref: Analyzing Influential Factors in Human Preference Judgments via GPT-4cited by 16
- 2024Measuring and Modifying the Readability of English Texts with GPT-4cited by 10
- 2024Knowledge-Aware Reasoning over Multimodal Semi-structured Tablescited by 9
- 2023Axiomatic Preference Modeling for Longform Question Answeringcited by 6
- 2022Explainable Misinformation Detection Across Multiple Social Media Platformscited by 5
- 2024MMCFND: Multimodal Multilingual Caption-aware Fake News Detection for Low-resource Indic Languagescited by 5
- 2023Prompt-Learning for Cross-Lingual Relation Extractioncited by 4
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2026Are Rationales Necessary and Sufficient? Tuning LLMs for Explainable Misinformation Detection
- 2024Mining the Explainability and Generalization: Fact Verification Based on Self-Instruction
- 2024Anchored Alignment for Self-Explanations Enhancement
- 2024MMIDR: Teaching Large Language Model to Interpret Multimodal Misinformation via Knowledge Distillation
- 2026LLM-MRD: LLM-Guided Multi-View Reasoning Distillation for Fake News Detection
- 2022Explanations from Large Language Models Make Small Reasoners Better
- 2025Llama-3-Nanda-10B-Chat: An Open Generative Large Language Model for Hindi
- 2024Insights into Alignment: Evaluating DPO and its Variants Across Multiple Tasks