arXiv:2510.11654v2 Announce Type: replace-cross 
Abstract: Financial markets face growing threats from misinformation that can trigger billions in losses in minutes. Most existing approaches lack transparency in their decision-making and provide limited attribution to credible sources. We introduce FinVet, a novel multi-agent framework that integrates two Retrieval-Augmented Generation (RAG) pipelines with external fact-checking through a confidence-weighted voting mechanism. FinVet employs adaptive three-tier processing that dynamically adjusts verification strategies based on retrieval confidence, from direct metadata extraction to hybrid reasoning to full model-based analysis. Unlike existing methods, FinVet provides evidence-backed verdicts, source attribution, confidence scores, and explicit uncertainty flags when evidence is insufficient. Experimental evaluation on the FinFact dataset shows that FinVet achieves an F1 score of 0.85, which is a 10.4% improvement over the best individual pipeline (fact-check pipeline) and 37% improvement over standalone RAG approaches.

تم تقديم FinVet كإطار عمل جديد متعدد الوكلاء مصمم لمكافحة المعلومات المضللة المالية، التي تمثل تهديدات كبيرة للأسواق. تفتقر الأساليب التقليدية غالبًا إلى الشفافية ولا تقدم نسبًا لمصادر موثوقة. يدمج FinVet بين خطي أنظمة استرجاع المعلومات المعززة (RAG) مع التحقق الخارجي من الحقائق، باستخدام آلية تصويت مرجحة بالثقة. تظهر النتائج التجريبية على مجموعة بيانات FinFact أن FinVet يحقق درجة F1 تبلغ 0.85، مما يمثل تحسنًا بنسبة 10.4% مقارنة بالأساليب السابقة.

FinVet es un nuevo marco multiagente diseñado para combatir la desinformación financiera, que representa riesgos significativos para los mercados. Los métodos tradicionales a menudo carecen de transparencia y no proporcionan atribución a fuentes creíbles. FinVet integra dos pipelines de Generación Aumentada por Recuperación (RAG) con verificación externa de hechos, utilizando un mecanismo de votación ponderada por confianza. Los resultados experimentales en el conjunto de datos FinFact indican un puntaje F1 de 0.85, marcando una mejora del 10.4% sobre métodos anteriores.

FinVet est un nouveau cadre multi-agents conçu pour lutter contre la désinformation financière, qui représente des risques importants pour les marchés. Les méthodes traditionnelles manquent souvent de transparence et ne fournissent pas d'attribution de sources crédibles. FinVet intègre deux pipelines de génération augmentée par récupération (RAG) avec une vérification externe des faits, utilisant un mécanisme de vote pondéré par la confiance. Les résultats expérimentaux sur le jeu de données FinFact indiquent un score F1 de 0,85, marquant une amélioration de 10,4 % par rapport aux méthodes pré…

FinVet is a new multi-agent framework designed to combat financial misinformation, which poses significant risks to markets. Traditional methods often lack transparency and fail to provide credible source attribution. FinVet integrates two Retrieval-Augmented Generation (RAG) pipelines with external fact-checking, utilizing a confidence-weighted voting mechanism. It features a three-tier processing system that adapts verification strategies based on retrieval confidence. Experimental results on the FinFact dataset indicate an F1 score of 0.85, marking a 10.4% improvement over previous methods.

FinVet: A Collaborative Framework of RAG and External Fact-Checking Agents for Financial Misinformation Detection

arXiv:2512.05012v1 Announce Type: new 
Abstract: This extended abstract introduces Self-Explaining Contrastive Evidence Re-Ranking (CER), a novel method that restructures retrieval around factual evidence by fine-tuning embeddings with contrastive learning and generating token-level attribution rationales for each retrieved passage. Hard negatives are automatically selected using a subjectivity-based criterion, forcing the model to pull factual rationales closer while pushing subjective or misleading explanations apart. As a result, the method creates an embedding space explicitly aligned with evidential reasoning. We evaluated our method on clinical trial reports, and initial experimental results show that CER improves retrieval accuracy, mitigates the potential for hallucinations in RAG systems, and provides transparent, evidence-based retrieval that enhances reliability, especially in safety-critical domains.

يقدم إدخال إعادة تصنيف الأدلة المتباينة ذاتية الشرح (CER) طريقة جديدة لتعزيز أنظمة الجيل المعزز بالاسترجاع (RAG) من خلال التركيز على الأدلة الواقعية وتحسين دقة الاسترجاع. تستخدم هذه الطريقة التعلم المتباين لضبط التضمينات وتوليد تبريرات على مستوى الرموز لكل مقطع مسترجع، مما يميز بفعالية بين المعلومات الواقعية والمعلومات المضللة.

La introducción de la Re-Ranking de Evidencia Contrastiva Auto-Explicativa (CER) presenta un nuevo método para mejorar los sistemas de Generación Aumentada por Recuperación (RAG) al centrarse en la evidencia fáctica y mejorar la precisión de la recuperación. Este método emplea el aprendizaje contrastivo para afinar los embeddings y generar justificaciones a nivel de token para los pasajes recuperados, distinguiendo eficazmente entre información fáctica y engañosa.

L'introduction de la méthode de réévaluation des preuves contrastées auto-explicatives (CER) représente une nouvelle approche pour améliorer les systèmes de génération augmentée par récupération (RAG) en se concentrant sur les preuves factuelles et en améliorant la précision de la récupération. Cette méthode utilise l'apprentissage contrastif pour affiner les embeddings et générer des justifications au niveau des tokens pour les passages récupérés, distinguant efficacement les informations factuelles des informations trompeuses.

The introduction of Self-Explaining Contrastive Evidence Re-Ranking (CER) presents a new method for enhancing Retrieval-Augmented Generation (RAG) systems by focusing on factual evidence and improving retrieval accuracy. This method employs contrastive learning to fine-tune embeddings and generate token-level rationales for retrieved passages, effectively distinguishing between factual and misleading information.

Factuality and Transparency Are All RAG Needs! Self-Explaining Contrastive Evidence Re-ranking

arXiv:2510.02967v2 Announce Type: replace 
Abstract: This paper presents the development and evaluation of a Retrieval-Augmented Generation (RAG) system for querying the United Kingdom's National Institute for Health and Care Excellence (NICE) clinical guidelines using Large Language Models (LLMs). The extensive length and volume of these guidelines can impede their utilisation within a time-constrained healthcare system, a challenge this project addresses through the creation of a system capable of providing users with precisely matched information in response to natural language queries. The system's retrieval architecture, composed of a hybrid embedding mechanism, was evaluated against a corpus of 10,195 text chunks derived from three hundred guidelines. It demonstrates high performance, with a Mean Reciprocal Rank (MRR) of 0.814, a Recall of 81% at the first chunk and of 99.1% within the top ten retrieved chunks, when evaluated on 7901 queries.
  The most significant impact of the RAG system was observed during the generation phase. When evaluated on a manually curated dataset of seventy question-answer pairs, RAG-enhanced models showed substantial gains in performance. Faithfulness, the measure of whether an answer is supported by the source text, was increased by 64.7 percentage points to 99.5% for the RAG-enhanced O4-Mini model and significantly outperformed the medical-focused Meditron3-8B LLM, which scored 43%. Clinical evaluation by seven Subject Matter Experts (SMEs) further validated these findings, with GPT-4.1 achieving 98.7% accuracy while reducing unsafe responses by 67% compared to O4-Mini (from 3.0 to 1.0 per evaluator). This study thus establishes RAG as an effective, reliable, and scalable approach for applying generative AI in healthcare, enabling cost-effective access to medical guidelines.

تم تطوير نظام جديد لتوليد البيانات المعزز بالاسترجاع (RAG) لتحسين استعلامات الإرشادات السريرية للمعهد الوطني للصحة والرعاية الممتازة (NICE) في المملكة المتحدة باستخدام نماذج اللغة الكبيرة (LLMs). يتناول هذا النظام التحديات التي تطرحها الطول الكبير للإرشادات، حيث يوفر للمستخدمين معلومات دقيقة استجابة لاستفسارات اللغة الطبيعية. حقق النظام مرتبة متوسطة عكسية (MRR) تبلغ 0.814 واسترجاع بنسبة 81% في الجزء الأول خلال التقييمات على 7901 استفسار.

Se ha desarrollado un nuevo sistema de Generación Aumentada por Recuperación (RAG) para mejorar la consulta de las directrices clínicas del Instituto Nacional para la Salud y la Excelencia en el Cuidado (NICE) del Reino Unido, utilizando Modelos de Lenguaje de Gran Tamaño (LLMs). Este sistema aborda los desafíos que presenta la extensa longitud de las directrices, proporcionando a los usuarios información precisa en respuesta a consultas en lenguaje natural. El sistema logró un rango recíproco medio (MRR) de 0.814 y un Recall del 81% en el primer fragmento durante las evaluaciones en 7901 cons…

Un nouveau système de génération augmentée par récupération (RAG) a été développé pour améliorer l'interrogation des directives cliniques de l'Institut national pour la santé et l'excellence des soins (NICE) au Royaume-Uni, en utilisant des modèles de langage de grande taille (LLMs). Ce système répond aux défis posés par la longueur des directives, fournissant aux utilisateurs des informations précises en réponse à des requêtes en langage naturel. Le système a atteint un rang réciproque moyen (MRR) de 0,814 et un rappel de 81 % au premier segment lors des évaluations sur 7901 requêtes.

A new Retrieval-Augmented Generation (RAG) system has been developed to enhance the querying of the UK National Institute for Health and Care Excellence (NICE) clinical guidelines using Large Language Models (LLMs). This system addresses the challenges posed by the extensive length of guidelines, providing users with accurate information in response to natural language queries. The system achieved a Mean Reciprocal Rank (MRR) of 0.814 and a Recall of 81% at the first chunk during evaluations on 7901 queries.

Grounding Large Language Models in Clinical Evidence: A Retrieval-Augmented Generation System for Querying UK NICE Clinical Guidelines

arXiv:2508.08785v2 Announce Type: replace 
Abstract: LLMs often suffer from hallucinations and outdated or incomplete knowledge. RAG is proposed to address these issues by integrating external knowledge like that in KGs into LLMs. However, leveraging private KGs in RAG systems poses significant privacy risks due to the black-box nature of LLMs and potential insecure data transmission, especially when using third-party LLM APIs lacking transparency and control. In this paper, we investigate the privacy-protected RAG scenario for the first time, where entities in KGs are anonymous for LLMs, thus preventing them from accessing entity semantics. Due to the loss of semantics of entities, previous RAG systems cannot retrieve question-relevant knowledge from KGs by matching questions with the meaningless identifiers of anonymous entities. To realize an effective RAG system in this scenario, two key challenges must be addressed: (1) How can anonymous entities be converted into retrievable information. (2) How to retrieve question-relevant anonymous entities. Hence, we propose a novel ARoG framework including relation-centric abstraction and structure-oriented abstraction strategies. For challenge (1), the first strategy abstracts entities into high-level concepts by dynamically capturing the semantics of their adjacent relations. It supplements meaningful semantics which can further support the retrieval process. For challenge (2), the second strategy transforms unstructured natural language questions into structured abstract concept paths. These paths can be more effectively aligned with the abstracted concepts in KGs, thereby improving retrieval performance. To guide LLMs to effectively retrieve knowledge from KGs, the two strategies strictly protect privacy from being exposed to LLMs. Experiments on three datasets demonstrate that ARoG achieves strong performance and privacy-robustness.

تم اقتراح نهج جديد لتوليد البيانات المعزز بالاسترجاع (RAG) يركز على حماية الخصوصية في إجابات الأسئلة المتعلقة برسوم المعرفة. تقوم هذه الطريقة بإخفاء هوية الكيانات داخل رسوم المعرفة، مما يمنع نماذج اللغة الكبيرة (LLMs) من الوصول إلى الدلالات الحساسة، مما يعالج المخاطر الكبيرة للخصوصية المرتبطة بأنظمة RAG التقليدية.

Se ha propuesto un nuevo enfoque para la Generación Aumentada por Recuperación (RAG) que se centra en la protección de la privacidad en la respuesta a preguntas sobre grafos de conocimiento. Este método anonimiza las entidades dentro de los grafos de conocimiento, impidiendo que los modelos de lenguaje de gran tamaño (LLMs) accedan a semánticas sensibles, lo que aborda riesgos significativos de privacidad asociados con los sistemas RAG tradicionales.

Une nouvelle approche de la génération augmentée par récupération (RAG) a été proposée, axée sur la protection de la vie privée dans le cadre des questions liées aux graphes de connaissances. Cette méthode anonymise les entités au sein des graphes de connaissances, empêchant les modèles de langage de grande taille (LLMs) d'accéder à des sémantiques sensibles, ce qui répond à des risques de confidentialité significatifs associés aux systèmes RAG traditionnels.

A new approach to Retrieval-Augmented Generation (RAG) has been proposed, focusing on privacy protection in knowledge graph question answering. This method anonymizes entities within knowledge graphs, preventing large language models (LLMs) from accessing sensitive semantics, which addresses significant privacy risks associated with traditional RAG systems.

FinVet: A Collaborative Framework of RAG and External Fact-Checking Agents for Financial Misinformation Detection

Was this article worth reading? Share it

finlight.me

ZeroGPT.org

Fincent