arXiv:2511.11978v1 Announce Type: new 
Abstract: Generative LLMs typically improve Named Entity Recognition (NER) performance through instruction tuning. They excel at generating entities by semantic pattern matching but lack an explicit, verifiable reasoning mechanism. This "cognitive shortcutting" leads to suboptimal performance and brittle generalization, especially in zero-shot and lowresource scenarios where reasoning from limited contextual cues is crucial. To address this issue, a reasoning framework is proposed for NER, which shifts the extraction paradigm from implicit pattern matching to explicit reasoning. This framework consists of three stages: Chain of Thought (CoT) generation, CoT tuning, and reasoning enhancement. First, a dataset annotated with NER-oriented CoTs is generated, which contain task-relevant reasoning chains. Then, they are used to tune the NER model to generate coherent rationales before deriving the final answer. Finally, a reasoning enhancement stage is implemented to optimize the reasoning process using a comprehensive reward signal. This stage ensures explicit and verifiable extractions. Experiments show that ReasoningNER demonstrates impressive cognitive ability in the NER task, achieving competitive performance. In zero-shot settings, it achieves state-of-the-art (SOTA) performance, outperforming GPT-4 by 12.3 percentage points on the F1 score. Analytical results also demonstrate its great potential to advance research in reasoningoriented information extraction. Our codes are available at https://github.com/HuiResearch/ReasoningIE.

تم اقتراح إطار جديد للتعرف على الكيانات المسماة (NER) لتحسين أداء نماذج اللغة الكبيرة التوليدية (LLMs) مثل GPT-4. على الرغم من أن هذه النماذج تتفوق في توليد الكيانات من خلال مطابقة الأنماط الدلالية، إلا أنها تفتقر غالبًا إلى آلية تفكير قوية، مما يؤدي إلى نتائج دون المستوى الأمثل، خاصة في السيناريوهات ذات الموارد المحدودة. ينتقل الإطار المقترح من مطابقة الأنماط الضمنية إلى التفكير الصريح، ويتكون من ثلاث مراحل: توليد سلسلة الأفكار (CoT)، وضبط CoT، وتعزيز التفكير، بهدف تحسين دقة NER.

Se ha propuesto un nuevo marco para el reconocimiento de entidades nombradas (NER) con el fin de mejorar el rendimiento de los grandes modelos de lenguaje generativos (LLMs) como GPT-4. Aunque estos modelos son excelentes en la generación de entidades mediante coincidencias semánticas, a menudo carecen de un mecanismo de razonamiento robusto, lo que lleva a resultados subóptimos, especialmente en escenarios de bajos recursos. El marco propuesto cambia el paradigma de la coincidencia de patrones implícitos a un razonamiento explícito, que consta de tres etapas: generación de cadenas de pensamie…

Un nouveau cadre pour la reconnaissance d'entités nommées (NER) a été proposé pour améliorer les performances des grands modèles de langage génératifs (LLMs) comme GPT-4. Bien que ces modèles excellent dans la génération d'entités par correspondance sémantique, ils manquent souvent d'un mécanisme de raisonnement robuste, ce qui entraîne des résultats sous-optimaux, en particulier dans des scénarios à faibles ressources. Le cadre proposé déplace le paradigme de la correspondance de motifs implicites vers un raisonnement explicite, impliquant trois étapes : génération de chaînes de pensée (CoT),…

A new framework for Named Entity Recognition (NER) has been proposed to enhance the performance of generative large language models (LLMs) like GPT-4. While these models excel at generating entities through semantic pattern matching, they often lack a robust reasoning mechanism, leading to suboptimal outcomes, particularly in low-resource scenarios. The proposed framework shifts the paradigm from implicit pattern matching to explicit reasoning, involving three stages: Chain of Thought (CoT) generation, CoT tuning, and reasoning enhancement, ultimately aiming to improve NER accuracy.

A Reasoning Paradigm for Named Entity Recognition

arXiv:2503.04971v2 Announce Type: replace 
Abstract: Foundation models (FMs) such as GPT-4 exhibit exceptional generative capabilities across diverse downstream tasks through fine-tuning. Split Federated Learning (SFL) facilitates privacy-preserving FM fine-tuning on resource-constrained local devices by offloading partial FM computations to edge servers, enabling device-edge synergistic fine-tuning. Practical edge networks often host multiple SFL tenants to support diversified downstream tasks. However, existing research primarily focuses on single-tenant SFL scenarios, and lacks tailored incentive mechanisms for multi-tenant settings, which are essential to effectively coordinate self-interested local devices for participation in various downstream tasks, ensuring that each SFL tenant's distinct FM fine-tuning requirements (e.g., FM types, performance targets, and fine-tuning deadlines) are met. To address this gap, we propose a novel Price-Incentive Mechanism (PRINCE) that guides multiple SFL tenants to offer strategic price incentives, which solicit high-quality device participation for efficient FM fine-tuning. Specifically, we first develop a bias-resilient global SFL model aggregation scheme to eliminate model biases caused by independent device participation. We then derive a rigorous SFL convergence bound to evaluate the contributions of heterogeneous devices to FM performance improvements, guiding the incentive strategies of SFL tenants. Furthermore, we model inter-tenant device competition as a congestion game for Stackelberg equilibrium (SE) analysis, deriving each SFL tenant's optimal incentive strategy. Extensive simulations involving four representative SFL tenant types (ViT, BERT, Whisper, and LLaMA) across diverse data modalities (text, images, and audio) demonstrate that PRINCE accelerates FM fine-tuning by up to 3.07x compared to state-of-the-art approaches, while consistently meeting fine-tuning performance targets.

تم اقتراح آلية جديدة للتسعير والتحفيز (PRINCE) لتحسين التعلم الفيدرالي متعدد المستأجرين (SFL) لنماذج الأساس (FMs) مثل GPT-4، مما يتيح ضبطًا فعالًا على الأجهزة ذات الموارد المحدودة مع الحفاظ على الخصوصية. تعالج هذه الآلية تحديات التنسيق التي تواجه عدة مستأجرين في SFL لديهم احتياجات ضبط متنوعة.

Se ha propuesto un nuevo Mecanismo de Precio-Incentivo (PRINCE) para mejorar el Aprendizaje Federado de Múltiples Inquilinos (SFL) para Modelos de Fundación (FMs) como GPT-4, permitiendo un ajuste eficiente en dispositivos con recursos limitados mientras se mantiene la privacidad. Este mecanismo aborda los desafíos de coordinación que enfrentan múltiples inquilinos de SFL con diversas necesidades de ajuste.

Un nouveau mécanisme de prix incitatif (PRINCE) a été proposé pour améliorer l'apprentissage fédéré à plusieurs locataires (SFL) pour les modèles de fondation (FMs) tels que GPT-4, permettant un ajustement efficace sur des appareils à ressources limitées tout en préservant la vie privée. Ce mécanisme répond aux défis de coordination rencontrés par plusieurs locataires SFL ayant des besoins d'ajustement divers.

A novel Price-Incentive Mechanism (PRINCE) has been proposed to enhance Multi-Tenant Split Federated Learning (SFL) for Foundation Models (FMs) like GPT-4, enabling efficient fine-tuning on resource-constrained devices while maintaining privacy. This mechanism addresses the coordination challenges faced by multiple SFL tenants with diverse fine-tuning needs.

Incentivizing Multi-Tenant Split Federated Learning for Foundation Models at the Network Edge

One More Thing in AI – Your Shortcut to AI Mastery

A Reasoning Paradigm for Named Entity Recognition

Was this article worth reading? Share it

One More Thing in AI

LucidQuery AI

Airparser

Humanize AI

Magicley AI

GPTHumanizer

Ready to build your own newsroom?