The 10th Workshop on Financial Technology and Natural Language Processing

EMNLP-2025, Nov. 8th, 2025, Suzhou, China

Workshop Description


The FinNLP workshop, as the annual event of ACL SIG-FinTech, focuses on applying NLP, Machine Learning, and Large Language Models to finance, economics, and law, aiming to drive interdisciplinary innovation and address unique challenges in these fields. Since 2019, FinNLP has served as a bridge between AI and NLP communities by collocating with major conferences. Its proceedings are available on the ACL Anthology.

FinNLP addresses the complexities of financial documents, which often contain tables and diverse multimodal data, as well as legal and compliance challenges in the financial industry. Key workshop topics include financial language modeling, multi-modal and graph representation learning, conversational agents, financial search and question answering, event discovery, compliance, and responsible AI practices in finance and law.

We have several interesting shared tasks in the collocated event, FinEval. Please refer to the FinEval page for more details. Here is the call for paper page: Call for Paper .



Program


Nov. 9 (UTC+8)

Time Topic
09:00 - 09:10Opening
09:10 - 10:00Invited Talk - Prof. Saeed Abdullah (Penn State): Following the Money: How Financial Data Can Transform Mental Health Research and Care
10:00 - 10:12Natural Language Inference as a Judge: Detecting Factuality and Causality Issues in Language Model Self-Reasoning for Financial Analysis
10:12 - 10:24SEC-QA: A Systematic Evaluation Corpus for Financial QA
10:24 - 10:36FinEval-KR: A Financial Domain Evaluation Framework for Large Language Models' Knowledge and Reasoning
10:36 - 11:00Coffee Break
11:00 - 11:12LAVA: Logic-Aware Validation and Augmentation Framework for Large-Scale Financial Document Auditing
11:12 - 11:24FinCoT: Grounding Chain-of-Thought in Expert Financial Reasoning
11:24 - 11:36Leveraging LLM-based sentiment analysis for portfolio optimization with proximal policy optimization
11:36 - 11:48Do Companies Reveal Their Own Fraud? - A Novel Data Set for Fraud Detection Based on 10-K Reports
11:48 - 12:00Synthesizing Behaviorally-Grounded Reasoning Chains: A Data-Generation Framework for Personal Finance LLMs
12:00 - 12:10LLM as a Guide: an Approach for Unsupervised Economic Relation Discovery in Administrative Documents
12:10 - 12:20Zero-Shot Extraction of Stock Relationship Graphs with LLMs
12:20 - 12:30Enhancing Financial RAG with Agentic AI and Multi-HyDE: A Novel Approach to Knowledge Retrieval and Hallucination Reduction
12:30 - 14:00Lunch
14:00 - 15:00Invited Talk - Dr. Shi-Xiong (Austin) Zhang & Dr. Sambit Sahu (Capital One): Orchestrating LLMs for Complex Financial Reasoning with Multi-Agentic Workflow
15:00 - 15:10Earnings2Insights: Analyst Report Generation for Investment Guidance
15:10 - 15:18DKE : FinDebate: Multi-Agent Collaborative Intelligence for Financial Analysis
15:18 - 15:24FinTurbo: Beyond Summaries: Multi-Agent Generation of Investment Reports with Text, Tables, and Charts
15:24 - 15:32LangKG: LangKG at the FinNLP 2025 - Earnings2Insights: Task-Adaptive LLMs To Generate Human-Persuasive Investment Reports
15:32 - 16:00Break
16:00 - 17:00Invited Talk - Prof. Andrea Rocci (Università della Svizzera italiana)
17:00 - 18:00 Poster Session
  • All Oral-Presented Papers
  • StockGenChaR: A Study on the Evaluation of Large Vision-Language Models on Stock Chart Captioning
  • Towards Efficient FinBERT via Quantization and Coreset for Financial Sentiment Analysis
  • Assessing RAG System Capabilities on Financial Documents (FiQA)
  • Detecting Evasive Answers in Financial Q&A: A Psychological Discourse Taxonomy and Lightweight Baselines
  • Enhancing Post Earnings Announcement Drift Measurement with Large Language Models
  • A Self-Improving Method for Generating Descriptions of Financial Data Quality Grading Using LLMs
  • From Earnings calls to Investment Reports: Evaluating Role based Multi-Agent LLM Systems
  • Structured Adversarial Synthesis: A Multi-Agent Framework for Generating Persuasive Financial Analysis from Earning Call Transcripts
  • Meta Prompting for Analyst Report Generation: Turning Earnings Calls into Investment Guidance
  • Jetsons at the FinNLP-2025 - Earnings2Insights: Persuasive Investment Report Generation Using Single And Multi-Agent Frameworks
  • Agentic LLMs for Analyst-Style Financial Insights: An LLM Pipeline for Persuasive Financial Analysis
  • SI4Fin at Earnings2Insights: LLM-Based Analyst Report Generation for Earnings Calls