Accepted Papers

Oral Papers

Oral papers are presented in two sessions (5 minutes each, followed by joint Q&A) and also present a poster during the lunch session. Papers marked Remote are presented online. See the programme for times.

Oral Session I: Financial Language Agents (09:45–10:30)

1. How LLM Financial Agents Respond to Stated Affect Depends on Their Assigned RoleLongDa Chen Lian, Yu-Lin Chang, Yen-Hsi Huang, Shu-Kai Hsieh
2. When Helpful Agents Cause Financial Harm: Evaluating LLM Agent Safety in Executable Economic WorldsLongRemoteYuzhu Ling, Jiale Han, Benyou Wang
3. DealTrace: Localizing and Intervening on Failures in Agentic Financial ReasoningLongAlina Hyk
4. MortarBench: Evaluating Mortgage Loan Origination AgentsLongMatthew Toles, Yunan Lu, Manav Munjal, Bojun Liu, Yuanhao Deng, Stephanie Selig, Derek Rindner, Cheng Li, Zhou Yu
5. Theatron: A Benchmark for Evaluating Persona Persistence in Long-Horizon LLM Decision-MakingLongYuyan Wang, Yuechen Jiang, Zhiwei Liu, Jimin Huang, Sophia Ananiadou
6. AlphaMemo: Structured Search-Process Memory for Self-Evolving Alpha Mining AgentsLongHang Yu, Zifan Zheng, Jeff Z. Pan, Tongliang Liu, Zhiyong Wang, Fengxiang He

Oral Session II: Financial Text Understanding and Reasoning (16:00–16:45)

1. FinLCR: Scaling Financial Long-Context Reasoning via Curriculum Reinforcement LearningShortWenjie Qiu, Yong Xie, Karan Aggarwal, Aitzaz Ahmad, Stephen Lau
2. Characterizing Ambiguity in Financial Document Question AnsweringLongKelsey Rook, Clare Arrington, Tomek Strzalkowski, Razvan Bunescu
3. Diagnosing Fine-Grained Inconsistency Classification in Financial Disclosure TextLongAman Kumar, Lasitha Vidyaratne, Dipanjan Dipak Ghosh, Arnab Chakrabarti, Ahmed K. Farahat
4. What Ensemble Brier Scores Hide: Evidence Routing at a Fixed Article-Exposure Budget in LLM Prediction-Market ForecastingLongRemoteAvinash Goutham Aluguvelly
5. GRPO for Financial Advice Generation: Outperforming Commercial LLMs under CATE EvaluationLongOfir Ben Shoham, Shrutendra Harsola, Vignesh Subrahmaniam, Shravan Mohan, Yakov Gazman, Oded Vainas
6. ShiJianBench: From Dialogue to Decision for Long-Horizon Evaluation of Investment AdvisorsLongJie Gong, Maowei Jiang, Zhiwei Liu, Yang Qiao, Wenxi Wu, Mengxi Xiao, Enze Zhang, Ziyan Kuang, Yankai Chen, Caishuang Huang, Meng Zhou, Xiku Du, Xue Liu, Guojun Xiong, Min Peng, Qianqian Xie, Sophia Ananiadou

Poster Papers

Presented at the poster session, 12:30–14:00 in Hall H. Listed alphabetically by title.

• Auditing Bank Statement PDFs: A Multi-Tier Parser and Design Choices for the NL-to-SQL Query Layer Above ItLongAnamul Hoque Emtiaj, Anindya Iqbal
• Auditing Scaffolding Gains in LLM Financial Extraction When the Label Is the MethodLongArthus le Vaillant de Chaudenay, Arij Riabi, Valentin Noël
• Bags of Words, Bags of Analysts: Two Event Spaces for Industry ClassificationShortArtem Lensky
• BankShield: Fairness-Aware Safety Guardrails for Indian-Language Banking AILongPiyush Agarwal
• Beyond Compressibility: Transferable Low-Dimensional Structure and the Limits of Compression in Financial RetrievalLongOddharak Tyagi, Shashishekar Ramakrishna, Chetan Harsha, Sridhar Dasaratha
• Can LLMs Catch a Rigged Backtest? A Clean-Control Calibration BenchmarkShortMakar Ulesov, Vladislav Smirnov, Omar Ibrahim, Arsenii Bobovnikov
• CARPAS: Towards Content-Aware Refinement of Provided Aspects for Summarization in Large Language ModelsLongYong-En Tian, Yu-Chien Tang, An-Zi Yen, Wen-Chih Peng
• CAST: An Interpretable Evidence-Based Confidence Estimation Tool for RAG with Human-in-the-Loop VerificationLongXutong Liu, Morris Groot Beumer, Joerg Robert Osterrieder, Max Baak
• CLAIR-Fin: An Adversarial Multi-Agent Framework for Claim-Level Verification and Adaptive Debate in Cross-Modal Financial QALongFatema Tuj Johora Faria, Mukaffi Bin Moin, Jubayer Al Mahmud, M. F. Mridha, Md. Alam Hossain
• Continuous Financial Knowledge Graph Construction with Citation-Grounded Agentic Question AnsweringLongArnau Rodon Comas
• CORAL: Closed-loop Oracle Alignment for Agentic Tax Engine ModernizationLongMaarten De Raedt, Ahmed Abdullah, Abhinav Agrawal
• Definitional or Computational? Decomposing LLM Numeric Fidelity in Financial-Statement Analysis with an XBRL Cascade OracleLongOussama Chahed
• DelistBench: Evaluating Search-Enabled LLMs for Auditable Corporate-Event Database CompletionLongXuan Yao, Shuping Li, Yang Dai, Yi Zhou, Ke-Wei Huang
• DeXposure-Claw: An Agent System for DeFi Risk SupervisionLongAijie Shu, Bowei Chen, Wenbin Wu, Cathy Yi Hsuan Chen, Fengxiang He
• Do Tools Matter, or Just Their Descriptions?LongAnanya Tomar, Ayush Nangia
• EVOQUANT: Self-Evolving Verifier-Guided Strategy Optimization for Robust Quantitative TradingLongJie Mao, Changlun Li, Xiang Li, Qiqi Duan, Jinhui Yuan, Xiang Liu, Yuyu Luo, Jing Tang, Nan Tang
• FinDRA: Multi-Agent Financial Reasoning under Asymmetric Information and Unequal AuthorityLongInyoung Won, Yunyoung Lee
• FinEngage: Benchmarking Multimodal Language Models on Content-Conditioned Human Feedback InferenceLongQiqi Qiang, Ying Peng, Yunmiao Zhang, Zhiwei Liu, Victor Gutierrez Basulto, Yifei Zhang, Dejun Cao, Jie Gong, Min Peng, Honghai Yu, Qianqian Xie
• FinEvo: Financial Language Agents Shape the Markets That Evaluate ThemLongMingxi Zou, Jiaxiang Chen, Langzhang Liang, Zhuo Wang, Aotian Luo, Chi Zhang, Danny Dongning Sun, Zenglin Xu
• FRED-SemBench: Evaluating Semantic Reliability in LLM Access to Macroeconomic DataShortWilson Wang, Chandler Han, Peter Zhang
• GreenAudit-Bench: Evidence-Grounded Auditing of Compositional Greenwashing in Corporate Sustainability DisclosuresLongShixiang You, Zeyuan Wang, Yuechen Jiang, Zhiwei Liu, Jie Gao, Sophia Ananiadou, Zhuoran Lu
• Helping Customers in Distress: An LLM-powered Agent that Converses, Probes, and RoutesShortAlankar Atreya, Stefan Sylvius Wagner, Devesh Batra, Robert Hankache, Cristovão Iglesias Jr, Patrick Sinclair, Giulio Pelosio, Michael McMillan, Greig A Cowan, Raad Khraishi
• Hidden Below the Landing Page: Tracking Linked Pages and Relation-Centered DOM Summaries for Financial Phishing DetectionShortHyeji Heo, Dae-Young Park, Minwoo Kim
• Human Agreement and Return Association Are Not Interchangeable CriteriaLongA S Aravinthakshan, Laven Srivastava, Harsh Nandwani
• Know Thyself: Concept-Level Diagnosis for Financial Large Language ModelsLongZiyan Kuang, Maowei Jiang, Feiyu Zhu, Ziyang Xu, Zhentao Li, Zhiwei Hu, Victor Gutierrez Basulto, Yanzhao Lai, Wenjin Feng, Zhitong Wang, Zelin Wang, Chen Xu, Meikang Qiu, Jiajia Huang, Min Peng, Qianqian Xie, Sophia Ananiadou
• LatentDemandBench: A Contamination-Resistant Diagnostic for Pre-Recognition Product–Need Ranking in Synthetic EconomiesLongAvinash Goutham Aluguvelly
• Market Signal Injection: Adversarial Context Manipulation of LLM Pricing AgentsLongDohun Lee, Hyunwoo Park
• NLP for Crowdfunding Success Prediction: A Systematic Mapping Review Exposes an Underexplored Equity DomainLongChaya Liebeskind, Hana Tzur, Shmuel Liebeskind
• One Advisor, Many Faces: How User Personas Shape LLM Stock RecommendationsLongYuxian He
• Perceived Professionalism in Financial Inquiry: An Interpretable and Empirically Grounded Linguistic Modeling of Textual PreferenceLongGiulia D'Agostino, Chung-Chi Chen
• Prediction-Market Seed Capital Recovery from Noise-Dominant FlowLongYankai Chen, Bowei He, Zhuohan Xie, Zhiwei Liu, Fan Zhang, Xueqing Peng, Yan Wang, Lingfei Qian, Xue Liu
• Preliminary Study on LLM-as-a-Judge Evaluation of Causal Mechanism Explanations in Financial News and its Bias AnalysisLongYucheng Xie, Yurina Kosai, Rikuto Tsuchida, Takehito Utsuro
• Regression-Aware Statistically Gated Policy Updating for Korean Financial LLM Input GuardrailsShortEunjung Cho, Dae-Young Park, Minwoo Kim
• Relevant Is Not Enough: Expert-Audited Decision-Grade Web Evidence for Financial Fact VerificationShortHyunkyung Park, Eui Hyun Kim, Wooah Choi, Arkaitz Zubiaga
• Same Financials, Different Identity: A Counterfactual Audit of Demographic Sensitivity in LLM-Based Credit-Risk ScoringShortMustafa Furkan Zenginoğlu, Zeynep Beyda Kara, Hasan Fatih Durkaya, Mehmet Haklidir, Tuna Çakar
• Same Meaning, Different Wording: Evaluating Paraphrase Robustness in Financial Natural Language InferenceShortYike Li, Adam Poliak
• Scripted Hesitation in a Financial Customer-Service Chatbot: A Brazilian Portuguese Case StudyShortValeria Vieira dos Santos
• Selective Compliance in Financial Rule Reasoning: A Solver-Verified Chinese Benchmark and Empirical AnalysisLongPeng Li, Yinchun Wang
• SpendSight: Privacy-Preserving Personal Finance Q&A with Sub-2B LLMsLongSajag Swami, Yash Pandey
• Topic‑Aware Environmental Claim Extraction from Corporate NewsLongTanay Kumar Gupta, Vaishali Aggarwal, Ishan Verma
• Trade-Adaptive Aggregation of Probabilistic Financial ForecastsLongYankai Chen, Rassul Magauin, Bowei He, Anuar Aimoldin, Sirui Song, Bin Xiao, Zangir Iklassov, Xue Liu
• Trust the Query, Not the Answer: Auditable Text-to-SPARQL over the GLEIF RegisterDemoJannic Alexander Cutura
• TRUST-BENCH: A Synthetic Japanese Financial-Group Benchmark for Enterprise Document Access in Coding AgentsLongShamik Kundu, Anand Mekala, Kazutaka Yamasaki
• Valid but Wrong: Domain Invariants and a Cross-Model Plan-Validity Benchmark for Household-Finance LLMsShortVignesh Coumarane
• What Deterministic Verification Actually Fixes in Financial QA: A Diagnostic Analysis of Rule Heterogeneity Across DatasetsLongAaditya Anand Thokal
• What the Window Does Not Contain: Auditing Provenance in a Document-Grounded Instability BenchmarkLongSeyed Mosayeb Alam
• When FEVER Meets Finance: A Controlled Benchmark Reveals Overlap-Driven Financial Claim VerificationShortSuprabhat Rapolu
• When Words Move Markets: Sentiment and Semantic Complexity in Central Bank Communication and Equity VolatilityLongArush Sachdeva, Anirban Dasgupta
• Where A Small Language Model Helps in Invoice Categorisation, Understood Through Embedding GeometryLongEmma Ceccherini, Daniel John Lawson, Anjulika Salhan

FinEval Shared Task

The FinEval shared task session runs 15:00–15:30. System papers are listed alphabetically by title.

Overview

• FinEval Challenge @ EMNLP 2026 - FinNLP Workshop: Multilingual Financial Language UnderstandingYupeng Cao, Haohang Li, Yuechen Jiang, Yueru He, Yuyan Wang, Ruoyu Xiang, Yangyang Yu

System Papers

• ADML at the FinEval 2026: Version-Aware Retrieval and Grounded Sparse Editing for Multilingual Financial QAGautier Marti, Khaled AlNuaimi
• AhsanBLAH at the FinEval 2026: Grounded Two-Stage Financial QA with LLM-Judge VerificationMuhammad Ahsan Atiq, Mohib Ahmad Butt
• Al-Qahera at the FinEval 2026: GEPA Prompt Evolution for Greek Financial NER and Japanese Implicit Communication RecognitionMohamed Mohamed, Hazem Ajlan
• Daikoku at the FinEval 2026: Multi-Stage Labelling with Forced-Choice AdjudicationKyosuke Morita
• Fin_Agent at FinEval 2026: Task-Aware Methods for Multilingual Financial NLPMohotarema Rashid, Tanzina Hossain, Xiaoying Song, Lingzi Hong
• Finaira at the FinEval 2026: Dense Retrieval and Validated Prompt Engineering for Cross-Lingual Financial Question AnsweringYara Mohamed Nasser, Ashraf Mohamed
• HX-Fin at FinEval 2026: Precision-Anchored Answer Refinement and Multi-Model Evidence Selection for Multilingual Financial Question AnsweringPeng Li, Yinchun Wang
• Jetsons at the FinEval 2026: Minimal-Pair Augmentation and Label-Routed Ensembling for Greek and Japanese Financial NLPParag Pravin Dakle, Sai Krishna Rallabandi, Preethi Raghavan
• LangKG at the FinEval 2026: Answer-Contract Knowledge over Instruction EngineeringShivika Prasanna, Jaeseong Lee, Hui Su, Yourong Xu
• Maddies NLP at the FinEval 2026: Proposition-Aware Tagged Reasoning for Japanese Implicit Communication RecognitionMadhav Kapoor, Aditya Channa, Jastej Singh
• Mircea Andrei Razvan at the FinEval 2026: A Unified Prompting Approach for Multilingual Financial Q&AMircea Andrei Razvan
• PIO at the FinEval 2026: An OpenRouter API Pipeline for Multilingual Financial UnderstandingIsrael Olanrewaju Odeajo
• shamikkundu at the FinEval 2026: Zero-Parameter Confidence Elicitation for Japanese Implicit Communication RecognitionShamik Kundu
• Token Crit at the FinEval 2026: Role-Specialized, Validation-Gated LLM Workflows for JF-ICR and Multilingual Financial Q&AAnke Xu, Wenbo Cao