FIRE 2026

Forum for Information Retrieval Evaluation

Indian Statistical Institute, Kolkata

17th-20th December

Regular Papers

  • Aggregation Matters Evaluating the Effect of Grouping Strategy on India-Centric LLM Bias Benchmarks
    Pranav Vachharajani (Amity University, Gurgaon)
  • From English Questions to Hindi Lessons: Compact Concept Anchors for Cross-Lingual Educational Retrieval
    Kapil Gupta (IIITD); Keshav Gupta (IIITD)
  • Generative Diversity Assessment in Ranking
    Nilanjan Sinhababu (Indian Institute of Technology Kharagpur); Abir Roy (Indian Institute of Technology Kharagpur); Rabina Khatun (Dr. B. C. Roy Engineering College Durgapur); Debasis Ganguly (University of Glasgow); Pabitra Mitra (Indian Institute of Technology Kharagpur)
  • Krishbuddhi: Knowledge-Grounded Multi-Hop VQA for Indian Crop Disease Diagnosis
    Deepak Sharma (IIITD); Utathya Aich (Jadavpur University); Rajiv Ratn Shah (IIITD)
  • Retrieval-Augmented Cross-Lingual Transfer for Abusive Content Detection in Indic Languages
    Galiveeti Poornima (S-Vyasa University)
  • Same Models, Opposite Conclusions: Testing the Robustness of Architecture Bias Comparisons
    Koushik Koushik Deb (Indian Institute of Information Technology Kalyani); Imon Mukherjee (Indian Institute of Information Technology Kalyani); Debarshi Kumar Sanyal (Indian Association for the Cultivation of Science)
  • Script Gap in the Judge: Script-Induced Bias in LLM-Generated Relevance Judgments for Indic Retrieval Evaluation
    Soham Biren Katlariwala (University of the People, Pasadena); Nirban Roy (Arizona State University, Tempe); Mahdisa Moradi (Herat University)
  • The Retriever’s Alibi: A Mechanism-Based Taxonomy for Diagnosing Context-Relevance Failures in RAG
    Deepak Rawat (EXL Services)
  • TM2-RAG: An Automated Framework for Temporal Multimodal Medical Retrieval-Augmented Generation
    Arindam Paul (University of Kalyani); Dipankar Das (Jadavpur University)

Resource and Demo Papers

  • LittiChoQA: Literary Texts in Indic Languages Chosen for Question Answering
    Aarya Khandelwal (IIIT Delhi); Ritwik Mishra (Accenture); Rajiv Ratn Shah (IIT Kanpur)
  • MiniClimate: Domain-Adaptive and Task-Adaptive Pretrained Model for Climate-Related Natural Language Processing Tasks
    Rajdeep Bose (NSHM Knowledge Campus Kolkata); Moumita Basu (NSHM Knowledge Campus Kolkata)
  • Multimodal Semantic Risk Engine: Assessing Macroeconomic Stress via Historical Analogue Retrieval
    Srijani Das (NSHM); Suparna Dhar (Bhawanipur Global Campus)
  • Ponni RAG: A Hybrid Retrieval-Augmented Generation System for the Tamil Ponni Archive
    Honika Sankar (Nunnari Labs); Abinaya Suresh (Nunnari Labs); Karthik Chidambaram (DCKAP); Karunakaran S (Madras Christian College); Srija Santhanam (DCKAP); Abinaya Mahendiran (Nunnari Labs)
  • RCID: A Parallel Corpus and Evaluation Framework for Roman-Script Code-Mixed Indic Detoxification
    RANIT BARAM (St. Thomas' College of Engineering & Technology); Nilanjan Mandal (Accenture); Silpi Ghosh (St. Thomas' College of Engineering & Technology); Sayani Bose (St. Thomas' College of Engineering & Technology); Ritesh Prasad (St. Thomas' College of Engineering & Technology)

Perspective Papers

  • Reconstructing a Researcher's Memory Palace
    Prantika Chakraborty (Indian Association for the Cultivation of Science); Subham Adhikary (Indian Association for the Cultivation of Science); Debarshi Kumar Sanyal (Indian Association for the Cultivation of Science)

Statistics:

This year, FIRE 2026 received 69 submissions across three paper categories. The selection process was competitive, with an overall acceptance rate of 21.7%. The category-wise acceptance statistics are as follows:

  • Regular Papers: 9 out of 52 accepted (17.3%).
  • Perspective Papers: 1 out of 10 accepted (10%).
  • Resource & Demo Papers: 5 out of 7 accepted (71.4%).

The submissions underwent an evaluation process, and the final decisions were made by the Program Committee based on the reviews and recommendations.



ACM SIGIR


TCS Research


To be announced soon.