Dear Colleagues,
We are pleased to announce the FIFI Shared Task @FIRE 2026 .
Track Title:
FIFI: Find It Fix It a
Shared Task on Scientific Title Retrieval & Rewriting
Track Website: https://fifi-2026.github.io/FIFI-2026/
To be organized in conjunction with FIRE 2026 (fire.irsi.org.in)
17th-20th December 2026, Kolkata, India
Large Language Models (LLMs) are increasingly used to rewrite, simplify, and adapt scientific content for different audiences. While these transformations improve accessibility and readability, they also introduce challenges related to content attribution, provenance tracking, and source recovery.
The FIFI@FIRE 2026 shared task aims to develop systems capable of identifying the original research paper title from multiple LLM-generated reformulations.
The dataset contains original scientific paper titles along with three stylistic variants:
Participants are required to retrieve or reconstruct the original title corresponding to a rewritten variant, addressing a novel challenge at the intersection of Information Retrieval, Natural Language Processing, and Generative AI.
Please register here:
https://fifi-2026.github.io/FIFI-2026/
We warmly invite researchers, students, and practitioners from academia and industry to participate and contribute to this emerging research area.