Search papers, labs, and topics across Lattice.
This paper presents a retrieval-augmented generation (RAG) system enhanced with reverse image search (RIS) for image-text fact-checking, achieving 3rd place in the AVerImaTeC shared task. The system combines textual retrieval via similarity search, image retrieval via API-accessed RIS, and generation using GPT5.1. The system achieves competitive performance with low cost ($0.013 per fact-check) and is presented as an accessible and reproducible baseline.
Achieve competitive image-text fact checking at just $0.013 per check by combining RAG with reverse image search, using a surprisingly simple and reproducible architecture.
In this paper, we present our 3rd place system in the AVerImaTeC shared task, which combines our last year's retrieval-augmented generation (RAG) pipeline with a reverse image search (RIS) module. Despite its simplicity, our system delivers competitive performance with a single multimodal LLM call per fact-check at just $0.013 on average using GPT5.1 via OpenAI Batch API. Our system is also easy to reproduce and tweak, consisting of only three decoupled modules - a textual retrieval module based on similarity search, an image retrieval module based on API-accessed RIS, and a generation module using GPT5.1 - which is why we suggest it as an accesible starting point for further experimentation. We publish its code and prompts, as well as our vector stores and insights into the scheme's running costs and directions for further improvement.
OpenAI