Fact Checking 🎸 Rocks!

Fact checking baseline combining dense retrieval and textual entailment

Last updated on Jul 23, 2023 Personal projects
Demo! (🤗 Spaces) Code Berlin Buzzwords 23 Presentation Haystack Project Highlight Presentation

This project aims to show that a naive and simple baseline for fact checking can be built by combining dense retrieval and a textual entailment task. In a nutshell, the flow is as follows:

  • the user enters a factual statement
  • the relevant passages are retrieved from the knowledge base using dense retrieval
  • the system computes the text entailment between each relevant passage and the statement, using a Natural Language Inference model
  • the entailment scores are aggregated to produce a summary score.

I also added the feature “Explain using a Large Language Model”, which is based on prompting google/flan-t5-large!

fact checking dense passage retrieval textual entailment natural language inference transformers Haystack Faiss Streamlit large language models
Stefano Fiorucci
Stefano Fiorucci
NLP Engineer, Craftsman and Explorer 🧭 | Contributing to Haystack, the NLP/LLM Framework 🏗️