PASTEL (Prompted weAk Supervision wiTh crEdibility signaLs) is a weakly supervised approach that leverages large language models to extract credibility signals from web content, then further combines them to predict content veracity without using any ground truth labels.

Credibility library

BC4AI：Blockchain Used to Guarantee Credibility of AI Model Evaluations;利用区块链来保证算法模型的真实性

SourceCred instance for Governauts

An application that allows the credibility of business partners to be determined by checking VAT status, publicly available VAT payer data and verifying bank account numbers.

Deep Fact Validation

A list of resources for students, collaborators, and myself.

Content-focused webpage credibility evaluation

Deep Reinforcement Learning (DRL) for the realization of non-player characters (NPC) in a video game such as to have a credible behavior and able to provide a stimulating game challenge for the users.

Build system for the Credibility Coalition website

Provides web credibility models (Likert scale) to assign a trustworthiness score to a given website.

Evaluation Measures for Relevance and Credibility in Ranked Lists

Materials for the September 2017 BITSS Research Transparency and Reproducibility Training (RT2)

accuFact is a tool built to shed light on real news.

a web application to detect fake news