Abschlussarbeit: Deep Learning Methods for Fake News Detection (FEVER international competition)

Moderator: Aktive Fachschaft

Forumsregeln
Auch ohne Registrierung können Beiträge in diesem Unterforum geschrieben werden.
andre_ha
Neuling
Neuling
Beiträge: 1
Registriert: 10. Apr 2018 10:27

Abschlussarbeit: Deep Learning Methods for Fake News Detection (FEVER international competition)

Beitrag von andre_ha » 10. Apr 2018 10:30

Motivation:
The spread of Fake News discredits reliable news networks, influences public opinion and
threatens the democratic order. Recently, there have been a number of initiatives to foster the
development of AI technology to help solve the fake news problem. The Fake News Challenge
was launched to boost the development of fake news detection systems. In the competition, the
UKP team Athene secured the second place out of 50 participating groups. In this project,
methods for the upcoming FEVER shared task shall be implemented. This competition is
concerned with the validation of controversial claim. For this purpose, deep learning approaches
for evidence identification, automated reasoning, and verdict prediction need to be developed.

Task description:
Development of deep learning methods for evidence identification
Implementation of algorithms for reasoning with evidence
Development of deep learning methods for the prediction of a verdict for a claim

Contact: Dr. Andreas Hanselowski (hanselowski@ukp.informatik.tu-darmstadt.de)
Original posting: https://public.ukp.informatik.tu-darmst ... ng_AnH.pdf
Competition website: https://sheffieldnlp.github.io/fever/task.html

Zurück zu „Abschlussarbeiten“