Stage-oe-small.jpg

Thema5026

Aus Aifbportal
Wechseln zu:Navigation, Suche



Multi-hop Retrieval and Reasoning with Language Models for Fake News Detection




Informationen zur Arbeit

Abschlussarbeitstyp: Master
Betreuer: Jin Liu
Forschungsgruppe: Web Science
Partner: FZI Forschungszentrum Informatik
Archivierungsnummer: 5026
Abschlussarbeitsstatus: Offen
Beginn: 11. April 2023
Abgabe: unbekannt

Weitere Informationen

The spread of fake news has caused more and more concern in the public. Reliable information retrieval is essential for fact-checking. Multi-hop information retrieval requires the extraction of information from several sources. The reasoning step processes the retrieved information and verifies the corresponding claim. In this thesis, we focus on the HOVER dataset with pre-trained large language models.


Ausschreibung: Download (pdf)