Automated Fact Checking: Task Formulations, Methods and Future Directions
5238Authors
nan
Abstract
AbstractThe recently increased focus on misinformation has stimulated research in fact checking, the task of assessing the truthfulness of a claim. Research in automating this task has been conducted in a variety of disciplines including natural language processing, machine learning, knowledge representation, databases, and journalism. While there has been substantial progress, relevant papers and articles have been published in research communities that are often unaware of each other and use inconsistent terminology, thus impeding understanding and further progress. In this paper we survey automated fact checking research stemming from natural language processing and related disciplines, unifying the task formulations and methodologies across papers and authors. Furthermore, we highlight the use of evidence as an important distinguishing factor among them cutting across task formulations and methods. We conclude with proposing avenues for future NLP research on automated fact checking.
Keywords
Full Content
Proceedings of the 27th International Conference on Computational Linguistics, pages 3346–3359 Santa Fe, New Mexico, USA, August 20-26, 2018. 3346 Automated Fact Checking: Task formulations, methods and future directions James Thorne Department of Computer Science University of Sheffield, UK j.thorne@sheffield.ac.uk Andreas Vlachos Department of Computer Science University of Sheffield, UK a.vlachos@sheffield.ac.uk Abstract The recently increased focus on misinformation has stimulated research in fact checking, the task of assessing the truthfulness of a claim. Research in automating this task has been conducted in a variety of disciplines including natural language processing, machine learning, knowledge repre- sentation, databases, and journalism. While there has been substantial progress, relevant papers and articles have been published in research communities that are often unaware of each other and use inconsistent terminology, thus impeding understanding and further progress. In this pa- per we survey automated fact checking research stemming from natural language processing and related disciplines, unifying the task formulations and methodologies across papers and authors. Furthermore, we highlight the use of evidence as an important distinguishing factor among them cutting across task formulations and methods. We conclude with proposing avenues for future NLP research on automated fact checking. Title and Abstract in Greek Αναγνώριση Ψευδών Ισχυρισμών: Ορισμοί, μέθοδοι και κατευθύνσης για μελλοντική έρευνα Το πρόσφατα αυξημένο ενδιαφέρον για το φαινόμενο της παραπληροφόρησης έχει κινητοποιή- σει ερευνητικές προσπάθειες για μεθόδους αναγνώρισης ψευδών ισχυρισμών. Η έρευνα σε αυτοματοποιημένες μεθόδους για το πρόβλημα αυτό διεξάγεται σε διάφορους επιστημονικούς κλάδους, όπως επεξεργασία φυσικής γλώσσας, μηχανική μάθηση, αναπαράσταση γνώσης, βά- σεις δεδομένων και δημοσιογραφία. Αν και έχει γίνει σημαντική πρόοδος, τα άρθρα σχετικά με το αντικείμενο δημοσιεύονται στα συνέδρια και τα περιοδικά του κάθε κλάδου και χρησι- μοποιούν διαφορετική ορολογία, με συνέπεια να δυσχεραίνεται η περαιτέρω πρόοδος. Στο παρόν άρθρο παρουσιάζουμε μια ανασκόπηση της έρευνας στον αυτόματη αναγνώριση ψευδών ισχυρισμών με κεντρικό άξονα την επεξεργασίας φυσικής γλώσσας, ενοποιώντας ορισμούς και μεθόδους που έχουν προταθεί στη βιβλιογραφία. Μια βασική διάκριση που διαχωρίζει τις ορισμούς και μεθόδους που έχουν προταθεί σε διαφορετικούς κλάδους είναι η χρήση ή μη αποδεικτικών στοιχείων. Η ανασκόπηση στο άρθρο αυτό καταλήγει με κατευθύνσεις για μελλοντική έρευνα στην επεξεργασία φυσικής γλώσσας σχετική με το πρόβλημα αυτό. 1 Introduction The ability to rapidly distribute information on the internet presents a great opportunity for individuals and organizations to disseminate and consume large amounts of data on almost any subject matter. As many of the barriers required to publish information are effectively removed through the advent of social media, it is possible for content written by any individual or organization to reach an audience of hun- dreds of millions of readers (Allcott and Gentzkow, 2017). This enhanced ability of reaching audiences This work is licenced under a Creative Commons Attribution 4.0 International Licence. Licence details: http: //creativecommons.org/licenses/by/4.0/ 3347 can be used to spread both true and false information, and has become an important concern as it is spec- ulated that this has affected decisions such as public votes, especially since recent studies have shown that false information reached greater audiences (Vosoughi et …
Content truncated to 500 words
Labeled Text
AbstractThe recently increased focus on misinformation has stimulated research in fact checking, the task of assessing the truthfulness of a claim. Research in automating this task has been conducted in a variety of disciplines including natural language processing, machine learning, knowledge representation, databases, and journalism. While there has been substantial progress, relevant papers and articles have been published in research communities that are often unaware of each other and use inconsistent terminology, thus impeding understanding and further progress. In this paper we survey automated fact checking research stemming from natural language processing and related disciplines, unifying the task formulations and methodologies across papers and authors. Furthermore, we highlight the use of evidence as an important distinguishing factor among them cutting across task formulations and methods. We conclude with proposing avenues for future NLP research on automated fact checking.
Publication Details
- Paper ID:
- 5238
- Year:
- 2018
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan