CoRNLP

Computational Rhetoric and Natural Language Processing

We develop computational linguistic models for discourse and argumentation and conduct research on assessing the linguistic capabilities of LLMs. 

____________________________________________________________

Our latest activities:

12/24: COLING2025 paper
The paper "Looking at the Unseen: Effective Sampling of Non-Related Propositions for Argument Mining" by Ramon Ruiz-Dolz, Debela Gemechu, Zlata Kikteva and Chris Reed accepted at COLING 2025 - way to go!

10/24: Two papers accepted @ EMNLP2024 workshops
Wassiliki and Ingrid Espinoza were successful with "So, are you a different person today? - Analyzing Bias in Questions during Parole Hearings", which has been accepted to the ICon2024 workshop  (Social Influence in Conversations) and ''Automated Anonymization of Parole Hearing Transcripts'' (Abed El Rahman Itani, Wassiliki and Nette) was accepted to the NLLP (Natural Legal Language Processing) workshop.

09/24: Computational Linguistics Fall School 2024
Two weeks of intense CL input for 35 international students - it's a great crowd!

09/24: Zlata @ SIGDial in Kyoto
Zlata presents our paper 'Question Type Prediction in Natural Debate' at SIGDial 2024! 

08/24: Nette teaching 'NLP: From Language to Information'
Eight days of teaching the most interdisciplinary group I ever had at the Sommerakademie of the Max Weber-Program (Studienstiftung des deutschen Volkes) in Ljubljana!

07/24: ECAI workshop paper accepted
A new strand of research coming out of CoRNLP, namely on ithe automatic identification of misinformation, with collaborators Bruna Paz Schmid and Steve Oswald (Université de Fribourg, Switzerland) and accepted at the ECAI 2024 workshop on 'Countering Disinformation with Artificial Intelligence'. 

07/24: SIGDIAL paper accepted
Looking forward to presenting our paper 'Question Type Prediction in Natural Debate' at SIGDIAL in Kyoto in September. Authors are Zlata Kikteva, Alex Trautsch, Steffen Herbold and Nette Hautli-Janisz. 

07/24: New work on LLM capabilities!
We're happy to announce our latest piece on assessing the capabilities of LLMs, this time on how good LLMs can impersonate politicians and other figures. Authors are Steffen Herbold, Alexander Trautsch, Zlata Kikteva and Nette Hautli-Janisz.  

06/24: PIZ Fachbeirat (Bayerisches Staatsministerium für Unterricht und Kultus)
From September 2024 onwards, Nette will be member of the "Profilschule für Informatik und Zukunftstechnologien" (PIZ) Fachbeirat which promotes computer science and technology in Bavarian secondary schools - way to go!

06/24: Nette gives keynote at ARGAGE 2024
Looking forward to giving one of the keynotes at ARGAGE 2024 in Fribourg, CH, plus a presentation with Thierry Hermann, Université de Neuchâtel.  

05/24: Nette @ OSSA2024 in Windsor, CA
Excited to talk about a paper by Erisa Bytyqi and me on 'The Facts Speak For Themselves': Fallacy Detection with GPT at OSSA 2024 in Canada. Also commenting on Oxana Pimenova's paper on argument continuity. 

05/24: Nette invited to Liverpool Symposium
I'm very happy to give a talk on 'GAI 1 – humans 0: Assessing the structure of automatically generated persuasive essays' at the “GenAI, Language, and Society” symposium organised by Rudi Palmieri and Elena Musi at the University of Liverpool.

05/24: Program out for DELITE at LREC-Coling 2024!
Looking forward to an exciting mix of papers on current work on deliberation technology in Torino at DELITE2024!

05/24: Nette invited to INDIGO Digital Tag
Happy to meet up with industry representatives to talk about 'Large Language Models & natural language' at the INDIGO Digital Tag 2024. 

05/24: Nette in ZEIT Sprachen
Happy to contribute to an article in ZEIT Sprachen on assessing the capabilities of Large Language Models. 

See here for more updates on conference trips, accepted papers etc.

____________________________________________________________________________________________________________________________