Schedule

All times are local Rabat (GMT/UTC +0h)

Time Programmed Session
14:00 - 14:10 Opening and welcome note
14:10 - 14:25 Oral presentation session 1 (best paper)
14:25 - 15:30 Invited Talk: Jose Camacho-Collados - Social Media Analysis in the Language Model Era: An Interdisciplinary Perspective
15:30 - 16:00 Coffee break
16:00 - 16:15 Oral presentation session 2
16:15 - 16:30 Oral presentation session 3
16:30 - 17:30 Poster session and closing remarks

Oral presentation session 1

  • The Impact of Highlighting Subjective Language on Perceived News Trustworthiness. Mohammad Shokri, Vivek Sharma, Emily Klapper, Shweta Jain, Elena Filatova, Sarah Ita Levitan. (remote)

Invited Talk: Jose Camacho-Collados

Social Media Analysis in the Language Model Era: An Interdisciplinary Perspective

Understanding what is going on in social media is not a simple task, even after the significant progress in NLP in recent years. Given the specific nature of social media and the large amount of content generated, efficient and specialised solutions are often necessary.

At the same time, language models, and in particular large generative language models (LLMs) are currently ubiquitous across the NLP research landscape and beyond. However, it isn’t clear how to effectively leverage them to process large amounts of unstructured data, such as the type usually found on social media, especially for non-expert users. There are also additional challenges when processing large amounts of text, such as ensuring transparency in the process and dealing with long context.

In this talk, I will explain my journey when it comes to social media analysis, from proposing efficient solutions for tasks such as sentiment analysis, hate speech detection or emotion detection, to sharing tips and lessons learned from interdisciplinary collaborations. I will try to revisit the importance of high-quality data and reliable evaluation benchmarks in the current LLM landscape, and some recent results suggesting promising areas for future research.

Bio

Jose Camacho-Collados is a UKRI Future Leaders Fellow and Professor at the School of Computer Science of Cardiff University, where he co-founded the Cardiff Natural Language Processing group (Cardiff NLP). Before joining Cardiff University, he completed his PhD in Sapienza University of Rome and was a Google AI PhD Fellow.

Jose has worked in multiple NLP areas with a particular focus on semantics, multilinguality and computational social science with an interdisciplinary perspective. In this area, he has been developing specialised and efficient NLP models for social media applications, such as TweetNLP and related efforts. His work has received several recognitions, including awards at top NLP conferences, or the 2023 AIJ Prominent Paper Award. He is also the co-author of the “Embeddings in Natural Language Processing” book.

Oral presentation session 2

  • Crowd-Based Evaluation of Emotion Preservation in Spanish–Basque Tweet Machine Translation. Nora Aranberri.

Oral presentation session 3

  • Disentangling Emotion Understanding and Generation in Large Language Models. Sadegh Jafari, Els Lefever, Veronique Hoste.

Poster session

In-person poster session:

  • Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance. Serena Tardelli, Lorenzo Alvisi, Lorenzo Cima, Stefano Cresci, Maurizio Tesconi

  • Measuring LLMs’ Sensitivity to Paraphrased Opinion Prompts. Bushra Alhetelah, Irfan Ahmad

  • Appraisal Trajectories in Narratives Reveal Distinct Patterns of Emotion Evocation. Johannes Schäfer, Janne Wagner, Roman Klinger

  • Exploring Subjective Tasks in Farsi: A Survey Analysis and Evaluation of Language Model. Donya Rooein, Flor Miriam Plaza-del-Arco, Debora Nozza, Dirk Hovy

  • Emotion-aware text simplification of user generated content using LLMs. Anastasiia Bezobrazova, Daria Sokova, Constantin Orasan

  • A Position Paper on Toxic Reasoning: Grounding Categories of Toxic Language in Implications and Attitudes. Stefan F. Schouten, Ilia Markov, Piek Vossen

  • Is Sentiment Banana-Shaped? Exploring the Geometry and Portability of Sentiment Concept Vectors. Laurits Lyngbaek, Pascale Feldkamp, Yuri Bizzoni, Kristoffer Nielbo, Kenneth Enevoldsen

  • News Credibility Assessment by LLMs and Humans: Implications for Political Bias. Pia Wenzel Neves, Charlott Jakob, Vera Schmitt

  • Towards Simulating Social Media Users with LLMs: Evaluating the Operational Validity of Conditioned Comment Prediction. Nils Schwager, Simon Münker, Alistair Plum, Achim Rettinger

  • Says Who? Argument Convincingness and Reader Stance Are Correlated with Perceived Author Personality. Sabine Weber, Lynn Greschner, Roman Klinger

  • Multimodal Claim Extraction for Fact-Checking. Joycelyn Teo, RUI CAO, Zhenyun Deng, Zifeng Ding, Michael Sejr Schlichtkrull, Andreas Vlachos

  • A Multi-Aspect Evaluation Framework for Synthetic Data: Case Study on Irony and Sarcasm. Laura Majer, Ana Barić, Florijan Sandalj, Ivan Unković, Bojan Puvača, Jan Šnajder

Remote poster session:

  • Council of LLMs: Evaluating Capability of Large Language Models to Annotate Propaganda. Vivek Sharma, Shweta Jain, Mohammad Shokri, Sarah Ita Levitan, Elena Filatova

  • Quantifying Social Sentiment in Hostels Using A Domain-Specific Transformer Pipeline. Ian W. McMurry

  • Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength. Bhuvanesh Verma, Mounika Marreddy, Alexander Mehler

  • Emotional Lexicons: How Large Language Models Predict Emotional Ratings of Russian Words. Polina V. Iaroshenko, Natalia V Loukachevitch

  • Label-Consistent Data Generation for Aspect-Based Sentiment Analysis Using LLM Agents. Mohammad Hossein Akbari Monfared, Lucie Flek, Akbar Karimi

  • Antisocial Behavior Prediction: A Survey and Practical Guide. Anaïs Ollagnier

  • Real-Time Mitigation of Negative Emotion in Customer Care Calls. Surupendu Gangopadhyay, Mahnoosh Mehrabani

  • A Transformer and Prototype-based Interpretable Model for Contextual Sarcasm Detection. Ximing Wen, Rezvaneh Rezapour