Collaborating with AI

This workshop, taking place in September at TU Wien, aims to explore emerging forms of human-AI collaboration, emphasizing the integration of Large Language Models with recommendation systems to enhance user experience and decision-making. New challenges and opportunities will be identified.


Source: https://www.tuwien.at/services/medien/fotodownload/gebaeude#c22687

RecSys Workshop @ TU Wien 2024

Date: September 9th and 10th, 2024
Location: TU Wien, Favoritenstrasse 9-11, 1040 Vienna, Austria

Chairs:

Julia Neidhardt (TU Wien)
Francesco Ricci (Free University of Bozen-Bolzano)
Wolfgang Wörndl (TU Munich)

Organized by the Christian Doppler Lab for Recommender Systems, TU Wien


New forms of Human and AI collaborations will emerge, and it is important to understand how this collaboration can evolve. While there is a long-standing discussion about the threats of AI substituting humans, we believe in exploring the different ways to collaborate. The core issue is understanding how this collaboration can be possible and fruitful, while also managing the risks.

Our workshop will cover several key areas, including:

  • The integration of Large Language Models (LLM) with recommendation systems (RSs), extracting item constraints from dialogues in LLM-based conversational systems, and improving diversity and serendipity in recommendations through generative models.
  • Addressing fairness and bias in generative recommender systems, evaluating explanations, and assessing subjective quality dimensions of conversations in these systems.
  • The simulation of RSs, from understanding their long-term impacts to designing sustainable RSs that improve user satisfaction and support sustainable policies, particularly in tourism and natural resources.
  • AI’s potential in predicting group choices to enhance decision-making, aiding single users in mediating group needs, and creating chat environments for mixed-initiative dialogues between humans and AI.
  • The legal implications of algorithmic decision-making in the EU post-AI and Digital Acts.

Format: Each session will feature a 20-30 minute presentation followed by a 1-hour discussion based on pre-shared questions. We will have no more than 4-5 presentations per day, ensuring ample time for discussion. 

Outcome: The aim of the workshop is to develop a coherent list of challenges and research hypotheses.

For more information, please contact us at: ai-workshop2024[ät]ec.tuwien.ac.at