Moderator
Aude Billard, EPFL
Panelists/Speakers
Renaud Detry, KU Leuven
Greg Dudek, McGill University
Nadia Figueroa, Univ. Pennsylvania
Dongheui Lee, TUM
Shigeki Sugano, Waseda University
Kunpeng Yao, University of Leeds
Description:
As the robotics community faces an unprecedented acceleration in publication volume, staying at the forefront of research has become a monumental challenge. As an example, in 2025 alone, IEEE published more than 50 thousands on robotics and automation, and IEEE publications represent only a fraction of the total research available online. Reading even one subfield fully can take weeks of full-time work. That leaves little room to explore beyond one’s specialty, even though major breakthroughs often come from cross-disciplinary idea.
To make things worse, the vast majority of researchers search for papers using Google Scholar, IEEEXplore, or simply via Google, arXiv, and social media. All these search methods put on par papers published in any venue, irrespective of the venue quality, and including papers that have not undergone a peer-review process. Important contributions are hence more and more likely to go unnoticed. It has also become harder for young researchers to grasp the literature, increasing the risk of revisiting already solved problems.
This panel will debate how to address this issue. The panelists will present different solutions, ranging from reducing the volume of publications and modifying paper formats to emphasize novelty and significance, to developing new search engines that rank papers according to the quality of the journal or conference, or based on peer-review evaluation scores. Others will discuss the role that LLMs and other AI tools can play in reviewing the literature, and present a case study comparing human versus LLM-generated review of the Learning from demonstration literature. The audience will be invited to participate in the debate through live polling and rank your question tool.