Proceedings of the 3rd Workshop on the Interactions between Analogical Reasoning and Machine Learning (IARML 2024) co-located with the International Joint Conference on Artificial Intelligence (IJCAI 2024)

Overview & Synopsis

Different ML & AI communities have a different views on "analogical reasoning". However, they all share the idea of transferring knowledge from a source domain to a target domain, by leveraging both the similarities and dissimilarities between the two. While early discussions on analogy trace back to Greek philosophy, in particular Aristotle's study of proportions, analogical reasoning has been intensively explored by cognitive sciences since the 1960s as a unique and remarkable capability of human cognition. The researches in artificial intelligence (AI) have since focused on either formalizing this ill-defined notion of "analogy" or attempting to replicate this very human capability. A significant shift has been observed lately with the rise of deep learning, opposing two trends. On the one hand, neuro-symbolic techniques pursue the systematization of analogical reasoning, by combining the strong representational power of symbolic AI with the computational efficiency of neural networks. On the other hand, analogy has been used more as a validation tool than as an end goal, helping to assess the reasoning performances of neural networks. This was the case already with the evaluation of the first embedding techniques like word2vec and, recently, in assessing the capabilities of Large Language Models.The series of workshops Interactions between Analogical Reasoning and Machine Learning (IARML), organized on behalf of the International Joint Conferences on Artificial Intelligence (IJCAI), aims to explore these two trends, to bridge gaps between the ML and the AI communities, and examine the status of analogical reasoning in the light of the current trends in AI and ML. Spefically, it aims to understand how analogical reasoning can be applied in machine learning, and conversely how machine learning techniques can enhance analogical reasoning. Over the last three editions, the workshop has brought together AI researchers at the cross roads of machine learning, cognitive sciences and knowledge representation and reasoning. This third edition of the IARML workshop was hosted by the 33rd International Joint Conference on Artificial Intelligence (IJCAI) and took place on August 4, 2024 in Jeju (South Korea). It welcomed submissions of research papers on all topics at the intersection of analogical reasoning and machine learning. The submissions were subjected to a strict double-blind reviewing process that resulted in the selection of six original contributions and of two plenary keynote talks.