associated with 21st International Conference on Principles of Knowledge Representation and Reasoning (KR 2024)
We are pleased to announce that our proceedings are now available here. Thank you for your continued interest and support.
The aim of this workshop is to bring together researchers and practitioners working on the intersections of language models, knowledge representation, and reasoning, particularly, but not exclusively, in medical, law, and science domains. We encourage submissions that discuss novel techniques, approaches, and innovative ideas related to this topic. Topics of interest include, but are not limited to, the following:
Contributions may be regular papers (up to 9 pages) or short/position papers (up to 5 pages), including abstract, figures, and appendices (if any) but excluding references and acknowledgements. Submissions should follow the KR 2024 formatting guidelines and be submitted through the submission page. Each submission will be reviewed by at least two program committee members.
Reasoning is an essential component of human intelligence as it plays a fundamental role in our ability to think critically, support responsible decisions, and solve challenging problems. Traditionally, AI has addressed reasoning in the context of logic-based representations of knowledge. However, the recent leap forward in natural language processing, with the emergence of language models based on transformers, is hinting at the possibility that these models exhibit reasoning abilities, particularly as they grow in size and are trained on more data. Despite ongoing discussions about what reasoning is in language models, it is still not easy to pin down to what extent these models are actually capable of reasoning.
The goal of this workshop is to create a platform for researchers from different disciplines and/or AI perspectives, to explore approaches and techniques with the aim to reconcile reasoning between language models using transformers and using logic-based representations. The specific objectives include analyzing the reasoning abilities of language models measured alongside KR methods, injecting KR-style reasoning abilities into language models (including by neuro-symbolic means), and formalizing the kind of reasoning language models carry out. This exploration aims to uncover how language models can effectively integrate and leverage knowledge and reasoning with it, thus improving their application and utility in areas where precision and reliability are a key requirement.
Registration
Morning Tea
Lunch Break
Afternoon Tea
Title: From Words to the Real World: Shared Knowledge, Grounding and Context for GenAI-based Dialogue Systems
Keywords: Generative AI, large language models, knowledge graphs, grounding, mutual understanding, dialogue system
Generative AI has radically changed the landscape in the fields of language and interaction research and the AI-powered applications have amazed the world by their fluent language capability and apparent usefulness for various tasks such as translation, summarization, question-answering etc. At the same time, issues related to their fabrication of facts, confabulation, and detachment of the real-world situations, have also raised questions of their limits and suitability for practical applications where reliability and trustworthyness are important. On the other hand, Knowledge-Graphs (KSs) with their long history in data science, semantic networks, and graph-based knowledge structures offer a powerful tool to represent world knowledge and reason about facts and consequences. It is thus natural to try to integrate KGs and LLMs, so as to support both fluent conversation and reliable task performance when building reliable applications. In this talk I will focus on the use of GenAI models in interactive tasks and discuss challenges and opportunities which have been brought forward to support trustworthyness and fluent conversation between AI systems and humans. In particular, I will emphasise the importance of mutual understanding in dialogue interactions, and explore grounding, cooperation and building of shared context based on real world facts and the reasoning capability of KGs. Using examples from my recent projects, I will argue that creating responsible AI for a safer and more secure world requires understanding of the capabilities and limitations of LLMs and ensuring they can converse with humans in a natural and reliable manner.
Biography: Kristiina Jokinen is Senior Researcher at AI Research Center (AIRC) at National Institute of Advanced Industrial Science and Technology (AIST) in Tokyo Waterfront, and Adjunct Professor at University of Helsinki. She is a member of the pan-European AI network of excellence ELLIS, Advisory Board for Japanese AIE (AI in Engineering) Programme, and Steering Committee for International dialogue workshop series IWSDS. She has led numerous national and international research projects, most recently collaborating in the EU-Japan project e-VITA. Her research concerns human-robot interaction, (Gen)AI-based dialogue modelling and multimodal communication, and she has published widely on these topics. She developed Constructive Dialogue Model as a general framework for interaction, and together with Graham Wilcock she developed the Wikipedia-based robot dialogue system WikiTalk, which won the Special Recognition for Best Robot Design (Software Category) at the International Conference of Social Robotics in 2017. She received her first degree at University of Helsinki, and her PhD from UMIST, Manchester. She was awarded a JSPS Fellowship for PostDoc research at NAIST (Nara Institute of Science and Technology), and was Invited Researcher at ATR Research Labs in Kyoto, and Visiting Professor at Doshisha University.