Invited Speakers
The Future of Personalized Universal Assistant
Abstract
We've moved way beyond the old days of building discovery, recommendation, decision support, and other AI tools using traditional ML and pattern recognition techniques. The future of universal personal assistance for discovery and learning is upon us. How will multimodality image, video, and audio understanding, and reasoning abilities of large foundation models change how we build these systems? I will shed some initial light on this topic by discussing 3 trends: First, the move to a single multimodal large model with reasoning abilities; Second, the fundamental research on personalization and user alignment; Third, the combination of System 1 and System 2 cognitive abilities into a single universal assistant.
Bio
Dr. Ed H. Chi is VP of Research at Google DeepMind, leading machine learning research teams working on large language models (from LaMDA leading to launching Bard/Gemini), and universal assistant agents. With 39 patents and ~200 research articles, he is also known for research on user behavior in web and social media. As the Research Platform Lead, he helped launched Bard/Gemini, a conversational chatbot experiment. His research also delivered significant improvements for YouTube, News, Ads, Google Play Store at Google with >950 product landings and ~$10.4B in annual revenue since 2013.
The Intelligence Feedback Loop: From Biological Inspiration to Augmented Cognition
Abstract
This talk presents an exploration of the intelligence feedback loop—a two-way street where biological inspiration drives AI development, and advanced AI, in turn, amplifies human cognitive capabilities. We begin with a concise history of AI’s evolution, emphasizing the influence of neuroscience on models ranging from classic symbolic agents to modern multimodal large language models (LLMs) that unite multisensory inputs with symbolic reasoning. For augmented cognition, we will discuss computer-use agents that operate in the complex digital world and automate information search and intent execution for human users. We then focus on new biologically-inspired developments, detailing our work on HippoRAG, a long-term memory framework for LLMs inspired by the hippocampal indexing theory.
Bio
Dr. Yu Su is a Distinguished Assistant Professor at the Ohio State University, where he co-directs the NLP group. He has broad interests in artificial intelligence, with a primary interest in the role of language as a vehicle for reasoning and communication. His group is a driving force on the emerging topic of LLM-based language agents, with seminal contributions such as Mind2Web, SeeAct, HippoRAG, LLM-Planner, and MMMU. He is a 2025 Sloan Fellow and has received multiple paper awards from CVPR and ACL.
Practice on OneRec - Unifying Retrieve and Rank with Generative Recommender and Preference Alignment
Abstract
Recently, generative retrieval-based recommendation systems (GRs) have emerged as a promising paradigm by directly generating candidate items in an autoregressive manner. However, most modern recommender systems adopt a retrieve-and-rank strategy, where the generative model functions only as a selector during the retrieval stage. In this talk, we will discuss the key limitations of current recommender systems, Onerec's industrial deployment insights, and the future challenges and opportunities in this evolving field.
Bio
Dr. Shiyao Wang is a researcher in the field of multimodal and recommendation systems. She is currently affiliated with Kuaishou Technology, where she leads innovative research projects that bridge the gap between retrieval, ranking, and generative recommendation systems. Her work includes contributions such as OneRec, which unifies retrieval and ranking with generative recommender and iterative preference alignment, and she continues to drive forward the emerging topics in multimodal and recommendation systems.
