​AGI is steered today by large-model approaches powered with similar technologies---the transformer architecture to name only one. The general consensus is that AGI needs to mature more, as an assistive technology and business component, but also as a model of artificial consciousness and hard concepts like strong AI.

​The Tokyo AI team is delighted to welcome speakers and an audience intrigued by what is coming on AGI and beyond. This event gathers thinkers and authors on very different approaches to AI, and some of what it will take to step forward.

​Speakers

• ​Carlos E. Perez (profile)

  • ​Carlos, visiting us from the US, is the co-founder of IntuitionMachine, an AI solution provider, a former IBM engineer, a researcher, and a very active voice on the future with AI. He is the author of several books including The Deep Learning Playbook and the Artificial series on Intuition, Empathy, and the latest volume "Artificial Fluency".
    

• ​Zhan (Cliff) Chen (profile):

  • ​Cliff, a 15-year search engine veteran, originally hailing from Bing, currently serves as the CEO of rinnaKK (link). His research interests span several areas, including the Free Energy Principle, Active Inference, Reasoning, and Collective Intelligence.
    

• ​Olaf Witkowski (profile):

  • ​Olaf is a leading expert in empathic AI and Artificial Life based in Kyoto, Japan. He is the Founding Director of Cross Labs, a research institute in Japan, focusing on the study of intelligence in biological and synthetic systems. He is the President of the Board of Directors of the International Society for Artificial Life, Executive Officer at leading AI company Cross Compass, and Lecturer at the University of Tokyo. His research focuses on a mathematical understanding of intelligence in any substrate, which he uses to design hybrid living systems, empathic human-machine interfaces, open-ended paradigms for collective computing, and theories of connected minds.
    

• ​Kai Arulkumaran (profile):

  • ​Kai is a research team leader at Araya, where he works on brain-controlled robots as part of the JST Moonshot R&D program, and a visiting researcher at Imperial College London. He has previously worked at DeepMind, Microsoft Research, Facebook AI Research, Twitter and NNAISENSE. His research interests are deep learning, reinforcement learning, evolutionary computation and theoretical neuroscience.

​Talks

​Talk #1 - Olaf Witkowski

​Title: Hybrid Intelligences: How to Communicate with Alien Minds

​Abstract: Intelligence can manifest in many different dimensions and grow in many substrates. Its properties are known to us mainly through human cognition, but can extend way beyond our humanly perception and understanding. How then, in interacting with other natural, artificial, or cybernetic beings, can we convey meaning and share values? Here, we explore the asymmetric, and yet acentric – as opposed to human-centric – relationship among diverse intelligences found on and beyond Earth. Our approach suggests thinking of diverse agents, including technological ones, not merely as tools for humans to use and offload their cognitive computation to, but as true partners together with whom humans can better perceive, understand, and influence the world. This perspective invites a future where diverse minds unify to expand their understanding of reality.

​Talk #2 - Carlos E. Perez

​Title: The Curious Notion of Slow Empathic Thinking

​Abstract: Quaternion Process Theory: Rethinking Creativity and Cognition with Carlos E. Perez. In this engaging talk, Carlos E. Perez expands upon Daniel Kahneman's foundational Dual Process Theory to introduce his innovative Quaternion Process Theory. This refined cognitive framework proposes four distinct modes of thinking, enhancing our understanding of both human and technological intelligence. Perez explores the critical role of slow, empathic thinking in fostering creativity, offering a new perspective on problem-solving and innovation. Delve into the theory's origins and its transformative potential for Deep Learning technologies. Join Carlos E. Perez for a concise exploration into the future of cognition and creative processes.

​Talk #3 - Zhan (Cliff) Chen

​Title: The 3rd way towards AGI - Active inference and collective intelligence.

​Abstract: In the talk, we bridge the realms of artificial intelligence, cognitive science, and collective intelligence. Active inference models how our brain perceives the world, updates beliefs, and makes decisions. It is a promising path toward creating more autonomous and adaptive AI agents. Collective intelligence emerges when groups of interacting agents perform beyond the sum of their individual abilities. When cognition, cooperation, and complexity converge, we see possibilities for achieving AGI.

​Talk #4 - Kai Arulkumaran

​Title: Neuro-inspired AI: Reinforcement Learning, Consciousness and Embodiment

​Abstract: What is the nature of intelligence? How does it differ in its biological and artificial forms? These are the questions that I have tried to address from the start of my research career. As such, I have been inspired by neuroscience in my approaches to AI. Firstly, I will talk about how reinforcement learning (RL) is used as a formal definition of intelligence, and how humans use a variety of different RL algorithms to accomplish their goals. Secondly, I will talk about theories of conscious function, and how they might relate to intelligent behaviour. Finally, I will talk about my current research focus on robotics, where we need to solve embodied intelligence.