​Max capacity: 60 (we will prioritize your fit to the topic, to allow for maximum value delivered).

​Location: Daimon, Tokyo.

​Topic

​Join us for the next session in our Advanced AI Series organized by the AI in Hardware and Robotics (AHR) subgroup of the Tokyo AI (TAI) community. This month’s event brings together four speakers at the forefront of robotics research to explore innovations in safe human-robot interaction, humanoid locomotion, and the evolution of humanoid robotics design.

​​Our Community

​​​​Tokyo AI (TAI) is a community composed of people based in Tokyo and working with, studying, or investing in AI. We are engineers, product managers, entrepreneurs, academics, and investors intending to build a strong “AI coreˮ in Tokyo. Find more in our overview: https://bit.ly/tai_overview

​Sponsor

​Thanks a lot to our supporters from the Startup Island TAIWAN, especially Chi Ko, for space, food, and drinks!

​​Schedule

​18:00 Doors open

​18:30 TITLE TBD - Ray Tai, CEO of Mighty Net (profile)

​18:45 Towards the Future of Safe Human-Robot Collaboration: Design and Control of Inflatable Robots with Deep Learning - Gangadhara Naga Sai Gubbala (profile)

​19:15 How to Train Your Humanoid - Rohan P Singh (link)

​19:45 History and Future of Tendon-driven Musculoskeletal Humanoids - Kento Kawaharazuka (link)

​20:15 Networking and Food

​21:00 Event ends

​Speakers

​Gangadhara Naga Sai Gubbala (profile)

​Title: Future of Safe Human-Robot Collaboration: Design and Control of Inflatable Robots with Deep Learning

​Abstract:

​Inflatable robots offer a promising solution for safe human-robot interaction, particularly in elderly care and environments requiring close physical collaboration. Their soft, lightweight structure and low inertia make them ideal for tasks that involve physical contact, minimizing the risk of injury. However, predicting and controlling deformations in these flexible robots is a complex challenge. We aim to address the challenges of controlling these robots by integrating deep learning models, enabling them to perform contact-based tasks suitable for domestic assistance—such as wiping surfaces, aiding in daily chores, and adapting to their surroundings. These robots are compact and easy to deploy, making them suitable for a wide range of applications, from healthcare to space exploration.

​Bio:

​I am a Doctor of Engineering candidate at Ogata Laboratory, Waseda University, focusing on AI and robotics, specifically the design and control of inflatable robots using deep learning. Before academia, I worked for two years as a chatbot developer in an IT company in India, where I led the development of customer-facing chatbot solutions, improving user interfaces and experiences. I envision a future where robotics plays a significant role in society, creating safe, collaborative systems that enhance daily life and interactions with technology

​Rohan P Singh (LINK)

​Title: How to Train Your Humanoid

​Abstract:

​Recent advances in deep reinforcement learning (RL) based techniques combined with training in simulation have offered a new approach to developing robust controllers for legged robots. However, the application of such approaches to real hardware has largely been limited to quadrupedal robots with direct-drive actuators and lightweight bipedal robots with low gear-ratio transmission systems. Application to real, life-sized humanoid robots has been less common arguably due to a large sim-to-real gap. In this talk, I will discuss our sim-to-real approach that successfully trains a unified, end-to-end policy in simulation that can be deployed on a real HRP-5P humanoid robot to achieve bipedal locomotion. This talk also covers policies for robust humanoid walking on compliant and uneven terrain. Finally, since the benefit of the humanoid morphology lies in the use of all its limbs, we investigate the use of deep reinforcement learning for multi-contact locomotion and present our preliminary findings.

​Bio:

​Rohan P. Singh is a Postdoctoral researcher at CNRS-AIST Joint Robotics Lab (JRL) in Tsukuba, Japan. He received the Ph.D. degree from the University of Tsukuba in April 2024 and the MS degree in April 2021. He has worked at AIST, Tsukuba as a Research Assistant since 2019. Earlier, he worked as a Robotics Engineer in the same group (then known as the Humanoid Robotics Group) from 2017 to 2019. His research work focuses on developing reinforcement learning-based locomotion controllers for humanoid robots for bipedal and multi-contact locomotion.

​Kento Kawaharazuka (LINK)

​Title: History and Future of Tendon-driven Musculoskeletal Humanoids

​Abstract:

​In this talk, I will focus on tendon-driven musculoskeletal humanoids and introduce the series of humanoids we have developed, such as Kenta, Kotaro, Kojiro, Kenzoh, Kenshiro, Kengoro, and Musashi. Additionally, I will present the Kangaroo robot, the jumping monopedal robot RAMIEL, the lightweight flexible manipulator SAQIEL, and the mobile tendon-driven robot CubiX, all of which utilize the advantages of tendon-driven mechanisms. Tendon-driven mechanisms are currently being incorporated into hands, manipulators, and leg systems to achieve quick, flexible, and delicate movements, and they are gaining popularity. I hope you will fully immerse yourself in their appeal.

​Bio:

​Kento Kawaharazuka is a Project Assistant Professor of JSK Robotics Laboratory in the Department of Mechano-Informatics at the University of Tokyo, Japan. He received his B.E., M.S., and Ph.D. degrees in Mechano-Informatics from the University of Tokyo in 2017, 2019, and 2022, respectively. His research interests include musculoskeletal humanoids that closely mimic human anatomy, wire-driven robots, soft robotics, machine learning-based controls such as imitation learning, reinforcement learning, and predictive model learning, and real-world robot applications of foundation models including large language models and vision-language models.

​Ray Tai (profile)

​Title: TBD

​Abstract: TBD

​Bio: TBD

​Organizers

​Russ Islam: Researcher on condensed and soft matter physics and reinforcement learning at the University of Tokyo, former robotics engineer and IoT entrepreneur. BS/MS in electrical engineering from Stanford.

​Apurv Saha: Researcher at Tokyo Institute of Information Technology, Working on fusing SLAM with vision language models for autonomous navigation of robots. 

​​Stefano Zangiacomi: ISP and camera tuning engineer at Forvia working on designing cameras and image quality tuning for autonomous driving applications. Previously Camera NPI and Optical engineer at Valeo.  MSc in Optical Engineering at Institut d’Optique graduate school and MSc double degree in Quantum Physics at Paris-Saclay University.

​Sponsor

​“Startup Island TAIWAN” is an initiative supported by Taiwan’s National Development Council (NDC) to foster international connections and growth opportunities for Taiwanese startups. Serving as the nation’s flagship startup brand, Startup Island TAIWAN represents Taiwan in global innovation events and works to enhance Taiwan’s presence in overseas markets. Key activities include organizing exhibitions, fostering cross-border collaborations, and connecting Taiwan’s startups with investors and industry partners worldwide.