Seminars (hosted by BML)

• Miran Lee (Microsoft Research Asia)
  The vision and approaches of Microsoft Research Asia in brain and neuroscience

• Dongsheng Li (Microsoft Research Asia)
  When machine learning meets the brain, Nov 22, 2023. (MSRA x KAIST BCS workshop)

• Dongqi Han (Microsoft Research Asia)
  A variational Bayesian perspective of habits and goals: insights for psychology and AI, Nov 22, 2023. (MSRA x KAIST BCS workshop)

• Doo-sup Choi (Mayo Clinic)
  Computational Neuroscience and Behavior Disorders: Neural Circuits and Artificial Intelligence in Predicting Behaviors, NOV 8, 2023. (BCS colloquium)

• Brenden Lake (NYU)
  Machine learning through the eyes and ears of a child, NOV 4, 2023. (BCS Symposium)

• Catherine A. Hartley (NYU)
  Neural and cognitive mechanisms of developmental change in goal-directed behavior, NOV 4, 2023. (BCS Symposium)

• Takuya Ito (IBM Research)
  Multitask and compositional capacities in human brains and machines, NOV 3, 2023. (BCS Symposium)

• Joel Z Leibo (Google DeepMind)
  A theory of appropriateness with applications to generative artificial intelligence, NOV 3, 2023. (BCS Symposium)

• Taro Toyoizumi (RIKEN)
  Information theoretical approaches to model synaptic plasticity, OCT 4, 2023. (BCS colloquium)

• Jonathan Rubin (University of Pittsburgh)
  Cortico-basal ganglia-thalamic control ensembles shift decision policies to maximize reward rate during learning, SEP 27, 2023. (BCS colloquium)

• Aurelio Cortese (ATR Computational Neuroscience Labs)
  Confidence, abstractions and reinforcement learning in humans and machines, SEP 6, 2023. (BCS Colloquium)

• Andrea Tacchetti (DeepMind)
  The good shepherd: Machine learning for mechanism design, Oct 5, 2022. (BBE/BCE seminar series + CNAI workshop)

• Doo-Sup Choi (Mayo Clinic)
  Neuroscience of breaking the bad habits: molecular and computational approaches focusing on striatum-pallidal circuits, OCT 27, 2021. (BBE seminar series)

• Bruno Averbeck (NIH/NIMH)
  Computational mechanisms and neural systems underlying reinforcement learning, OCT 13, 2021. (BBE/BCE seminar series)

• Ben Seymour (Oxford)
  Looking for Pain in the Brain, April 14, 2021. (BBE/BCE seminar series)

• Andrew Saxe (Oxford)
  Dynamics of learning contextual, controlled and abstract representations in deep neural networks, Nov 5, 2020. (KI AI + CNAI workshop)

• Andrea Tacchetti (DeepMind)
  Learning in multi-agent systems, Nov 5 2020. (KI AI + CNAI workshop)

• Joel Z Leibo (DeepMind)
  Multi-agent reinforcement learning, Dec 3, 2019. (KI AI seminar series)

• Gabriel Kreiman (Harvard Medical School)
  Peeking inside the brain to develop the next generation of AI, OCT, 30, 2019. (KI AI + CNAI workshop).

• Rahul Bhui (Harvard Univeristy)
  Efficient coding in economic judgment, OCT, 30, 2019. (KI AI + CNAI workshop).

• Anil Yaman (Eindhoven University of Technology)
  Evolution of biologically inspired learning in artificial neural networks, Aug 22, 2019. (CNAI seminar)

• Hyojin Park (University of Birmingham)
  Neural oscillatory mechanisms in dynamic information representation during natural audio-visual speech perception, Aug 14, 2019. (CNAI seminar)

• Joel Z Leibo (DeepMind)
  Autocurricula and the emergence of innovation from social interaction, May 16, 2019. (Bio-IT/BBE/BCE seminar series)

• Zeb Kurth-Nelson (DeepMind)
  Distributions from dopamine and factorized replay, MAY 1, 2019. (Bio-IT/BBE/BCE seminar series)

• YoungGyun Park (MIT)
  Toward integrative brain mapping via intact tissue processing and phenotyping techniques, MAY 1, 2019. (BCE seminar series)

• Xavier Boix (MIT)
  Making a science from the computer vision zoo, NOV 15, 2018. (MIR-MSREP seminar series)

• Hiroyuki Nakahara (RIKEN)
  Neural mechanism and computations for social decision-making , OCT 24, 2018. (Bio-IT half-day workshop)

• Ales Leonardis (University of Birmingham)
  Combining vision and physics to explore synergies in scene understanding, AUG 14, 2018. (KI for AI seminar series)

• Daeyeol Lee (Yale University)
  Future of AI: Is the brain a computer?, AUG 1, 2018. (Bio-IT inspiring talk series)

• Minjoon Kouh (Drew)
  Trade-offs in neural computation, June 27, 2018. (Bio-IT inspiring talk series)

• Keun-Ah Cheon (Yonsei University College of Medicine)
  Neural basis of aberrant social communication in autism spectrum disorder, May 9, 2018. (BBE/BCE seminar series)

• Yoonsuck Choe (Texas A&M) (BBE/BCE seminar series)
  Overcoming limitations of deep learning, April 25, 2018. (BBE/BCE seminar series)

• Joel Z Leibo (Google DeepMind)
  The interplay of competition and cooperation in shaping intelligence, MAR 28, 2018. (BBE seminar series)

• Choong-Wan Woo (Institute for Basic Science)
  Pain neuroimaging, DEC 1, 2017. (Computational psychiatry seminar series)

• Ben Seymour (University of Cambridge; CINN/ATR/Osaka Univ.)
  Pain and aversive learning: from computational neuroscience to clinical neuroengineering, NOV 16, 2017. (Computational psychiatry seminar series)

• Benedetto Martinos (University College London)
  Decision uncertainty, OCT 12, 2017. (Computational psychiatry seminar series)

• Rongjun Yu (National University of Singapore)
  The neural basis of decision making under uncertainty, SEP 28, 2017. (BML computational psychiatry seminar series)

• Christopher Summerfield (Oxford/DeepMind)
  Neural and computational mechanisms of human decision-making, SEP 13, 2017.  (Computational psychiatry seminar series)

• Kóczy T. László (Budapest University of Technology and Economics)
  Fuzzy signature, July 3, 2017.

• Erie D. Boorman (University of California at Davis)
  Computational and representational approaches to associative learning, June 21, 2017. (Computational psychiatry seminar series)

• Hyun Kook Lim (Catholic University Saint Vincent Hospital)
  Alzheimer's disease, Mar 16, 2017.

• Seung-Tae Lee (Yonsei University College of Medicine)
  Next-generation sequencing, Mar 24, 2017.

• Minlie Huang (Tsinghua University)
  New Approaches for Representing Text and Knowledge, NOV 22, 2016.

• Mattia Rigotti (IBM TJ Watson)
  High and low dimensional neural responses for learning and implementing context-dependent behavior, NOV 2, 2016. (Neural computation workshop)

• Jinseob Kim (Korea Brain Research Institute)
  Neural codes of visual perception: single cells and neural circuits in the retina, NOV 2, 2016. (Neural computation workshop)

• JeeHang Lee (Yonsei University; University of Bath)
  Normative decision making, OCT 19, 2016.

• Benedetto Martinos (University of Cambridge)
  The construction of confidence and its role in guiding behavior, OCT 5, 2016. (Computational psychiatry workshop)

• Robb Rutledge (University College London)
  A computational and neural model of momentary subjective well-being, OCT 5, 2016. (Computational psychiatry workshop)

• Shinsuke Suzuki (Tohoku University)
  Value computation in the human brain: its basis and contagious nature, OCT 5, 2016. (Computational psychiatry workshop)

• Sukbin Lim (NYU Shanghai)
  Balanced cortical microcircuitry for working memory and revised NMDA hypothesis, OCT 5, 2016. (Computational psychiatry workshop)

• Heyeon Park (Seoul National University Bundang Hospital)
  Multiple effects of stress on reinforcement learning in a changing environment, SEP 22, 2016.

• Yongsek Yoo (Hongik University)
  A computational model of the medial temporal lobe, AUG 11, 2016.

• Demis Hassabis (Google DeepMind - Founder & CEO)
  Artificial Intelligence and the Future, MAR 11, 2016. (Bio-IT seminar series)

Lab workshops

2018 Model-based deep reinforcement learning (PDF flyer)

Date: Mon, FEB 12, 2018 (15:00-18:00)
Venue: #205 (E16-1 YBS Bldg.)

Reinforcement learning + deep learning + Bayesian game theory. This half-day workshop aims to review recent studies about model-based deep reinforcement learning (RL). Model-based RL refers to a class of reinforcement learning algorithms that learn the model of the environment. For example, model-based RL agents are expected to rapidly adapt to the change of the environment structure. It addresses the Bayesian game problem. Imagine you play a Tic Tac Toe, Chess, or GO with the model-based RL agent. It can dominate the game by taking advantage of your game strategy. However, the conventional model-free RL agent (e.g., DQN, SARSA, TD, and etc.) can be fooled by sudden changes of a goal or deliberate changes in your game strategy. This approach offers enormous potential for solving general problems.