Brain x Machine Intelligence Lab

Brain x Machine Intelligence Lab @ KAIST

Laboratory for brain and machine intelligence, KAIST

 Featured publications 

Computational neuroscience

  • Stably flexible learning in the prefrontal cortex. Nature Communications, 2025. (link)

  • Few-shot adaptation in the striatum. Nature Communications, 2025. (link; featured in Nature Communications editors’ highlights)

  • In-silico task design for controlling human inference. Cell Report, 2024. (link)

  • Neural computations underlying meta reinforcement learning (version 3). Cell Reports, 2021. (link)

  • Neural computations underlying meta reinforcement learning (version 2). Nature Communications, 2019. (link) (featured in Top 50 Physics articles)

  • High-performance, memory-efficient, and fast reinforcement learning in the brain. Science Robotics 2019. (link)

  • Neural computations mediating human one-shot inference. PLOS Biology 2015. (link; synopsis)

  • Neural computations underlying meta reinforcement learning (version 1). Neuron 2014. (link, cover; preview)

Machine learning

  • Mitigating plasticity loss through architectural design in continual learning. ICML 2026. (link)

  • Stable and scalable deep predictive coding networks with meta prediction errors. ICLR 2026. (link)

  • Neural value alignment, IEEE Transactions on Cybernetics 2026. (link)

  • Self-supervised debiasing using low rank regularization. CVPR 2024. (link)

  • Energy-based cross attention for diffusion models. NeurIPS 2023. (link)

  • Training debiased subnetworks with contrastive weight pruning. CVPR 2023. (link)

  • Reliably fast adversarial training via latent adversarial perturbation. ICCV 2021. (Oral presentation; link)

  • Human uncertainty inference via deterministic ensemble neural networks. AAAI 2021. (link)

  • Controlling model-free learning using model-based learning. IEEE Transactions on Knowledge and Data Engineering 2010. (link)

  • Fisher criterion function for kernel methods. IEEE Transactions on Neural Networks 2010. (link)