The human cognitive system learns quickly, generalizes easily and robustly to new situations, enabling rapid decision-making even in critical situations. These amazing capabilities are driven by life-long learning on dynamic and multi-modal information coming from the different sensory inputs.

How is the brain able to learn and process such dynamic, multi-modal data given the massive amounts of input streaming from its sensors? Can we teach artificial cognitive systems similar capabilities? Once we do, how do we align artificial intelligence systems better to humans, where alignment spans all aspects from perceptual, cognitive, and emotional aspects all the way up to moral reasoning and high-level decision-making?

In the Cognitive Systems Lab, we address these questions aiming to enhance the understanding of the algorithms employed by the human cognitive system. For this, we make use of cutting-edge methods from machine learning (AI) and mixed reality coupled with neuroimaging and psychophysics experiments.

We currently conduct research in four main areas:

Decision making (combining virtual/mixed reality, neuroscience, and deep learning to investigate decision-making in high-risk, complex scenarios)

Human-centered AI (evaluation, validation, and improvement of algorithms from computer graphics, computer vision, visualization, and deep learning in human-centered applications, alignment)

Social face processing (neuroimaging of face and facial expression processing, computational and experimental investigation of conversational facial expressions)

Multisensory and multimodal object processing (combining vision and touch for object processing, shape representations in the brain, multisensory skills in robotics, mixed reality studies using vision and touch)

NEWS

• June 12th, 2026: New paper accepted in Medical Image Analysis (a top 10 journal in the AI-category) by Daehyun Cho entitled "Establishing the Robustness Metric as a Scalable Proxy for Clinical Relevance in Medical AI Explainability". Everyone wants explainable models - but which measure do we actually use? Can we validate these? Daehyun conducted a whole series of experiments on two popular deep learning tasks based on structural MRI, showing that "Robustness" wins out in terms of independent clinical markers (including radiologist experts) among established XAI metrics. Really awesome work!!

• June 12th, 2026: We are celebrating three graduations: Jiyoon Park (work on decoding anxiety via high-immersive VR / physiology / deep learning - check her profile on LinkedIn) and Jaehyoung Bae (work on affective touch decoding using innovative multimodal experiments - check his profile on LinkedIn) for their PhD theses and Jinhee Jeon (work on improving foundation models for segmentation of transparent objects) for her Master thesis. Couldn't be prouder of everyone's work. 

• May 30th, 2026: Congratulations to Jaehyoung Bae, Hoesung Ryu, and Daehyun Cho on their collaborative paper on decoding emotional communication via touch. A really interesting project, in which we recorded a dataset of how people conveyed different emotions via touch to a receiving target. This will be presented at IEEE RO-MAN 2026. 

• May 30th, 2026: One of our interns (Youhyun Sung), has managed to snag an oral presentation at the Visual Science of Art Conference (VSAC 2026) from her work so far - impressive!

• April 28th, 2026: Congratulations to Cansu Malak on her new paper on fine-grained ethnicity categorization in Perception. This has been a pet-peeve project of mine for a long time - we basically checked whether Koreans can reliably categorize Korean versus Japanese versus Chinese faces. Spoiler: they can't really ^^. Great work!

• April 13th, 2026: New paper by Jaehyoung Bae on comparing active and passive touch for exploration of different materials versus the human hand in IEEE Transactions on Haptics. Check it out - we did get some nice results.

• March 3rd, 2026: Congratulations to Yiyu Chen for her latest publication on decoding of deceit from our challenging EEG dataset. Very nice results! Chen et al., Unmasking Deceptive Decisions in a Two-Player Competitive Game: Integrating ERPs, Spectral Perturbation Signatures, and Interpretable Neural-Network Decoding. Biological Psychology, accepted.

• February 1st, 2026: I am part of the World's Top 2% scientists - interesting ^^: https://topscinet.com/scientist_profile/Wallraven,%20Christian/2001/?stype=career_data

• January 17th, 2026: Jinhee Jeon's paper on a better architecture for segmentation of transparent objects has been accepted to ICASSP - congrats! 

• December 5th, 2025: Congratulations to Hoe Sung Ryu for successfully defending his PhD and to Jaehoon Jang for successfully defending his Masters!! Onwards and forwards ^^

• November 17th, 2025: Congrats to Hoe Sung Ryu for his second publication in a Top Conference this year. His study looked at attention (mis!)alignment in large VLMs and will go to AAAI in Singapore (Mind the Gap: Quantifying and Aligning Human-AI Visual Attention for Accident Anticipation). UPDATE: This was just confirmed to be an oral presentation! Yay!

• October 15, 2025: Jaehoon Jang and his team made 2nd place in Korea in the 7th KB AI challenge - congratulations!!

• September 18, 2025: Paper from Dilara Derya on decoding of facial expressions in the brain accepted.

• September 1, 2025: Two papers and one collaboration paper at ISMAR 2025. Jiyoon Park, Christian Wallraven. Investigating Dynamics of Subjective Anxiety and Behavior due to Personal Space Violations and COVID-19-related Stressors in a Social VR Simulation. Hoe Sung Ryu, Uijong Ju, Christian Wallraven. Seeing What Matters: Attentional (Mis-)Alignment Between Humans and AI in VR-Simulated Prediction of Driving Accidents (invited also into special issue of IEEE TVCG). 

• June 6, 2025: Successful PhD defense by Dilara Derya, and two successful Master defenses by Hyunseung Lee and Cansu Malak - congratulations!

• May 13, 2025: New publications by Jaehyoung Bae in PLoS One, and a publication with Uijong Ju in TRF: Traffic Psychology and Behaviour. Congrats ^^

• April 28, 2025: Podcast interview.

• April 22, 2025: New webpage online on google sites.

• March 1, 2025: Selected as one of the (unfortunately becoming fewer) recipients of funding from NRF on "Advanced Deep Learning Models of Risky Decision Making in Simulated Driving Integrating Behavioral and EEG Data" - excited for the upcoming 3 years about this research!

• October 21, 2024 → October 25, 2024: Three poster presentations at ISMAR2024 by Hoesung Ryu, Jiyoon Park, and Jaehoon Jang, check https://ieeeismar.org/

• October 6, 2024 → October 10, 2024: Paper presentation at MICCAI Workshop by Daehyeon Cho, see https://sites.google.com/view/ml-cdsmiccai2024/home

• September 10, 2024: Publication from Taeho Kang and Yiyu Chen in Journal of Neural Engineering, available at https://doi.org/10.1088/1741-2552/ad788e

• July 8, 2024 → July 14, 2024: Poster presentation in collaboration with Taeho Kang at NeuroErgonomics, see https://neuroergonomics2024.inria.fr/program/

• July 3, 2024 → July 6, 2024: ICPRAI2024 in Jeju - Christian Wallraven is co-chairing this conference on pattern recognition and AI. Sponsored by IAPR and published in LNCS in Springer, check http://www.icprai2024.org