Douglas Tilley in Tokyo at the National Institute of Informatics [Japan]  

Doug Tilley from Cohort 3 has recently come back from Tokyo and writes about his experience here.

Doug Tilley from Cohort 3 has recently come back from Tokyo. He writes:

“That wraps up 6 incredible months in Tokyo at the National Institute of Informatics [Japan] 国立情報学研究所, where I had the privilege of working under Professor Taisuke Kobayashi. While the collaboration continues, my time on the ground in Japan has come to a close, at least for now.

Anyone who knows me knows that I tend to throw myself fully into a range of ambitious projects, and this experience was no exception. I’ve had the chance to work alongside some brilliant minds, early-career researchers, seasoned academics, and industry professionals, united by a shared drive to push boundaries.

I primarily developed a tool for exploring the latent space of Variational AutoEncoders (VAEs), aiming to better interpret robotics competitions like RoboCup. This work integrated a range of different explainable techniques, both old and new, with domain adaption into a functional UI. Our goal is to shed light on the underlying evaluation process, bridging the gap between human intuition and algorithmic scoring.

In parallel to this work, we’ve been developing a novel optimizer tailored for real-world, noisy data, rather than standard benchmark datasets. This is a collaborative effort, and we’re currently running ablation studies with plans to release the results soon. We have also been working on a neuromorphic architecture and a few biologically plausible learning mechanisms, which although at the early stages, have shown some promising results.

One personal highlight has been co-founding the NII International Intern Alumni Network, which is still in its early stages. We hope it will grow into a meaningful group for world-class researchers to connect and collaborate. I’m excited for what lies ahead and for the continued collaborations that will grow from this experience.

I would like to specifically thank the various individuals that have helped me get to this stage and for the internship, Professor Taisuke Kobayashi for his mentorship, Dr Uriel Martinez-Hernandez for his continued support as a primary supervisor, Dr Marina De Vos for her help and support in collaborating, as well as Professor Emmanuel Planas and the Global Liason Team who help to support and run the NII International Internship Program. A huge thanks to all the colleagues, mentors, and friends who made this time so rewarding.”

No alternative text description for this image
Keep Reading

AI Ethics, Risks, and Safety Conference

The second edition of the AI Ethics, Risks, and Safety Conference will be held on Thursday 15th of May, in Bristol.

Read More

Policy Modeling and Reasoning in Sociotechnical Systems

ART-AI's Director of Training & Academic Supervisor Marina De Vos is organising an invitation only seminar at Schloss Dagstuhl.

Read More

Joseph Marvin Imperial at the Annual AAAI Conference on Artificial Intelligence

Cohort 4's Joseph Marvin Imperial has just come back from the Annual AAAI Conference on Artificial Intelligence (AAAI) and writes about his experience.

Read More

Algorithmic Recommendations: What’s the problem? with Silvia Milano

We are pleased to have Silvia Milano, who is a Humboldt Fellow in the Munich Center for Mathematical Philosophy at LMU Munich and a Senior Lecturer in Philosophy at the University of Exeter (UK), join us for this ART-AI seminar entitled 'Algorithmic Recommendations: What’s the problem?' on 17th June 2025.

Read More