Blog
Write-ups of work-in-progress, early results and the thinking behind our open problems. The code and data behind every post link back to the problem board.
Robotics
A robot that cannot tell when its own perception is unreliable is a hazard. We argue for a self-trust layer that reasons about contradictions across sensors — and acts on them.
Cybersecurity
Hand-written security rules go stale the moment a deployment changes. We report early lessons from building a security baseline that learns itself — and the poisoning traps we are designing against.
AI
Keyword spotting on a microcontroller used to justify throwing the model away. Our 4.2-bit BERT reaches 94.1% accuracy inside 256 KB of RAM. Here is what the accuracy-vs-bits curve actually looks like.
AI
Differential privacy gives federated learning a formal guarantee — and a brutal trade-off screen. We walk through what DP budgets mean when twelve simulated hospitals train one medical imaging model.
AI Safety
We catalogued the prompt shapes that steer LLM code assistants toward insecure output — and found that the defense that mattered was not the one everyone expected.
Cyber-Physical Security
Between the first anomalous signal and the physical damage there is a window. We want to measure it, score it in real time, and make it a first-class security metric.
Lab Culture
Every problem in our board is published with its research gap, its weaknesses and its unanswered questions. That is deliberate — here is what it buys the lab and our collaborators.
AI Research
ASTRA is our research project on self-learning AI — a model built to remember its experience, judge what worked, and improve behaviour across episodes. Here is the loop that makes it work.
Every article maps to an open problem with a scoped deliverable. Pick one and the lab will help you run the study.
