Nikolaos Tsagkas
Nikolaos Tsagkas is a PhD candidate at the School of Informatics, University of Edinburgh, affiliated with the Edinburgh Centre for Robotics. He is known for his research on leveraging pre-trained visual representations for robot learning, specifically for developing the w2VLA model that decouples declarative and procedural knowledge to facilitate zero-shot skill transfer in visuomotor policies. Part of this research was conducted during his internship at the Samsung AI Center in Cambridge, UK.
“the primary contribution of this work is a novel, end-to-end trainable VLA model that, to the best of our knowledge, is the first to achieve zero-shot skill transfer to unseen objects”
Source→“w²VLA restructures information flow inside the model. It sequentially modulates robot proprioceptive states with three signals: visual context (from a frozen VFM), spatial localization ("where" — from VLM attention heatmaps), and skill intent ("what" — from a text embedding).”
Source→“Part of this work was conducted while Nikolaos Tsagkas was conducting an internship at the Samsung AI Center in Cambridge, UK”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.