We are experiencing a paradigm shift to learned nonlinear models in signal-image-video processing. There are no theoretical bounds on the performance limits of learned models, it is not possible to tell how much the results can be further improved. In order to fill this gap in the state of the art, this project aims following.

Our preliminary results demonstrate that even though these tasks are very intuitive for humans, the implemented baselines could not cope with the underlying challenges.

Recent paper by Zana Buçinca, Yücel Yemez, Engin Erzin, and Metin Sezgin: "AffectON: Incorporating Affect Into Dialog Generation" has been accepted at IEEE Transactions on Affective Computing.

Recent paper "Cyberphysical Blockchain-Enabled Peer-to-Peer Energy Trading" is published as part of the Cover Feature of IEEE Computer. Three blockchain-based energy-trading models are proposed to overcome the technical challenges and market barriers as well as enhance the adoption of this disruptive technology.

A new single input multiple output deep temporal regression network (DTRN) to detect the vocal tract (VT) contour and the separation boundary between different articulators.

We're looking for highly motivated graduate students for our project "The Road Less Travelled: One-Shot Learning of Rare Events in Autonomous Driving", also in collaboration with Joao Henriques from the University of Oxford and Luca Bertinetto from the autonomous driving startup Five.

We're looking for talented and motivated post-doctoral researchers in multiple areas of AI and ML. We provide a competitive benefits package, please check our page for details and spread the word!

Assistant Professor Didem Unat, Department of Computer Engineering, is awarded with the 2019-2020 College of Engineering Outstanding Faculty Award.

We introduce a reward learning framework that extracts dense rewards from learned perceptual goals in a robotic skill learning from demonstration setting. We show the efficacy of the learned rewards on various Policy Search methods both in simulation and real robot.

Dr. Fatma Guney has received the Newton Advanced Fellowships, from the British Royal Society for the project titled "The Road Less Travelled: One-Shot Learning of Rare Events in Autonomous Driving" with a budget of £111K for 3 years.

Dr. Didem Unat has received the Newton Advanced Fellowships, from the British Royal Society for the project titled "Diagnostic Tools for Communication Pathologies in Parallel Architectures" with a budget of £111K for 3 years. She will be collaborating with Prof. Paul Kelly from Imperial College-London in this project.

Given recent advances in learned video prediction, we investigate whether a simple video codec using a pre-trained deep model for next frame prediction based on previously encoded/decoded frames without sending any motion side information can compete with standard video codecs based on block-motion compensation. Frame differences given learned frame predictions are encoded by a standard still-image (intra) codec