AI-Cafe: Deep Ensembles for the Prediction of Media Interestingness


The AI-Cafe presents the Speaker: Dr Mihai Gabriel Constantin (Researcher at the AI Multimedia Lab, University Politehnica of Bucharest, Romania)
Description: In the context of the ever-growing quantity of multimedia content from social, news and educational platforms, generating meaningful recommendations, ratings, and filters now requires a more advanced understanding of their impact on the user, such as their subjective perception. Visual interestingness is one of the most important subjective perception concepts and is currently a popular avenue of research in affective computing.

However, given the high degree of complexity of this subject caused by its inherent multimodality and subjectivity, classical single-system deep neural network-based approaches currently show relatively low prediction capabilities when compared with other computer vision concepts. We will therefore present our advances in deploying a set of larger ensembling based architectures that use late fusion for greatly increasing single-system results. This session will present one of the first attempts at using deep neural network as the primary ensembling engine, as well as introduce some novel neural network layers and architectures that are specially designed for ensembling.
CV: Dr Mihai Gabriel Constantin is a researcher at the AI Multimedia Lab, University Politehnica of Bucharest, Romania, and got his PhD at the Faculty of Electronics, Telecommunications, and Information Technology at the same university.  His PhD topic was “Automatic Analysis of the Visual Impact of Multimedia Data”.  He has authored over 30 scientific papers in international conferences and high impact journals, with an emphasis on the prediction of the subjective impact of multimedia items on human viewers and deep ensembles, having a Google Scholar h-index of 11. 

Mihai Gabriel Constantin