no code implementations • 30 Jan 2024 • Fotios Lygerakis, Vedant Dave, Elmar Rueckert
One of the most critical aspects of multimodal Reinforcement Learning (RL) is the effective integration of different observation modalities.
no code implementations • 22 Jan 2024 • Vedant Dave, Fotios Lygerakis, Elmar Rueckert
In evaluating our methodology, we focus on two distinct tasks: material classification and grasping success prediction.
1 code implementation • 6 Sep 2023 • Fotios Lygerakis, Elmar Rueckert
The Variational Autoencoder (VAE) is known to suffer from the phenomenon of \textit{posterior collapse}, where the latent representations generated by the model become independent of the inputs.
no code implementations • 22 Jun 2021 • Debapriya Banerjee, Fotios Lygerakis, Fillia Makedon
Multi-modal sentiment analysis plays an important role for providing better interactive experiences to users.