We propose a more realistic, physics-based color data augmentation - which we call Planckian Jitter.
Provides a video similarity learning approach using self-supervision.
A self-supervised learning method aiming to alleviate the inherent false-negative problem in contrastive learning framework.
Discover one of the breakthrough achievements of the ESFA project, a pioneer project in the field of Artificial Intelligence applied to Climate Prediction. Leveraging Geoskop’s revolutionary AI algorithms, housed in a Docker image, we provide unprecedente...
Novel framework for Playable Video Generation that is trained in a self-supervised manner on a large dataset of unlabelled videos. We employ an encoder-decoder architecture where the predicted action labels act as bottleneck. The network is constrained to...
A new loss function for self-supervised representation learning, which is based on the whitening of the latent space features.
A 4 hour tutorial presented in October 2021 at ACAI 2021 on Multimodal Perception and Interaction with Transformers
A tutorial from the AI in the Industry course from University of Bologna, covering neural network based methods for anomaly detection
Course notes and Zoom lecture recording of a 4 hour tutorial on machine learning with Neural Networks presented at the Advanced Course on Artificial Intelligence, ACAI 2021, in October 2021.
This micro-project surveys tools and data sets for experiments for demonstrating the potential use of transformers for multimodal perception and multimodal interactions.