This poster illustrates how we redesigned the CS1 course for Math undergraduates to be held online but reflecting the face-to-face (F2F) experience as much as possible. We describe the course structure and the strategies we implemented to maintain the benefits of a synchronous experience. We present the positive and negative aspects that emerged from the students' opinion analysis. We highlight what worked, what did not, and what can be improved to strengthen the perception of a F2F experience and mitigate the "presence paradox" we found: although students are enthusiastic about the online format, most would still prefer a F2F course.

In this paper, we propose to exploit the side-tuning framework for multimodal document classification. Side-tuning is a methodology for network adaptation recently introduced to solve some of the problems related to previous approaches. Thanks to this technique it is actually possible to overcome model rigidity and catastrophic forgetting of transfer learning by fine-tuning. The proposed solution uses off-the-shelf deep learning architectures leveraging the side-tuning framework to combine a base model with a tandem of two side networks. We show that side-tuning can be successfully employed also when different data sources are considered, e.g. text and images in document classification. The experimental results show that this approach pushes further the limit for document classification accuracy with respect to the state of the art.