Schedule

Welcome dinner

Social Event

Welcome to GTTI MMSP 2022

Premio di Laurea Magistrale, IEEE SPS Italy Chapter

• Davide Salvi, Politecnico di Milano, "Modal analysis and optimization of the top plate of string instruments through a parametric control of their shape", [video]

• Abrham Shiferaw, Università di Genova, "Learning of Linear Video Prediction Models in a Multi-Modal Framework for Anomaly Detection", [video]

Keynote speech: Yuejie Chi, "Policy Optimization in Reinforcement Learning: A Tale of Preconditioning and Regularization"

Policy optimization, which learns the policy of interest by maximizing the value function via large-scale optimization techniques, lies at the heart of modern reinforcement learning (RL). In addition to value maximization, other practical considerations arise commonly as well, including the need of encouraging exploration, and that of ensuring certain structural properties of the learned policy due to safety, resource and operational constraints. These considerations can often be accounted for by resorting to regularized RL, which augments the target value function with a structure-promoting regularization term, such as Shannon entropy, Tsallis entropy, and log-barrier functions. Focusing on an infinite-horizon discounted Markov decision process, this talk first shows that entropy-regularized natural policy gradient methods converge globally at a linear convergence that is near independent of the dimension of the state-action space, whereas the vanilla softmax policy gradient method may take an exponential time to converge. Next, a generalized policy mirror descent algorithm is proposed to accommodate a general class of convex regularizers beyond Shannon entropy, even when the regularizer lacks strong convexity and smoothness. Time permitting, we will discuss how these ideas can be leveraged to solve zero-sum Markov games. Our results accommodate a wide range of learning rates, and shed light upon the role of regularization in enabling fast convergence in RL.

Coffee Break

Oral Session -
"Biomedical image processing: recent trends and developments", organised by Silvana Dellepiane.

• Vittorio Accomazzi (IMS-Canada) - Steven Borg (Director Microsoft Healthcare), "How DICOMWeb is unlocking the value of medical images"

• Gloria Menegaz (Università Verona), "eXplainable AI in Neuroimaging: state-of-the-art, open issues and perspectives"

• Alberto Signoroni (Università Brescia), "Medical image understanding at the time of pandemics: actions and reactions"

• Silvana Dellepiane (Università Genova), "Graph-based segmentation"

Social Dinner

Ristorante La Grangia - Pian del Sole

19.30: Departure from Chairlift Smith 4 (Seggiovia Smith 4) in Campo Smith.
Be on time because the lift opens for our event only and closes afterwards.

22.30: Return on the chairlift

Show&Tell / Poster Session

Posters

• Marco Olivieri, Mirco Pezzoli, Fabio Antonacci, Augusto Sarti, "A Physics-Informed Neural Network Approach for Nearfield Acoustic Holography"

• Davide Salvi, Clara Borrelli, "Deepfake speech detection through emotion recognition: a semantic approach"

• Paolo Testolina, Francesco Barbato, Umberto Michieli, Marco Giordani, Pietro Zanuttigh, Michele Zorzi, "SVTW: Sensor, View, Time, and Weather -- An Extensive Road-View MultiTask Dataset in CARLA"

• Weiwei Xu, Vincenzo Lipari, Paolo Bestagini, Stefano Tubaro, "Interpolation of Missing Shots via Plug and Play Method with CSGs Trained Deep Denoiser"

• Andrea Maracani, Umberto Michieli, Marco Toldo and Pietro Zanuttigh, "RECALL: Replay-based Continual Learning in Semantic Segmentation"

• Daniele Mari, Elena Camuffo, Simone Milani, "Content-aware compression for Lidar point clouds"

• Antonio Giganti, Daniele Ugo Leonzio, Paolo Bestagini, Luca Cuccovillo, Stefano Tubaro, "Audio Splicing Detection Based on Acquisition Device Traces"

• Donald Shenaj, Francesco Barbato, Umberto Michieli, Pietro Zanuttigh, "Continual Coarse-to-Fine Domain Adaptation in Semantic Segmentation"

• Riccardo Giampiccolo, "Multiphysics Modeling of Audio Systems in the Wave Digital Domain"

• Abrham Shiferaw Alemaw, Giulia Slavic, Lucio Marcenaro, Carlo Regazzoni, "Learning relative distance based multi-agent interaction model from multi-sensorial data"

• Luca Maiano, Irene Amerini, Aris Anagnostopoulos, "Social media platform identification of videos"

• Francesco Paissan, Alberto Ancilotto, Alessio Brutti, Elisabetta Farella, "PhiNets: a scalable backbone for multimedia analytics at the edge"

• Francesca Pistilli, Diego Valsesia, Giulia Fracastoro, Enrico Magli, "Signal compression via neural implicit representations"

Demos

• Niccolò Bisagno, Nicola Conci, "The role of simulation in image and video processing: state of the art, and the use case of crowd analytics"

• Zeno Sambugaro, and Nicola Conci,"Towards single image 3D shapes generation"

• Giulia Slavic, Pablo Marin, Lucio Marcenaro, David Martin, Carlo Regazzoni, "Anomaly Detection using Vehicle Localization from First Person Video Data through a Coupled Dynamic Bayesian Network Model"

• Ilaria Mannocchi, Sara Baldoni, Michele Brizzi, Kamal Lamichhane, Michael Neri, "Automatic emotion classification based on human skeleton features - “I know that feel, bro " "

• Alberto Ancilotto, Francesco Paissan, Elisabetta Farella, "Exploiting PhiNets for face swapping"

• Antonio Montanaro, Diego Valsesia, Enrico Magli, "Exploring the solution space of linear inverse problems with GAN latent geometry"

Coffee Break

Interactive Session with PhD Students, organised by Lucio Marcenaro

Lunch Break

IEEE SPS Italy Chapter Meeting