PhDInfo Seminar: Francesco Marchetti, “The Memory in Deep learning: from Recurrent Neural Network to Memory Network”

Posted Posted in PhDinfo Seminars

(the slides of this and other presentations are available in the PhDinfo restricted area) PhDinfo seminar held on April 16th, 2021 by Francesco Marchetti Abstract. The memory module in Deep Learning architecture is fundamental to solve sequential andtemporal task such as speech recognition, language modelling, sentiment analysis.The Recurrent Neural Networks (RNN), analysing recurrently each element of […]

PhDinfo Seminar: Lorenzo Biotti, “Molecular and Nano-scale Communications”, April 16, 2021

Posted Posted in PhDinfo Seminars

(the slides of this and other presentations are available in the PhDinfo restricted area) PhDinfo seminar held on April 16th, 2021 by Lorenzo Biotti, XXXVI cycle Abstract. In this presentation I’m going to give a quick overview of Molecular Communications.In the first part, I present a recent work, so I want to thank Gian Marco Roselli, […]

PhDinfo Seminar: Pietro Bongini, “Visual Question Answering: an overview”

Posted Posted in PhDinfo Seminars

Abstract: Visual Question Answering (VQA) is an emerging topic which aims at automatically answering questions referred to a specific image. Together with image captioning, VQA is the main point of contact between the two heterogeneous communities of Natural Language Processing (NLP) and Computer Vision. VQA datasets contain different types of questions which require reasoning about […]

PhDinfo Seminar: Alessio Sortino on “An efficient optimization approach for best subset selection in linear regression”

Posted Posted in PhDinfo Seminars

Abstract The problem of selecting the best sparse linear regression model is considered a relevant optimization problem. Usually this problem is solved through a sequence of separate steps, alternating between choosing a subset of features and then finding a best fit regression. In this seminar a new approach, based on a mixed-integer nonlinear optimization problem, […]

PhDinfo Seminars: Marco Papini “Reliability-based prognostics applications”.

Posted Posted in PhDinfo Seminars

Reliability is a fundamental attribute of industrial systems and can be analytically computed using reliability models. Industrial systems are nowadays complex systems, i.e. they are characterized by a large number of components. This characteristic generally makes the analytical analysis unfeasible. Generally, reliability is expressed, under the assumption of constant failure rate, through the Mean Time […]

PhDinfo Seminars: Lorenzo Bongini “Sub-Synchronous Torsional Interaction (SSTI) in Oil and Gas Plant”

Posted Posted in PhDinfo Seminars

My presentation provides theoretical determination and experimental validation of the interaction between Thyristor Variable Frequency Drive (TVFD) and turbine generator units in Oil&Gas plant. The voltage oscillations caused by torsional and speed oscillation of a turbine generator unit become disturbances for the power electronics devices. They react injecting back into the grid current disturbances, that […]

PhDinfo Seminar: Benedetta Picano on “Predicting passenger demand problem for Didi, a Chinese ride-sharing company”

Posted Posted in PhDinfo Seminars

  The work addresses the predicting passenger demand problem for Didi, a Chinese ride-sharing company. The passenger demand forecasting has been pursued by resorting to chaos theory. More in detail, in order to predict a nonlinear time series exploiting its chaotic features, its chaotic behavior has to be verified. In this respect, I’ll explain how […]

PhDinfo internal seminars: Giulio Galvan on “Decomposition methods: ADMM & friends”

Posted Posted in PhDinfo Seminars

Giulio gave us a very nice presentation on his research topic, based on the ADMM (Alternating Direction Method of Multipliers) algorithm. The presentation was nicely divided into 4 parts:1 – Love (decomposition methods)2 – The other man (ADMM)3 – Getting the girl back (ADMM+decomposition, or something similar…)4 – No title (I suggest: daring more difficult […]