Ricerc@Sapienza

A novel deep learning approach in proposed in this paper for multivariate prediction of energy time series. It is developed by using Convolutional Neural Network and Long Short-Term Memory models, in such a way that several correlated time series can be joined and filtered together considering the long term dependencies on the whole information. The learning scheme can be viewed as a stacked deep neural network where one or more layers are superposed, feeding their output in the sequent layer's input.

In smart grids and microgrids, time series prediction is a fundamental tool for enabling intelligent energy resource management and advanced interactions between heterogeneous agents. In this work, we propose a solution to the energy forecasting problem based on two machine learning techniques: Convolutional Neural Network and Long Short-Term Memory Network. These techniques are combined with a new embedding format to appropriately feed the time series to the stacked network architecture.

We propose a deep learning approach for multivariate forecasting of energy time series. It is developed by using Long Short-Term Memory deep neural networks so that different related time series, incorporating information of longterm dependencies, can be joined together as a multidimensional input of the deep neural network. The learning scheme can be represented as a stacked LSTM network in which one or more layers are cascaded, feeding their output to the input of the sequent layer.

Imaging Cherenkov detectors are largely used for particle identification (PID) in nuclear and particle physics experiments, where developing fast reconstruction algorithms is becoming of paramount importance to allow for near real time calibration and data quality control, as well as to speed up offline analysis of large amount of data. In this paper we present DeepRICH, a novel deep learning algorithm for fast reconstruction which can be applied to different imaging Cherenkov detectors.

Palmprint verification is one of the most significant and popular approaches for personal authentication due to its high accuracy and efficiency. Using deep region of interest (ROI) and feature extraction models for palmprint verification, a novel approach is proposed where convolutional neural networks (CNNs) along with transfer learning are exploited. The extracted palmprint ROIs are fed to the final verification system, which is composed of two modules. These modules are (i) a pre-trained CNN architecture as a feature extractor and (ii) a machine learning classifier.

Deep neural networks are generally designed as a stack of differentiable layers, in which a prediction is obtained only after running the full stack. Recently, some contributions have proposed techniques to endow the networks with early exits, allowing to obtain predictions at intermediate points of the stack. These multi-output networks have a number of advantages, including (i) significant reductions of the inference time, (ii) reduced tendency to overfitting and vanishing gradients, and (iii) capability of being distributed over multi-tier computation platforms.

Convolutional Neural Networks (CNNs) have been widely used in the field of audio recognition and classification, since they often provide positive results. Motivated by the success of this kind of approach and the lack of practical methodologies for the monitoring of construction sites by using audio data, we developed an application for the classification of different types and brands of construction vehicles and tools, which operates on the emitted audio through a stack of convolutional layers.

In this paper, we first describe the basics of the field of cancer diagnosis, which includes steps of cancer diagnosis followed by the typical classification methods used by doctors, providing a historical idea of cancer classification techniques to the readers. These methods include Asymmetry, Border, Color and Diameter (ABCD) method, seven-point detection method, Menzies method, and pattern analysis. They are used regularly by doctors for cancer diagnosis, although they are not considered very efficient for obtaining better performance.

Brain tumors are considered to be one of the most lethal types of tumor. Accurate segmentation of brain MRI is an important task for the analysis of neurological diseases. The mortality rate of brain tumors is increasing according to World Health Organization. Detection at early stages of brain tumors can increase the expectation of the patients’ survival. Concerning artificial intelligence approaches for clinical diagnosis of brain tumors, there is an increasing interest in segmentation approaches based on deep learning because of its ability of self-learning over large amounts of data.

Gliomas are considered as the most aggressive and commonly found type among brain tumors. This leads to the shortage of lives of oncological patients. These tumors are mostly by magnetic resonance imaging (MRI) from which the segmentation becomes a big problem because of the large structural and spatial variability. In this study, we propose a 2D-UNET model based on convolutional neural networks (CNN). The model is trained, validated and tested on BRATS 2019 dataset. The average dice coefficient achieved is 0.9694.