Progetto e implementazione di un decodificatore ldpc su architettura gpu

In accordo con la filosofia della Software Defined Radio è stato progettato un decoder LDPC software che utilizza una GPU per ottenere prestazioni migliori. Il lavoro, che comprende anche l'encoder e un simulatore di canale AWGN, può essere utilizzato sia per eseguire simulazioni che per elaborare dati in real time. Come caso di studio si sono considerati i codici LDPC dello standard DVB-S2.

Polar Codes for error correction: Analysis and Decoding Algorithms

I Polar Codes sono la prima classe di codici a correzione d’errore di cui è stato dimostrato il raggiungimento della capacità per ogni canale simmetrico, discreto e senza memoria, grazie ad un nuovo metodo introdotto recentemente, chiamato ”Channel Polarization”. In questa tesi verranno descritti in dettaglio i principali algoritmi di codifica e decodifica. In particolare verranno confrontate le prestazioni dei simulatori sviluppati per il ”Successive Cancellation Decoder” e per il ”Successive Cancellation List Decoder” rispetto ai risultati riportati in letteratura. Al fine di migliorare la distanza minima e di conseguenza le prestazioni, utilizzeremo uno schema concatenato con il polar code come codice interno ed un CRC come codice esterno. Proporremo inoltre una nuova tecnica per analizzare la channel polarization nel caso di trasmissione su canale AWGN che risulta il modello statistico più appropriato per le comunicazioni satellitari e nelle applicazioni deep space. In aggiunta, investigheremo l’importanza di una accurata approssimazione delle funzioni di polarizzazione.

prism: rinfrangere distribuzioni di probabilità in riassunti astrattivi mediante strategie di decoding

Negli ultimi quattro anni la summarization astrattiva è stata protagonista di una evoluzione senza precedenti dettata da nuovi language model neurali, architetture transformer-based, elevati spazi dimensionali, ampi dataset e innovativi task di pre-training. In questo contesto, le strategie di decoding convertono le distribuzioni di probabilità predette da un modello in un testo artificiale, il quale viene composto in modo auto regressivo. Nonostante il loro cruciale impatto sulla qualità dei riassunti inferiti, il ruolo delle strategie di decoding è frequentemente trascurato e sottovalutato. Di fronte all'elevato numero di tecniche e iperparametri, i ricercatori necessitano di operare scelte consapevoli per ottenere risultati più affini agli obiettivi di generazione. Questa tesi propone il primo studio altamente comprensivo sull'efficacia ed efficienza delle strategie di decoding in task di short, long e multi-document abstractive summarization. Diversamente dalle pubblicazioni disponibili in letteratura, la valutazione quantitativa comprende 5 metriche automatiche, analisi temporali e carbon footprint. I risultati ottenuti dimostrano come non vi sia una strategia di decoding dominante, ma come ciascuna possieda delle caratteristiche adatte a task e dataset specifici. I contributi proposti hanno l'obiettivo di neutralizzare il gap di conoscenza attuale e stimolare lo sviluppo di nuove tecniche di decoding.

Packet erasure correcting codes for wireless sensor networks: implementation and field trial measurements

This thesis regards the Wireless Sensor Network (WSN), as one of the most important technologies for the twenty-first century and the implementation of different packet correcting erasure codes to cope with the ”bursty” nature of the transmission channel and the possibility of packet losses during the transmission. The limited battery capacity of each sensor node makes the minimization of the power consumption one of the primary concerns in WSN. Considering also the fact that in each sensor node the communication is considerably more expensive than computation, this motivates the core idea to invest computation within the network whenever possible to safe on communication costs. The goal of the research was to evaluate a parameter, for example the Packet Erasure Ratio (PER), that permit to verify the functionality and the behavior of the created network, validate the theoretical expectations and evaluate the convenience of introducing the recovery packet techniques using different types of packet erasure codes in different types of networks. Thus, considering all the constrains of energy consumption in WSN, the topic of this thesis is to try to minimize it by introducing encoding/decoding algorithms in the transmission chain in order to prevent the retransmission of the erased packets through the Packet Erasure Channel and save the energy used for each retransmitted packet. In this way it is possible extend the lifetime of entire network.

Near-Optimum Coding for q-ary Symmetric Channel with Application to Space Communications

Questa Tesi aspira a mostrare un codice a livello di pacchetto, che abbia performance molto vicine a quello ottimo, per progetti di comunicazioni Satellitari. L’altro scopo di questa Tesi è quello di capire se rimane ancora molto più difficile maneggiare direttamente gli errori piuttosto che le erasures. Le applicazioni per comunicazioni satellitari ora come ora usano tutte packet erasure coding per codificare e decodificare l’informazione. La struttura dell’erasure decoding è molto semplice, perché abbiamo solamente bisogno di un Cyclic Redundancy Check (CRC) per realizzarla. Il problema nasce quando abbiamo pacchetti di dimensioni medie o piccole (per esempio più piccole di 100 bits) perché in queste situazioni il costo del CRC risulta essere troppo dispendioso. La soluzione la possiamo trovare utilizzando il Vector Symbol Decoding (VSD) per raggiungere le stesse performance degli erasure codes, ma senza la necessità di usare il CRC. Per prima cosa viene fatta una breve introduzione su come è nata e su come si è evoluta la codifica a livello di pacchetto. In seguito è stato introdotto il canale q-ary Symmetric Channel (qSC), con sia la derivazione della sua capacità che quella del suo Random Coding Bound (RCB). VSD è stato poi proposto con la speranza di superare in prestazioni il Verification Based Decoding (VBD) su il canale qSC. Infine, le effettive performance del VSD sono state stimate via simulazioni numeriche. I possibili miglioramenti delle performance, per quanto riguarda il VBD sono state discusse, come anche le possibili applicazioni future. Inoltre abbiamo anche risposto alla domande se è ancora così tanto più difficile maneggiare gli errori piuttosto che le erasure.

Coded slotted aloha with multipacket reception over block fading channels (for satellite networks)

Random access (RA) protocols are normally used in a satellite networks for initial terminal access and are particularly effective since no coordination is required. On the other hand, contention resolution diversity slotted Aloha (CRDSA), irregular repetition slotted Aloha (IRSA) and coded slotted Aloha (CSA) has shown to be more efficient than classic RA schemes as slotted Aloha, and can be exploited also when short packets transmissions are done over a shared medium. In particular, they relies on burst repetition and on successive interference cancellation (SIC) applied at the receiver. The SIC process can be well described using a bipartite graph representation and exploiting tools used for analyze iterative decoding. The scope of my Master Thesis has been to described the performance of such RA protocols when the Rayleigh fading is taken into account. In this context, each user has the ability to correctly decode a packet also in presence of collision and when SIC is considered this may result in multi-packet reception. Analysis of the SIC procedure under Rayleigh fading has been analytically derived for the asymptotic case (infinite frame length), helping the analysis of both throughput and packet loss rates. An upper bound of the achievable performance has been analytically obtained. It can be show that in particular channel conditions the throughput of the system can be greater than one packets per slot which is the theoretical limit of the Collision Channel case.

L'influenza della preferenza manuale nell'interpretazione di conferenza: un contributo sperimentale

This study is based on a former student’s work, aimed at examining the influence of handedness on conference interpreting. In simultaneous interpreting (IS) both cerebral hemispheres participate in the decoding of the incoming message and in the activation of the motor functions for the production of the output signal. In right-handers language functions are mainly located in the left hemisphere, while left-handers have a more symmetrical representation of language functions. Given that with the development of interpreting skills and a long work experience the interpreters’ brain becomes less lateralized for language functions, in an initial phase left-handers may be «neurobiologically better suited for interpreting tasks» (Gran and Fabbro 1988: 37). To test this hypothesis, 9 students (5 right-handers and 4 left-handers) participated in a dual test of simultaneous and consecutive interpretation (CI) from English into Italian. The subjects were asked to interpret one text with their preferred ear and the other with the non-preferred one, since according neuropsychology aural symmetry reflects cerebral symmetry. The aim of this study was to analyze:1) the differences between the number of errors in consecutive and simultaneous interpretation with the preferred and non-preferred ear; 2) the differences in performance (in terms of number of errors) between right-handed and left-handed, both with the preferred and non-preferred ear; 3) the most frequent types of errors in right and left-handers; 4) the influence of the degree of handedness on interpreting quality. The students’ performances were analyzed in terms of errors of meaning, errors of numbers, omissions of text, omissions of numbers, inaccuracies, errors of nexus, and unfinished sentences. The results showed that: 1) in SI subjects committed fewer errors interpreting with the preferred ear, whereas in CI a slight advantage of the non-preferred ear was observed. Moreover, in CI, right-handers committed fewer mistakes with the non-preferred ear than with the preferred one. 2) The total performance of left-handers proved to be better than that of right-handers. 3) In SI left-handers committed fewer errors of meaning and fewer errors of number than right-handers, whereas in CI left-handers committed fewer errors of meaning and more errors of number than right-handers 4) As the degree of left-handedness increases, the number of errors committed also increases. Moreover, there is a statistically significant left-ear advantage for right-handers and a right-ear one for left-handers. Finally, those who interpreted with their right ear committed fewer errors of number than those who have used their left ear or both ears.

"Romanzo criminale - la serie". Analisi del sottotitolaggio spagnolo amatoriale.

This dissertation focuses on the phenomenon of amateur subtitling, known as fansubbing. Although this phenomenon began in the late ‘80s, in recent years amateur subtitling has spread worldwide, thanks to both Internet and fan communities, also known as fandoms. At first, amateur subtitling was mainly centred on the translation of Japanese cartoons, but nowadays fandoms also tend to subtitle other kinds of audiovisual products, such as American TV series. Thanks to fansubbing, which is created by fans for other fans, fandoms claim that they would prefer to have subtitled rather than dubbed versions of audiovisual products, which is the norm in Italy and Spain. The dissertation provides a linguistic analysis of the fansubbing in Spanish of the Italian TV series Romanzo Criminale. The purpose of this dissertation is to analyse fansubbing from the linguistic point of view, as well as from the point of view of the translation. Furthermore, it aims to evaluate to what extent this translation can be compared to professional subtitling. The first chapter offers an introduction to the TV series and provides an overview of the main events and characters. The second chapter deals with an analysis of the strategies that fansubbers use to translate cultural elements from Italian into Spanish. The third chapter focuses on linguistic mistakes due to calques and linguistic interference between Italian and Spanish. The fourth chapter provides an analysis of some translation errors which occurred during the decoding of the original text. The aim is to understand if this kind of mistake might jeopardize the comprehension of the original message.

Predicting corticospinal excitability from pre-TMS EEG using convolutional neural networks

The amplitude of motor evoked potentials (MEPs) elicited by transcranial magnetic stimulation (TMS) of the primary motor cortex (M1) shows a large variability from trial to trial, although MEPs are evoked by the same repeated stimulus. A multitude of factors is believed to influence MEP amplitudes, such as cortical, spinal and motor excitability state. The goal of this work is to explore to which degree the variation in MEP amplitudes can be explained by the cortical state right before the stimulation. Specifically, we analyzed a dataset acquired on eleven healthy subjects comprising, for each subject, 840 single TMS pulses applied to the left M1 during acquisition of electroencephalography (EEG) and electromyography (EMG). An interpretable convolutional neural network, named SincEEGNet, was utilized to discriminate between low- and high-corticospinal excitability trials, defined according to the MEP amplitude, using in input the pre-TMS EEG. This data-driven approach enabled considering multiple brain locations and frequency bands without any a priori selection. Post-hoc interpretation techniques were adopted to enhance interpretation by identifying the more relevant EEG features for the classification. Results show that individualized classifiers successfully discriminated between low and high M1 excitability states in all participants. Outcomes of the interpretation methods suggest the importance of the electrodes situated over the TMS stimulation site, as well as the relevance of the temporal samples of the input EEG closer to the stimulation time. This novel decoding method allows causal investigation of the cortical excitability state, which may be relevant for personalizing and increasing the efficacy of therapeutic brain-state dependent brain stimulation (for example in patients affected by Parkinson’s disease).

Performance Analysis of Topological Quantum Error Correcting Codes

In the last few years there has been a great development of techniques like quantum computers and quantum communication systems, due to their huge potentialities and the growing number of applications. However, physical qubits experience a lot of nonidealities, like measurement errors and decoherence, that generate failures in the quantum computation. This work shows how it is possible to exploit concepts from classical information in order to realize quantum error-correcting codes, adding some redundancy qubits. In particular, the threshold theorem states that it is possible to lower the percentage of failures in the decoding at will, if the physical error rate is below a given accuracy threshold. The focus will be on codes belonging to the family of the topological codes, like toric, planar and XZZX surface codes. Firstly, they will be compared from a theoretical point of view, in order to show their advantages and disadvantages. The algorithms behind the minimum perfect matching decoder, the most popular for such codes, will be presented. The last section will be dedicated to the analysis of the performances of these topological codes with different error channel models, showing interesting results. In particular, while the error correction capability of surface codes decreases in presence of biased errors, XZZX codes own some intrinsic symmetries that allow them to improve their performances if one kind of error occurs more frequently than the others.

Motor Unit Spike Trains Recontruction from sEMG Signals for Gesture Classification on a Low-Power Processing Platform

Hand gesture recognition based on surface electromyography (sEMG) signals is a promising approach for the development of intuitive human-machine interfaces (HMIs) in domains such as robotics and prosthetics. The sEMG signal arises from the muscles' electrical activity, and can thus be used to recognize hand gestures. The decoding from sEMG signals to actual control signals is non-trivial; typically, control systems map sEMG patterns into a set of gestures using machine learning, failing to incorporate any physiological insight. This master thesis aims at developing a bio-inspired hand gesture recognition system based on neuromuscular spike extraction rather than on simple pattern recognition. The system relies on a decomposition algorithm based on independent component analysis (ICA) that decomposes the sEMG signal into its constituent motor unit spike trains, which are then forwarded to a machine learning classifier. Since ICA does not guarantee a consistent motor unit ordering across different sessions, 3 approaches are proposed: 2 ordering criteria based on firing rate and negative entropy, and a re-calibration approach that allows the decomposition model to retain information about previous sessions. Using a multilayer perceptron (MLP), the latter approach results in an accuracy up to 99.4% in a 1-subject, 1-degree of freedom scenario. Afterwards, the decomposition and classification pipeline for inference is parallelized and profiled on the PULP platform, achieving a latency < 50 ms and an energy consumption < 1 mJ. Both the classification models tested (a support vector machine and a lightweight MLP) yielded an accuracy > 92% in a 1-subject, 5-classes (4 gestures and rest) scenario. These results prove that the proposed system is suitable for real-time execution on embedded platforms and also capable of matching the accuracy of state-of-the-art approaches, while also giving some physiological insight on the neuromuscular spikes underlying the sEMG.