On Reasonable Space and Time Cost Models for the λ-Calculus

Slot and van Emde Boas Invariance Thesis states that a time (respectively, space) cost model is reasonable for a computational model C if there are mutual simulations between Turing machines and C such that the overhead is polynomial in time (respectively, linear in space). The rationale is that under the Invariance Thesis, complexity classes such as LOGSPACE, P, PSPACE, become robust, i.e. machine independent. In this dissertation, we want to find out if it possible to define a reasonable space cost model for the lambda-calculus, the paradigmatic model for functional programming languages. We start by considering an unusual evaluation mechanism for the lambda-calculus, based on Girard's Geometry of Interaction, that was conjectured to be the key ingredient to obtain a space reasonable cost model. By a fine complexity analysis of this schema, based on new variants of non-idempotent intersection types, we disprove this conjecture. Then, we change the target of our analysis. We consider a variant over Krivine's abstract machine, a standard evaluation mechanism for the call-by-name lambda-calculus, optimized for space complexity, and implemented without any pointer. A fine analysis of the execution of (a refined version of) the encoding of Turing machines into the lambda-calculus allows us to conclude that the space consumed by this machine is indeed a reasonable space cost model. In particular, for the first time we are able to measure also sub-linear space complexities. Moreover, we transfer this result to the call-by-value case. Finally, we provide also an intersection type system that characterizes compositionally this new reasonable space measure. This is done through a minimal, yet non trivial, modification of the original de Carvalho type system.

Numerical and analytical vibro-acoustic modelling of porous materials: the Cell Method for poroelasticity and the case of inclusions

Porous materials are widely used in many fields of industrial applications, to achieve the requirements of noise reduction, that nowadays derive from strict regulations. The modeling of porous materials is still a problematic issue. Numerical simulations are often problematic in case of real complex geometries, especially in terms of computational times and convergence. At the same time, analytical models, even if partly limited by restrictive simplificative hypotheses, represent a powerful instrument to capture quickly the physics of the problem and general trends. In this context, a recently developed numerical method, called the Cell Method, is described, is presented in the case of the Biot's theory and applied for representative cases. The peculiarity of the Cell Method is that it allows for a direct algebraic and geometrical discretization of the field equations, without any reduction to a weak integral form. Then, the second part of the thesis presents the case of interaction between two poroelastic materials under the context of double porosity. The idea of using periodically repeated inclusions of a second porous material into a layer composed by an original material is described. In particular, the problem is addressed considering the efficiency of the analytical method. A analytical procedure for the simulation of heterogeneous layers based is described and validated considering both conditions of absorption and transmission; a comparison with the available numerical methods is performed. ---------------- I materiali porosi sono ampiamente utilizzati per diverse applicazioni industriali, al fine di raggiungere gli obiettivi di riduzione del rumore, che sono resi impegnativi da norme al giorno d'oggi sempre più stringenti. La modellazione dei materiali porori per applicazioni vibro-acustiche rapprensenta un aspetto di una certa complessità. Le simulazioni numeriche sono spesso problematiche quando siano coinvolte geometrie di pezzi reali, in particolare riguardo i tempi computazionali e la convergenza. Allo stesso tempo, i modelli analitici, anche se parzialmente limitati a causa di ipotesi semplificative che ne restringono l'ambito di utilizzo, rappresentano uno strumento molto utile per comprendere rapidamente la fisica del problema e individuare tendenze generali. In questo contesto, un metodo numerico recentemente sviluppato, il Metodo delle Celle, viene descritto, implementato nel caso della teoria di Biot per la poroelasticità e applicato a casi rappresentativi. La peculiarità del Metodo delle Celle consiste nella discretizzazione diretta algebrica e geometrica delle equazioni di campo, senza alcuna riduzione a forme integrali deboli. Successivamente, nella seconda parte della tesi viene presentato il caso delle interazioni tra due materiali poroelastici a contatto, nel contesto dei materiali a doppia porosità. Viene descritta l'idea di utilizzare inclusioni periodicamente ripetute di un secondo materiale poroso all'interno di un layer a sua volta poroso. In particolare, il problema è studiando il metodo analitico e la sua efficienza. Una procedura analitica per il calcolo di strati eterogenei di materiale viene descritta e validata considerando sia condizioni di assorbimento, sia di trasmissione; viene effettuata una comparazione con i metodi numerici a disposizione.

Preventing worker’s musculoskeletal disorders with wearable inertial sensors: ergonomics and modelling of physical human-robot interactions

The most widespread work-related diseases are musculoskeletal disorders (MSD) caused by awkward postures and excessive effort to upper limb muscles during work operations. The use of wearable IMU sensors could monitor the workers constantly to prevent hazardous actions, thus diminishing work injuries. In this thesis, procedures are developed and tested for ergonomic analyses in a working environment, based on a commercial motion capture system (MoCap) made of 17 Inertial Measurement Units (IMUs). An IMU is usually made of a tri-axial gyroscope, a tri-axial accelerometer, and a tri-axial magnetometer that, through sensor fusion algorithms, estimates its attitude. Effective strategies for preventing MSD rely on various aspects: firstly, the accuracy of the IMU, depending on the chosen sensor and its calibration; secondly, the correct identification of the pose of each sensor on the worker’s body; thirdly, the chosen multibody model, which must consider both the accuracy and the computational burden, to provide results in real-time; finally, the model scaling law, which defines the possibility of a fast and accurate personalization of the multibody model geometry. Moreover, the MSD can be diminished using collaborative robots (cobots) as assisted devices for complex or heavy operations to relieve the worker's effort during repetitive tasks. All these aspects are considered to test and show the efficiency and usability of inertial MoCap systems for assessing ergonomics evaluation in real-time and implementing safety control strategies in collaborative robotics. Validation is performed with several experimental tests, both to test the proposed procedures and to compare the results of real-time multibody models developed in this thesis with the results from commercial software. As an additional result, the positive effects of using cobots as assisted devices for reducing human effort in repetitive industrial tasks are also shown, to demonstrate the potential of wearable electronics in on-field ergonomics analyses for industrial applications.

Early Warning For Large Earthquakes: Observations, Models and Real-Time Data Analysis

This thesis is a collection of works focused on the topic of Earthquake Early Warning, with a special attention to large magnitude events. The topic is addressed from different points of view and the structure of the thesis reflects the variety of the aspects which have been analyzed. The first part is dedicated to the giant, 2011 Tohoku-Oki earthquake. The main features of the rupture process are first discussed. The earthquake is then used as a case study to test the feasibility Early Warning methodologies for very large events. Limitations of the standard approaches for large events arise in this chapter. The difficulties are related to the real-time magnitude estimate from the first few seconds of recorded signal. An evolutionary strategy for the real-time magnitude estimate is proposed and applied to the single Tohoku-Oki earthquake. In the second part of the thesis a larger number of earthquakes is analyzed, including small, moderate and large events. Starting from the measurement of two Early Warning parameters, the behavior of small and large earthquakes in the initial portion of recorded signals is investigated. The aim is to understand whether small and large earthquakes can be distinguished from the initial stage of their rupture process. A physical model and a plausible interpretation to justify the observations are proposed. The third part of the thesis is focused on practical, real-time approaches for the rapid identification of the potentially damaged zone during a seismic event. Two different approaches for the rapid prediction of the damage area are proposed and tested. The first one is a threshold-based method which uses traditional seismic data. Then an innovative approach using continuous, GPS data is explored. Both strategies improve the prediction of large scale effects of strong earthquakes.

Variational and learning models for image and time series inverse problems

Inverse problems are at the core of many challenging applications. Variational and learning models provide estimated solutions of inverse problems as the outcome of specific reconstruction maps. In the variational approach, the result of the reconstruction map is the solution of a regularized minimization problem encoding information on the acquisition process and prior knowledge on the solution. In the learning approach, the reconstruction map is a parametric function whose parameters are identified by solving a minimization problem depending on a large set of data. In this thesis, we go beyond this apparent dichotomy between variational and learning models and we show they can be harmoniously merged in unified hybrid frameworks preserving their main advantages. We develop several highly efficient methods based on both these model-driven and data-driven strategies, for which we provide a detailed convergence analysis. The arising algorithms are applied to solve inverse problems involving images and time series. For each task, we show the proposed schemes improve the performances of many other existing methods in terms of both computational burden and quality of the solution. In the first part, we focus on gradient-based regularized variational models which are shown to be effective for segmentation purposes and thermal and medical image enhancement. We consider gradient sparsity-promoting regularized models for which we develop different strategies to estimate the regularization strength. Furthermore, we introduce a novel gradient-based Plug-and-Play convergent scheme considering a deep learning based denoiser trained on the gradient domain. In the second part, we address the tasks of natural image deblurring, image and video super resolution microscopy and positioning time series prediction, through deep learning based methods. We boost the performances of supervised, such as trained convolutional and recurrent networks, and unsupervised deep learning strategies, such as Deep Image Prior, by penalizing the losses with handcrafted regularization terms.