User Behavioral Modeling of Web-based Systems for Continuous User Authentication

Authentication plays an important role in how we interact with computers, mobile devices, the web, etc. The idea of authentication is to uniquely identify a user before granting access to system privileges. For example, in recent years more corporate information and applications have been accessible via the Internet and Intranet. Many employees are working from remote locations and need access to secure corporate files. During this time, it is possible for malicious or unauthorized users to gain access to the system. For this reason, it is logical to have some mechanism in place to detect whether the logged-in user is the same user in control of the user's session. Therefore, highly secure authentication methods must be used. We posit that each of us is unique in our use of computer systems. It is this uniqueness that is leveraged to "continuously authenticate users" while they use web software. To monitor user behavior, n-gram models are used to capture user interactions with web-based software. This statistical language model essentially captures sequences and sub-sequences of user actions, their orderings, and temporal relationships that make them unique by providing a model of how each user typically behaves. Users are then continuously monitored during software operations. Large deviations from "normal behavior" can possibly indicate malicious or unintended behavior. This approach is implemented in a system called Intruder Detector (ID) that models user actions as embodied in web logs generated in response to a user's actions. User identification through web logs is cost-effective and non-intrusive. We perform experiments on a large fielded system with web logs of approximately 4000 users. For these experiments, we use two classification techniques; binary and multi-class classification. We evaluate model-specific differences of user behavior based on coarse-grain (i.e., role) and fine-grain (i.e., individual) analysis. A specific set of metrics are used to provide valuable insight into how each model performs. Intruder Detector achieves accurate results when identifying legitimate users and user types. This tool is also able to detect outliers in role-based user behavior with optimal performance. In addition to web applications, this continuous monitoring technique can be used with other user-based systems such as mobile devices and the analysis of network traffic.

An internet-based control engineering laboratory

This paper presents LABNET, an internet-based remote laboratory for control engineering education developed at UEM-University. At present, the remote laboratory integrates three basic physical systems (level control, temperature control and ship stabilizing system). In this paper, the LABNET architecture is presented and discussed in detail. Issues concerned with concurrent user access, local or remote feedback, automatic report generating and reusing of experiment’s templates have been addressed. Furthermore, the experiences gained developing, testing and using the system will be also presented and their consequences for future design.

Development and Validation of a Model-Based Combustion Controller for Water Injection Management

This manuscript reports the overall development of a Ph.D. research project during the “Mechanics and advanced engineering sciences” course at the Department of Industrial Engineering of the University of Bologna. The project is focused on the development of a combustion control system for an innovative Spark Ignited engine layout. In details, the controller is oriented to manage a prototypal engine equipped with a Port Water Injection system. The water injection technology allows an increment of combustion efficiency due to the knock mitigation effect that permits to keep the combustion phasing closer to the optimal position with respect to the traditional layout. At the beginning of the project, the effects and the possible benefits achievable by water injection have been investigated by a focused experimental campaign. Then the data obtained by combustion analysis have been processed to design a control-oriented combustion model. The model identifies the correlation between Spark Advance, combustion phasing and injected water mass, and two different strategies are presented, both based on an analytic and semi-empirical approach and therefore compatible with a real-time application. The model has been implemented in a combustion controller that manages water injection to reach the best achievable combustion efficiency while keeping knock levels under a pre-established threshold. Three different versions of the algorithm are described in detail. This controller has been designed and pre-calibrated in a software-in-the-loop environment and later an experimental validation has been performed with a rapid control prototyping approach to highlight the performance of the system on real set-up. To further make the strategy implementable on an onboard application, an estimation algorithm of combustion phasing, necessary for the controller, has been developed during the last phase of the PhD Course, based on accelerometric signals.

Model-Based validation of Driver Drowsiness Detection System for ADAS

The work described in this Master’s Degree thesis was born after the collaboration with the company Maserati S.p.a, an Italian luxury car maker with its headquarters located in Modena, in the heart of the Italian Motor Valley, where I worked as a stagiaire in the Virtual Engineering team between September 2021 and February 2022. This work proposes the validation using real-world ECUs of a Driver Drowsiness Detection (DDD) system prototype based on different detection methods with the goal to overcome input signal losses and system failures. Detection methods of different categories have been chosen from literature and merged with the goal of utilizing the benefits of each of them, overcoming their limitations and limiting as much as possible their degree of intrusiveness to prevent any kind of driving distraction: an image processing-based technique for human physical signals detection as well as methods based on driver-vehicle interaction are used. A Driver-In-the-Loop simulator is used to gather real data on which a Machine Learning-based algorithm will be trained and validated. These data come from the tests that the company conducts in its daily activities so confidential information about the simulator and the drivers will be omitted. Although the impact of the proposed system is not remarkable and there is still work to do in all its elements, the results indicate the main advantages of the system in terms of robustness against subsystem failures and signal losses.

Massive MIMO-based system analysis and comparison with MIMO based-systems

The thesis is focused on introducing basic MIMO-based and Massive MIMO-based systems and their possible benefits. Then going through the implementation options that we have, according to 3GPP standards, for 5G systems and how the transition is done from a non-standalone 5G RAN to a completely standalone 5G RAN. Having introduced the above-mentioned subjects and providing some definition of telecommunications principles, we move forward to a more technical analysis of the Capacity, Throughput, Power consumption, and Costs. Comparing all the mentioned parameters between a Massive-MIMO-based system and a MIMO-based system. In the analysis of power consumption and costs, we also introduce the concept of virtualization and its benefits in terms of both power and costs. Finally, we try to justify a trade-off between having a more reliable system with a high capacity and throughput while keeping the costs as low as possible.

Model-Based Contrastive Explanations for XAIP: Towards a General Model and Prototype

Planning is an important sub-field of artificial intelligence (AI) focusing on letting intelligent agents deliberate on the most adequate course of action to attain their goals. Thanks to the recent boost in the number of critical domains and systems which exploit planning for their internal procedures, there is an increasing need for planning systems to become more transparent and trustworthy. Along this line, planning systems are now required to produce not only plans but also explanations about those plans, or the way they were attained. To address this issue, a new research area is emerging in the AI panorama: eXplainable AI (XAI), within which explainable planning (XAIP) is a pivotal sub-field. As a recent domain, XAIP is far from mature. No consensus has been reached in the literature about what explanations are, how they should be computed, and what they should explain in the first place. Furthermore, existing contributions are mostly theoretical, and software implementations are rarely more than preliminary. To overcome such issues, in this thesis we design an explainable planning framework bridging the gap between theoretical contributions from literature and software implementations. More precisely, taking inspiration from the state of the art, we develop a formal model for XAIP, and the software tool enabling its practical exploitation. Accordingly, the contribution of this thesis is four-folded. First, we review the state of the art of XAIP, supplying an outline of its most significant contributions from the literature. We then generalise the aforementioned contributions into a unified model for XAIP, aimed at supporting model-based contrastive explanations. Next, we design and implement an algorithm-agnostic library for XAIP based on our model. Finally, we validate our library from a technological perspective, via an extensive testing suite. Furthermore, we assess its performance and usability through a set of benchmarks and end-to-end examples.

Melphalan dosing regimens for management of recurrent melanoma by isolated limb perfusion: Application of a physiological pharmacokinetic model based on melphalan distribution in the isolated perfused rat hindlimb

The optimal dosing schedule for melphalan therapy of recurrent malignant melanoma in isolated limb perfusions has been examined using a physiological pharmacokinetic model with data from isolated rat hindlimb perfusions (IRHP), The study included a comparison of melphalan distribution in IRHP under hyperthermia and normothermia conditions. Rat hindlimbs were perfused with Krebs-Henseleit buffer containing 4.7% bovine serum albumin at 37 or 41.5 degrees C at a flow rate of 4 ml/min. Concentrations of melphalan in perfusate and tissues were determined by high performance liquid chromatography with fluorescence detection, The concentration of melphalan in perfusate and tissues was linearly related to the input concentration. The rate and amount of melphalan uptake into the different tissues was higher at 41.5 degrees C than at 37 degrees C. A physiological pharmacokinetic model was validated from the tissue and perfusate time course of melphalan after melphalan perfusion. Application of the model involved the amount of melphalan exposure in the muscle, skin and fat in a recirculation system was related to the method of melphalan administration: single bolus > divided bolus > infusion, The peak concentration of melphalan in the perfusate was also related to the method of administration in the same order, Infusing the total dose of melphalan over 20 min during a 60 min perfusion optimized the exposure of tissues to melphalan whilst minimizing the peak perfusate concentration of melphalan. It is suggested that this method of melphalan administration may be preferable to other methods in terms of optimizing the efficacy of melphalan whilst minimizing the limb toxicity associated with its use in isolated limb perfusion.

Diagnostic model based on Raman spectra of normal, hyperplasia and prostate adenocarcinoma tissues in vitro

This study evaluated the use of Raman spectroscopy to identify the spectral differences between normal (N), benign hyperplasia (BPH) and adenocarcinoma (CaP) in fragments of prostate biopsies in vitro with the aim of developing a spectral diagnostic model for tissue classification. A dispersive Raman spectrometer was used with 830 nm wavelength and 80 mW excitation. Following Raman data collection and tissue histopathology (48 fragments diagnosed as N, 43 as BPH and 14 as CaP), two diagnostic models were developed in order to extract diagnostic information: the first using PCA and Mahalanobis analysis techniques and the second one a simplified biochemical model based on spectral features of cholesterol, collagen, smooth muscle cell and adipocyte. Spectral differences between N, BPH and CaP tissues, were observed mainly in the Raman bands associated with proteins, lipids, nucleic and amino acids. The PCA diagnostic model showed a sensitivity and specificity of 100%, which indicates the ability of PCA and Mahalanobis distance techniques to classify tissue changes in vitro. Also, it was found that the relative amount of collagen decreased while the amount of cholesterol and adipocyte increased with severity of the disease. Smooth muscle cell increased in BPH tissue. These characteristics were used for diagnostic purposes.

Model based procedure for scale-up of wet overflow ball mills Part 1: Outline of the methodology

Bond's method for ball mill scale-up only gives the mill power draw for a given duty. This method is incompatible with computer modelling and simulation techniques. It might not be applicable for the design of fine grinding ball mills and ball mills preceded by autogenous and semi-autogenous grinding mills. Model-based ball mill scale-up methods have not been validated using a wide range of full-scale circuit data. Their accuracy is therefore questionable. Some of these methods also need expensive pilot testing. A new ball mill scale-up procedure is developed which does not have these limitations. This procedure uses data from two laboratory tests to determine the parameters of a ball mill model. A set of scale-up criteria then scales-up these parameters. The procedure uses the scaled-up parameters to simulate the steady state performance of full-scale mill circuits. At the end of the simulation, the scale-up procedure gives the size distribution, the volumetric flowrate and the mass flowrate of all the streams in the circuit, and the mill power draw.

A mixture model-based approach to the clustering of microarray expression data

Motivation: This paper introduces the software EMMIX-GENE that has been developed for the specific purpose of a model-based approach to the clustering of microarray expression data, in particular, of tissue samples on a very large number of genes. The latter is a nonstandard problem in parametric cluster analysis because the dimension of the feature space (the number of genes) is typically much greater than the number of tissues. A feasible approach is provided by first selecting a subset of the genes relevant for the clustering of the tissue samples by fitting mixtures of t distributions to rank the genes in order of increasing size of the likelihood ratio statistic for the test of one versus two components in the mixture model. The imposition of a threshold on the likelihood ratio statistic used in conjunction with a threshold on the size of a cluster allows the selection of a relevant set of genes. However, even this reduced set of genes will usually be too large for a normal mixture model to be fitted directly to the tissues, and so the use of mixtures of factor analyzers is exploited to reduce effectively the dimension of the feature space of genes. Results: The usefulness of the EMMIX-GENE approach for the clustering of tissue samples is demonstrated on two well-known data sets on colon and leukaemia tissues. For both data sets, relevant subsets of the genes are able to be selected that reveal interesting clusterings of the tissues that are either consistent with the external classification of the tissues or with background and biological knowledge of these sets.