Methods for integrating machine learning and constrained optimization

In the framework of industrial problems, the application of Constrained Optimization is known to have overall very good modeling capability and performance and stands as one of the most powerful, explored, and exploited tool to address prescriptive tasks. The number of applications is huge, ranging from logistics to transportation, packing, production, telecommunication, scheduling, and much more. The main reason behind this success is to be found in the remarkable effort put in the last decades by the OR community to develop realistic models and devise exact or approximate methods to solve the largest variety of constrained or combinatorial optimization problems, together with the spread of computational power and easily accessible OR software and resources. On the other hand, the technological advancements lead to a data wealth never seen before and increasingly push towards methods able to extract useful knowledge from them; among the data-driven methods, Machine Learning techniques appear to be one of the most promising, thanks to its successes in domains like Image Recognition, Natural Language Processes and playing games, but also the amount of research involved. The purpose of the present research is to study how Machine Learning and Constrained Optimization can be used together to achieve systems able to leverage the strengths of both methods: this would open the way to exploiting decades of research on resolution techniques for COPs and constructing models able to adapt and learn from available data. In the first part of this work, we survey the existing techniques and classify them according to the type, method, or scope of the integration; subsequently, we introduce a novel and general algorithm devised to inject knowledge into learning models through constraints, Moving Target. In the last part of the thesis, two applications stemming from real-world projects and done in collaboration with Optit will be presented.

The three-dimensional single-bin-size bin packing problem: combining metaheuristic and machine learning approaches

The Three-Dimensional Single-Bin-Size Bin Packing Problem is one of the most studied problem in the Cutting & Packing category. From a strictly mathematical point of view, it consists of packing a finite set of strongly heterogeneous “small” boxes, called items, into a finite set of identical “large” rectangles, called bins, minimizing the unused volume and requiring that the items are packed without overlapping. The great interest is mainly due to the number of real-world applications in which it arises, such as pallet and container loading, cutting objects out of a piece of material and packaging design. Depending on these real-world applications, more objective functions and more practical constraints could be needed. After a brief discussion about the real-world applications of the problem and a exhaustive literature review, the design of a two-stage algorithm to solve the aforementioned problem is presented. The algorithm must be able to provide the spatial coordinates of the placed boxes vertices and also the optimal boxes input sequence, while guaranteeing geometric, stability, fragility constraints and a reduced computational time. Due to NP-hard complexity of this type of combinatorial problems, a fusion of metaheuristic and machine learning techniques is adopted. In particular, a hybrid genetic algorithm coupled with a feedforward neural network is used. In the first stage, a rich dataset is created starting from a set of real input instances provided by an industrial company and the feedforward neural network is trained on it. After its training, given a new input instance, the hybrid genetic algorithm is able to run using the neural network output as input parameter vector, providing as output the optimal solution. The effectiveness of the proposed works is confirmed via several experimental tests.

A multidisciplinary approach to the study of protein dynamics and signal transmission

Allostery is a phenomenon of fundamental importance in biology, allowing regulation of function and dynamic adaptability of enzymes and proteins. Despite the allosteric effect was first observed more than a century ago allostery remains a biophysical enigma, defined as the “second secret of life”. The challenge is mainly associated to the rather complex nature of the allosteric mechanisms, which manifests itself as the alteration of the biological function of a protein/enzyme (e.g. ligand/substrate binding at the active site) by binding of “other object” (“allos stereos” in Greek) at a site distant (> 1 nanometer) from the active site, namely the effector site. Thus, at the heart of allostery there is signal propagation from the effector to the active site through a dense protein matrix, with a fundamental challenge being represented by the elucidation of the physico-chemical interactions between amino acid residues allowing communicatio n between the two binding sites, i.e. the “allosteric pathways”. Here, we propose a multidisciplinary approach based on a combination of computational chemistry, involving molecular dynamics simulations of protein motions, (bio)physical analysis of allosteric systems, including multiple sequence alignments of known allosteric systems, and mathematical tools based on graph theory and machine learning that can greatly help understanding the complexity of dynamical interactions involved in the different allosteric systems. The project aims at developing robust and fast tools to identify unknown allosteric pathways. The characterization and predictions of such allosteric spots could elucidate and fully exploit the power of allosteric modulation in enzymes and DNA-protein complexes, with great potential applications in enzyme engineering and drug discovery.

Mathematical methods to analyze and interpret calcium signals of astrocytes

Astrocytes are the most numerous glial cell type in the mammalian brain and permeate the entire CNS interacting with neurons, vasculature, and other glial cells. Astrocytes display intracellular calcium signals that encode information about local synaptic function, distributed network activity, and high-level cognitive functions. Several studies have investigated the calcium dynamics of astrocytes in sensory areas and have shown that these cells can encode sensory stimuli. Nevertheless, only recently the neuro-scientific community has focused its attention on the role and functions of astrocytes in associative areas such as the hippocampus. In our first study, we used the information theory formalism to show that astrocytes in the CA1 area of the hippocampus recorded with 2-photon fluorescence microscopy during spatial navigation encode spatial information that is complementary and synergistic to information encoded by nearby "place cell" neurons. In our second study, we investigated various computational aspects of applying the information theory formalism to astrocytic calcium data. For this reason, we generated realistic simulations of calcium signals in astrocytes to determine optimal hyperparameters and procedures of information measures and applied them to real astrocytic calcium imaging data. Calcium signals of astrocytes are characterized by complex spatiotemporal dynamics occurring in subcellular parcels of the astrocytic domain which makes studying these cells in 2-photon calcium imaging recordings difficult. However, current analytical tools which identify the astrocytic subcellular regions are time consuming and extensively rely on user-defined parameters. Here, we present Rapid Astrocytic calcium Spatio-Temporal Analysis (RASTA), a novel machine learning algorithm for spatiotemporal semantic segmentation of 2-photon calcium imaging recordings of astrocytes which operates without human intervention. We found that RASTA provided fast and accurate identification of astrocytic cell somata, processes, and cellular domains, extracting calcium signals from identified regions of interest across individual cells and populations of hundreds of astrocytes recorded in awake mice.

How to explain: from theory to practice

Today we live in an age where the internet and artificial intelligence allow us to search for information through impressive amounts of data, opening up revolutionary new ways to make sense of reality and understand our world. However, it is still an area of improvement to exploit the full potential of large amounts of explainable information by distilling it automatically in an intuitive and user-centred explanation. For instance, different people (or artificial agents) may search for and request different types of information in a different order, so it is unlikely that a short explanation can suffice for all needs in the most generic case. Moreover, dumping a large portion of explainable information in a one-size-fits-all representation may also be sub-optimal, as the needed information may be scarce and dispersed across hundreds of pages. The aim of this work is to investigate how to automatically generate (user-centred) explanations from heterogeneous and large collections of data, with a focus on the concept of explanation in a broad sense, as a critical artefact for intelligence, regardless of whether it is human or robotic. Our approach builds on and extends Achinstein’s philosophical theory of explanations, where explaining is an illocutionary (i.e., broad but relevant) act of usefully answering questions. Specifically, we provide the theoretical foundations of Explanatory Artificial Intelligence (YAI), formally defining a user-centred explanatory tool and the space of all possible explanations, or explanatory space, generated by it. We present empirical results in support of our theory, showcasing the implementation of YAI tools and strategies for assessing explainability. To justify and evaluate the proposed theories and models, we considered case studies at the intersection of artificial intelligence and law, particularly European legislation. Our tools helped produce better explanations of software documentation and legal texts for humans and complex regulations for reinforcement learning agents.

Computational methods applied to undeciphered scripts from the Aegean and Cyprus: Linear A and Cypro-Minoan

The study of ancient, undeciphered scripts presents unique challenges, that depend both on the nature of the problem and on the peculiarities of each writing system. In this thesis, I present two computational approaches that are tailored to two different tasks and writing systems. The first of these methods is aimed at the decipherment of the Linear A afraction signs, in order to discover their numerical values. This is achieved with a combination of constraint programming, ad-hoc metrics and paleographic considerations. The second main contribution of this thesis regards the creation of an unsupervised deep learning model which uses drawings of signs from ancient writing system to learn to distinguish different graphemes in the vector space. This system, which is based on techniques used in the field of computer vision, is adapted to the study of ancient writing systems by incorporating information about sequences in the model, mirroring what is often done in natural language processing. In order to develop this model, the Cypriot Greek Syllabary is used as a target, since this is a deciphered writing system. Finally, this unsupervised model is adapted to the undeciphered Cypro-Minoan and it is used to answer open questions about this script. In particular, by reconstructing multiple allographs that are not agreed upon by paleographers, it supports the idea that Cypro-Minoan is a single script and not a collection of three script like it was proposed in the literature. These results on two different tasks shows that computational methods can be applied to undeciphered scripts, despite the relatively low amount of available data, paving the way for further advancement in paleography using these methods.

Weak reasoning argumentative theory. A dialogical approach for the discovery of philosophy in high schools

In the literature on philosophical practices, despite the crucial role that argumentation plays in these activities, no specific argumentative theories have ever been proposed to assist the figure of the facilitator in conducting philosophical dialogue and to enhance student’s critical thinking skills. The dissertation starts from a cognitive perspective that challenges the classic Cartesian notion of rationality by focusing on limits and biases of human reasoning. An argumentative model (WRAT – Weak Reasoning Argumentative Theory) is then outlined in order to respond to the needs of philosophical dialogue. After justifying the claim that this learning activity, among other inductive methodologies, is the most suitable for critical thinking education, I inquired into the specific goal of ‘arguing’ within this context by means of the tools provided by Speech Act Theory: the speaker’s intention is to construct new knowledge by questioning her own and other’s beliefs. The model proposed has been theorized on this assumption, starting from which the goals, and, in turn, the related norms, have been pinpointed. In order to include all the epistemic attitudes required to accomplish the complex task of arguing in philosophical dialogue, I needed to integrate two opposed cognitive accounts, Dual Process Theory and Evolutionary Approach, that, although they provide incompatible descriptions of reasoning, can be integrated to provide a normative account of argumentation. The model, apart from offering a theoretical contribution to argumentation studies, is designed to be applied to the Italian educational system, in particular to classes in technical and professional high schools belonging to the newly created network Inventio. This initiative is one of the outcomes of the research project by the same name, which also includes an original Syllabus, research seminars, a monitoring action and publications focused on introducing philosophy, in the form of workshop activities, into technical and professional schools.

A deep learning model via thermodynamics and condensed matter physics

Deep Learning architectures give brilliant results in a large variety of fields, but a comprehensive theoretical description of their inner functioning is still lacking. In this work, we try to understand the behavior of neural networks by modelling in the frameworks of Thermodynamics and Condensed Matter Physics. We approach neural networks as in a real laboratory and we measure the frequency spectrum and the entropy of the weights of the trained model. The stochasticity of the training occupies a central role in the dynamics of the weights and makes it difficult to assimilate neural networks to simple physical systems. However, the analogy with Thermodynamics and the introduction of a well defined temperature leads us to an interesting result: if we eliminate from a CNN the "hottest" filters, the performance of the model remains the same, whereas, if we eliminate the "coldest" ones, the performance gets drastically worst. This result could be exploited in the realization of a training loop which eliminates the filters that do not contribute to loss reduction. In this way, the computational cost of the training will be lightened and more importantly this would be done by following a physical model. In any case, beside important practical applications, our analysis proves that a new and improved modeling of Deep Learning systems can pave the way to new and more efficient algorithms.

Computational investigation of a stereoselective synthesis of atropisomeric hydrazides possessing an N–N stereogenic axis

Axially chiral substrates are an interesting and widely studied class of compounds as they can be found in bioactive natural products and are employed as functional materials or as ligands in asymmetric catalytic processes. One branch of this family is the well-known world of the atropisomers. Among them, atropisomeric compounds possessing an N–N stereogenic axis are one truthfully fascinating system but not completely understood yet. In this thesis, we computationally investigated the mechanism of the diastereoselective formation of the N – N chiral axis of a hydrazide under asymmetric phase transfer catalytic conditions. Moreover, during this study, torsional barriers have been calculated for both the reagent and the product at the density functional theory (DFT). These values turned out to suitably match the experimental values and observations. Finally, Electronic Circular Dichroism (ECD) spectra have been simulated in order to assign the chiral absolute configuration to the products.

A biophysical model for valence-based navigational learning: the role of D-Serine

Recent experiments have revealed the fundamental importance of neuromodulatory action on activity-dependent synaptic plasticity underlying behavioral learning and spatial memory formation. Neuromodulators affect synaptic plasticity through the modification of the dynamics of receptors on the synaptic membrane. However, chemical substances other than neuromodulators, such as receptors co-agonists, can influence the receptors' dynamics and thus participate in determining plasticity. Here we focus on D-serine, which has been observed to affect the activity thresholds of synaptic plasticity by co-activating NMDA receptors. We use a computational model for spatial value learning with plasticity between two place cell layers. The D-serine release is CB1R mediated and the model reproduces the impairment of spatial memory due to the astrocytic CB1R knockout for a mouse navigating in the Morris water maze. The addition of path-constraining obstacles shows how performance impairment depends on the environment's topology. The model can explain the experimental evidence and produce useful testable predictions to increase our understanding of the complex mechanisms underlying learning.

[the Teaching Of Social Sciences In Health: Between Practice And Theory].

The models of teaching social sciences and clinical practice are insufficient for the needs of practical-reflective teaching of social sciences applied to health. The scope of this article is to reflect on the challenges and perspectives of social science education for health professionals. In the 1950s the important movement bringing together social sciences and the field of health began, however weak credentials still prevail. This is due to the low professional status of social scientists in health and the ill-defined position of the social sciences professionals in the health field. It is also due to the scant importance attributed by students to the social sciences, the small number of professionals and the colonization of the social sciences by the biomedical culture in the health field. Thus, the professionals of social sciences applied to health are also faced with the need to build an identity, even after six decades of their presence in the field of health. This is because their ambivalent status has established them as a partial, incomplete and virtual presence, requiring a complex survival strategy in the nebulous area between social sciences and health.

Atomic Charge Transfer-counter Polarization Effects Determine Infrared Ch Intensities Of Hydrocarbons: A Quantum Theory Of Atoms In Molecules Model.

Atomic charge transfer-counter polarization effects determine most of the infrared fundamental CH intensities of simple hydrocarbons, methane, ethylene, ethane, propyne, cyclopropane and allene. The quantum theory of atoms in molecules/charge-charge flux-dipole flux model predicted the values of 30 CH intensities ranging from 0 to 123 km mol(-1) with a root mean square (rms) error of only 4.2 km mol(-1) without including a specific equilibrium atomic charge term. Sums of the contributions from terms involving charge flux and/or dipole flux averaged 20.3 km mol(-1), about ten times larger than the average charge contribution of 2.0 km mol(-1). The only notable exceptions are the CH stretching and bending intensities of acetylene and two of the propyne vibrations for hydrogens bound to sp hybridized carbon atoms. Calculations were carried out at four quantum levels, MP2/6-311++G(3d,3p), MP2/cc-pVTZ, QCISD/6-311++G(3d,3p) and QCISD/cc-pVTZ. The results calculated at the QCISD level are the most accurate among the four with root mean square errors of 4.7 and 5.0 km mol(-1) for the 6-311++G(3d,3p) and cc-pVTZ basis sets. These values are close to the estimated aggregate experimental error of the hydrocarbon intensities, 4.0 km mol(-1). The atomic charge transfer-counter polarization effect is much larger than the charge effect for the results of all four quantum levels. Charge transfer-counter polarization effects are expected to also be important in vibrations of more polar molecules for which equilibrium charge contributions can be large.

Beliefs Related To Adherence To Oral Antidiabetic Treatment According To The Theory Of Planned Behavior.

to identify salient behavioral, normative, control and self-efficacy beliefs related to the behavior of adherence to oral antidiabetic agents, using the Theory of Planned Behavior. cross-sectional, exploratory study with 17 diabetic patients in chronic use of oral antidiabetic medication and in outpatient follow-up. Individual interviews were recorded, transcribed and content-analyzed using pre-established categories. behavioral beliefs concerning advantages and disadvantages of adhering to medication emerged, such as the possibility of avoiding complications from diabetes, preventing or delaying the use of insulin, and a perception of side effects. The children of patients and physicians are seen as important social references who influence medication adherence. The factors that facilitate adherence include access to free-of-cost medication and taking medications associated with temporal markers. On the other hand, a complex therapeutic regimen was considered a factor that hinders adherence. Understanding how to use medication and forgetfulness impact the perception of patients regarding their ability to adhere to oral antidiabetic agents. medication adherence is a complex behavior permeated by behavioral, normative, control and self-efficacy beliefs that should be taken into account when assessing determinants of behavior.

Ecological Literacy And Beyond: Problem-based Learning For Future Professionals.

Ecological science contributes to solving a broad range of environmental problems. However, lack of ecological literacy in practice often limits application of this knowledge. In this paper, we highlight a critical but often overlooked demand on ecological literacy: to enable professionals of various careers to apply scientific knowledge when faced with environmental problems. Current university courses on ecology often fail to persuade students that ecological science provides important tools for environmental problem solving. We propose problem-based learning to improve the understanding of ecological science and its usefulness for real-world environmental issues that professionals in careers as diverse as engineering, public health, architecture, social sciences, or management will address. Courses should set clear learning objectives for cognitive skills they expect students to acquire. Thus, professionals in different fields will be enabled to improve environmental decision-making processes and to participate effectively in multidisciplinary work groups charged with tackling environmental issues.

Uma observação sobre o lugar da teoria da ligação e do critério temático dentro do programa minimalista

In this paper I present some evidencie that forces us to conclude that within the Minimalist Program (Chomsky 1993; 1995), Binding Theory (BT) should be computed after LF (Logical Form). I show that derivations leading to structures containing violations of BT-Principles must converge at LF, since less economical alternative derivations respecting those principles are also ungrammatical. Being irrelevant to the notion of convergence, BT must apply after LF. A similar reasoning reveals that the Theta-Criterion should have the status of a bare output condition appling at LF, since less economical derivations are allowed by the computational system to prevent violations of it.