Convex Optimization Algorithms and Statistical Bounds for Learning Structured Models

Thesis (Ph.D.)--University of Washington, 2016-08

Structure learning of context-specific graphical models

Abstract The ultimate problem considered in this thesis is modeling a high-dimensional joint distribution over a set of discrete variables. For this purpose, we consider classes of context-specific graphical models and the main emphasis is on learning the structure of such models from data. Traditional graphical models compactly represent a joint distribution through a factorization justi ed by statements of conditional independence which are encoded by a graph structure. Context-speci c independence is a natural generalization of conditional independence that only holds in a certain context, speci ed by the conditioning variables. We introduce context-speci c generalizations of both Bayesian networks and Markov networks by including statements of context-specific independence which can be encoded as a part of the model structures. For the purpose of learning context-speci c model structures from data, we derive score functions, based on results from Bayesian statistics, by which the plausibility of a structure is assessed. To identify high-scoring structures, we construct stochastic and deterministic search algorithms designed to exploit the structural decomposition of our score functions. Numerical experiments on synthetic and real-world data show that the increased exibility of context-specific structures can more accurately emulate the dependence structure among the variables and thereby improve the predictive accuracy of the models.

The introduction of e-learning in European universities: models and strategies

The aim of this paper is to present the models and the strategies of adoption of e-learning in a group of European universities, most of them located in the regions called “the four motors of Europe” (Baden-Württenberg, Catalunya, Lombardy and Rhône-Alpes) and in Switzerland. Our analysis focuses on four dimensions: the rationale behind the introduction of e-learning, the organisation of the activities and, in particular, the existence of a university centre for e-learning, the type of activities, and, finally, the type of public reached by e-learning. The majority of campus universities in our sample introduced e-learning to improve the quality of education of their students and, for the most part, as a support for existing courses. Some of the campus universities went even further insofar as they have introduced some online courses into their curricula. This has led to forms of cooperation where different universities share some of their courses. Finally, a small number of campus universities have included as part of their educational offer full distance degree programs which can be attended also by non residential students. The above cases show that there is no general move from campus universities towards distance education, but rather a more selective behaviour. Thus we conclude that e-learning, although it is undoubtedly spreading in both distance and presence universities, is not yet bringing fundamental changes in the institutions themselves. E-learning is at the moment integrated into the existing organization and educational offer. (DIPF/Orig.)

Learning data-driven models of non-verbal behaviors for building rapport using an intelligent virtual agent

There is a growing societal need to address the increasing prevalence of behavioral health issues, such as obesity, alcohol or drug use, and general lack of treatment adherence for a variety of health problems. The statistics, worldwide and in the USA, are daunting. Excessive alcohol use is the third leading preventable cause of death in the United States (with 79,000 deaths annually), and is responsible for a wide range of health and social problems. On the positive side though, these behavioral health issues (and associated possible diseases) can often be prevented with relatively simple lifestyle changes, such as losing weight with a diet and/or physical exercise, or learning how to reduce alcohol consumption. Medicine has therefore started to move toward finding ways of preventively promoting wellness, rather than solely treating already established illness.^ Evidence-based patient-centered Brief Motivational Interviewing (BMI) interventions have been found particularly effective in helping people find intrinsic motivation to change problem behaviors after short counseling sessions, and to maintain healthy lifestyles over the long-term. Lack of locally available personnel well-trained in BMI, however, often limits access to successful interventions for people in need. To fill this accessibility gap, Computer-Based Interventions (CBIs) have started to emerge. Success of the CBIs, however, critically relies on insuring engagement and retention of CBI users so that they remain motivated to use these systems and come back to use them over the long term as necessary.^ Because of their text-only interfaces, current CBIs can therefore only express limited empathy and rapport, which are the most important factors of health interventions. Fortunately, in the last decade, computer science research has progressed in the design of simulated human characters with anthropomorphic communicative abilities. Virtual characters interact using humans’ innate communication modalities, such as facial expressions, body language, speech, and natural language understanding. By advancing research in Artificial Intelligence (AI), we can improve the ability of artificial agents to help us solve CBI problems.^ To facilitate successful communication and social interaction between artificial agents and human partners, it is essential that aspects of human social behavior, especially empathy and rapport, be considered when designing human-computer interfaces. Hence, the goal of the present dissertation is to provide a computational model of rapport to enhance an artificial agent’s social behavior, and to provide an experimental tool for the psychological theories shaping the model. Parts of this thesis were already published in [LYL+12, AYL12, AL13, ALYR13, LAYR13, YALR13, ALY14].^

Self-supervised information retrieval: a novel approach based on Deep Metric Learning and Neural Language Models

Most of the existing open-source search engines, utilize keyword or tf-idf based techniques to find relevant documents and web pages relative to an input query. Although these methods, with the help of a page rank or knowledge graphs, proved to be effective in some cases, they often fail to retrieve relevant instances for more complicated queries that would require a semantic understanding to be exploited. In this Thesis, a self-supervised information retrieval system based on transformers is employed to build a semantic search engine over the library of Gruppo Maggioli company. Semantic search or search with meaning can refer to an understanding of the query, instead of simply finding words matches and, in general, it represents knowledge in a way suitable for retrieval. We chose to investigate a new self-supervised strategy to handle the training of unlabeled data based on the creation of pairs of ’artificial’ queries and the respective positive passages. We claim that by removing the reliance on labeled data, we may use the large volume of unlabeled material on the web without being limited to languages or domains where labeled data is abundant.

Pedagogia do esporte : pressupostos para uma teoria da avaliação da aprendizagem = Sport pedagogy: presuppositions for a learning assessment theory

Universidade Estadual de Campinas . Faculdade de Educação Física

LEARNING A COMPLEX MOTOR SKILL FROM VIDEO AND POINT-LIGHT DEMONSTRATIONS

The aim of this Study was to compare the learning process of a highly complex ballet skill following demonstrations of point light and video models 16 participants divided into point light and video groups (ns = 8) performed 160 trials of a pirouette equally distributed in blocks of 20 trials alternating periods of demonstration and practice with a retention test a day later Measures of head and trunk oscillation coordination d1 parity from the model and movement time difference showed similarities between video and point light groups ballet experts evaluations indicated superiority of performance in the video over the point light group Results are discussed in terms of the task requirements of dissociation between head and trunk rotations focusing on the hypothesis of sufficiency and higher relevance of information contained in biological motion models applied to learning of complex motor skills

Parameter setting and statistical learning

Three main models of parameter setting have been proposed: the Variational model proposed by Yang (2002; 2004), the Structured Acquisition model endorsed by Baker (2001; 2005), and the Very Early Parameter Setting (VEPS) model advanced by Wexler (1998). The VEPS model contends that parameters are set early. The Variational model supposes that children employ statistical learning mechanisms to decide among competing parameter values, so this model anticipates delays in parameter setting when critical input is sparse, and gradual setting of parameters. On the Structured Acquisition model, delays occur because parameters form a hierarchy, with higher-level parameters set before lower-level parameters. Assuming that children freely choose the initial value, children sometimes will miss-set parameters. However when that happens, the input is expected to trigger a precipitous rise in one parameter value and a corresponding decline in the other value. We will point to the kind of child language data that is needed in order to adjudicate among these competing models.

A new graphic format to facilitate the understanding of technological innovation models: the seesaw of competitiveness

Many models exist in the literature to explain the success of technological innovation. However, no studies have been made regarding graphic formats representing the technological innovation models and their impact, or on the understanding of these models by non-specialists in technology management. Thus, the main objective of this paper is to propose a new graphic configuration to represent the technological innovation management. Based on the literature, the innovation model is presented in the traditional format. Next, the same model is designed in the graphic format - named `the see-saw of competitiveness` - showing the interfaces among the identified factors. The two graphic formats were compared by a group of graduate students in terms of the ease in understanding the conceptual model of innovation. The statistical analysis shows that the seesaw of competitiveness is preferred.

Special issue on methods and learning in functional MRI: Introductory remarks

This special issue represents a further exploration of some issues raised at a symposium entitled “Functional magnetic resonance imaging: From methods to madness” presented during the 15th annual Theoretical and Experimental Neuropsychology (TENNET XV) meeting in Montreal, Canada in June, 2004. The special issue’s theme is methods and learning in functional magnetic resonance imaging (fMRI), and it comprises 6 articles (3 reviews and 3 empirical studies). The first (Amaro and Barker) provides a beginners guide to fMRI and the BOLD effect (perhaps an alternative title might have been “fMRI for dummies”). While fMRI is now commonplace, there are still researchers who have yet to employ it as an experimental method and need some basic questions answered before they venture into new territory. This article should serve them well. A key issue of interest at the symposium was how fMRI could be used to elucidate cerebral mechanisms responsible for new learning. The next 4 articles address this directly, with the first (Little and Thulborn) an overview of data from fMRI studies of category-learning, and the second from the same laboratory (Little, Shin, Siscol, and Thulborn) an empirical investigation of changes in brain activity occurring across different stages of learning. While a role for medial temporal lobe (MTL) structures in episodic memory encoding has been acknowledged for some time, the different experimental tasks and stimuli employed across neuroimaging studies have not surprisingly produced conflicting data in terms of the precise subregion(s) involved. The next paper (Parsons, Haut, Lemieux, Moran, and Leach) addresses this by examining effects of stimulus modality during verbal memory encoding. Typically, BOLD fMRI studies of learning are conducted over short time scales, however, the fourth paper in this series (Olson, Rao, Moore, Wang, Detre, and Aguirre) describes an empirical investigation of learning occurring over a longer than usual period, achieving this by employing a relatively novel technique called perfusion fMRI. This technique shows considerable promise for future studies. The final article in this special issue (de Zubicaray) represents a departure from the more familiar cognitive neuroscience applications of fMRI, instead describing how neuroimaging studies might be conducted to both inform and constrain information processing models of cognition.

Single Intranasal Administration of 1-Methyl-4-Phenyl-1,2,3,6-Tetrahydropyridine in C57BL/6 Mice Models Early Preclinical Phase of Parkinson`s Disease

Many studies have shown that deficits in olfactory and cognitive functions precede the classical motor symptoms seen in Parkinson`s disease (PD) and that olfactory testing may contribute to the early diagnosis of this disorder. Although the primary cause of PD is still unknown, epidemiological studies have revealed that its incidence is increased in consequence of exposure to certain environmental toxins. In this study, most of the impairments presented by C57BL/6 mice infused with a single intranasal (i.n.) administration of the proneurotoxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (1 mg/nostril) were similar to those observed during the early phase of PD, when a moderate loss of nigral dopamine neurons results in olfactory and memory deficits with no major motor impairments. Such infusion decreased the levels of the enzyme tyrosine hydroxylase in the olfactory bulb, striatum, and substantia nigra by means of apoptotic mechanisms, reducing dopamine concentration in different brain structures such as olfactory bulb, striatum, and prefrontal cortex, but not in the hippocampus. These findings reinforce the notion that the olfactory system represents a particularly sensitive route for the transport of neurotoxins into the central nervous system that may be related to the etiology of PD. These results also provide new insights in experimental models of PD, indicating that the i.n. administration of MPTP represents a valuable mouse model for the study of the early stages of PD and for testing new therapeutic strategies to restore sensorial and cognitive processes in PD.

Enriched behavioral prediction equation and its impact on structured learning and the dynamic calculus

This theoretical note describes an expansion of the behavioral prediction equation, in line with the greater complexity encountered in models of structured learning theory (R. B. Cattell, 1996a). This presents learning theory with a vector substitute for the simpler scalar quantities by which traditional Pavlovian-Skinnerian models have hitherto been represented. Structured learning can be demonstrated by vector changes across a range of intrapersonal psychological variables (ability, personality, motivation, and state constructs). Its use with motivational dynamic trait measures (R. B. Cattell, 1985) should reveal new theoretical possibilities for scientifically monitoring change processes (dynamic calculus model; R. B. Cattell, 1996b), such as encountered within psycho therapeutic settings (R. B. Cattell, 1987). The enhanced behavioral prediction equation suggests that static conceptualizations of personality structure such as the Big Five model are less than optimal.

Input-driven Language Learning

Input-driven models provide an explicit and readily testable account of language learning. Although we share Ellis's view that the statistical structure of the linguistic environment is a crucial and, until recently, relatively neglected variable in language learning, we also recognize that the approach makes three assumptions about cognition and language learning that are not universally shared. The three assumptions concern (a) the language learner as an intuitive statistician, (b) the constraints on what constitute relevant surface cues, and (c) the redescription problem faced by any system that seeks to derive abstract grammatical relations from the frequency of co-occurring surface forms and functions. These are significant assumptions that must be established if input-driven models are to gain wider acceptance. We comment on these issues and briefly describe a distributed, instance-based approach that retains the key features of the input-driven account advocated by Ellis but that also addresses shortcomings of the current approaches.

LEARNING ENVIRONMENTS IN THE CONTEXT OF A MASTER DEGREE IN ANIMATION AND ILLUSTRATION

This paper aims to describe the processes of teaching illustration and animation, together, in the context of a masters degree program. In Portugal, until very recently, illustration and animation higher education courses, were very scarce and only provided by a few private universities, which offered separated programs - either illustration or animation. The MA in Illustration and Animation (MIA) based in the Instituto Politécnico do Cávado e Ave in Portugal, dared to join these two creative areas in a common learning model and is already starting it’s third edition with encouraging results and will be supported by the first international conference on illustration and animation (CONFIA). This masters program integrates several approaches and techniques (in illustration and animation) and integrates and encourages creative writing and critique writing. This paper describes the iterative process of construction, and implementation of the program as well as the results obtained on the initial years of existence in terms of pedagogic and learning conclusions. In summary, we aim to compare pedagogic models of animation or illustration teaching in higher education opposed to a more contemporary and multidisciplinary model approach that integrates the two - on an earlier stage - and allows them to be developed separately – on the second part of the program. This is based on the differences and specificities of animation (from classic techniques to 3D) and illustration (drawing the illustration) and the intersection area of these two subjects within the program structure focused on the students learning and competencies acquired to use in professional or authorial projects.