No description available.

abstract: Smart home assistants are becoming a norm due to their ease-of-use. They employ spoken language as an interface, facilitating easy interaction with their users. Even with their obvious advantages, natural-language based interfaces are not prevalent outside the domain of home assistants. It is hard to adopt them for computer-controlled systems due to the numerous complexities involved with their implementation in varying fields. The main challenge is the grounding of natural language base terms into the underlying system's primitives. The existing systems that do use natural language interfaces are specific to one problem domain only. In this thesis, a domain-agnostic framework that creates natural language interfaces for computer-controlled systems has been developed by making the mapping between the language constructs and the system primitives customizable. The framework employs ontologies built using OWL (Web Ontology Language) for knowledge representation purposes and machine learning models for language processing tasks. It has been evaluated within a simulation environment consisting of objects and a robot. This environment has been deployed as a web application, providing anonymous user testing for evaluation, and generating training data for machine learning components. Performance evaluation has been done on metrics such as time taken for a task or the number of instructions given by the user to the robot to accomplish a task. Additionally, the framework has been used to create a natural language interface for a database system to demonstrate its domain independence. / Dissertation/Thesis / Masters Thesis Software Engineering 2020

In education and research, references play a key role. They give credit to prior works, and provide support for reviews, discussions, and arguments. The set of references attached to a publication can help describe that publication, can aid with its categorization and retrieval, can support bibliometric studies, and can guide interested readers and researchers. If suitably analyzed, that set can aid with the analysis of the publication itself, especially regarding all its citing passages. However, extracting and parsing references are difficult problems. One concern is that there are many styles of references, and identifying what style was employed is problematic, especially in heterogeneous collections of theses and dissertations, which cover many fields and disciplines, and where different styles may be used even in the same publication. We address these problems by drawing upon suitable knowledge found in the WWW. In particular, we use appropriate lists (e.g., of names, cities, and other types of entities). We use available information about the many reference styles found, in a type of reverse engineering. We use available references to guide machine learning. In particular, we research a two-stage classifier approach, with multi-class classification with respect to reference styles, and partially solve the problem of parsing surface representations of references. We describe empirical evidence for the effectiveness of our approach and plans for improvement of our method. / Ph. D.

In mobile computing systems, predicting the next location of a mobile wireless user has gained interest over the past decade. Location prediction may have a wide-range of application areas such as network load balancing, advertising and web page prefetching. In the literature, there exist many location predictors which are divided into two main classes: domain-independent and domain-specific. Song et al. compare the prediction accuracy of the domain-independent predictors from four major families, namely, Markov-based, compression-based, PPM and SPM predictors on Dartmouth&#039 / s campus-wide Wi-Fi mobility data. As a result, the low-order Markov predictors are found as the best predictor. In another work, Bayir et al. propose a domain-specific location predictor (LPMP) as the application of a framework used for discovering mobile cell phone user profiles. In this thesis, we evaluate LPMP and the best Markov predictor with Dartmouth&#039 / s campus-wide Wi-Fi mobility data in terms of accuracy. We also propose a simple method which improves the accuracy of LPMP slightly in the location prediction part of LPMP. Our results show that the accuracy of the best Markov predictor is better than that of LPMP in total. However, interestingly, LPMP yields more accurate results than the best Markov predictor does for the users with the low prediction accuracy.

No description available.

Title E cient Representations and Conversions of Planning Problems Author Daniel Toropila Department Department of Theoretical Computer Science and Mathematical Logic Supervisor prof. RNDr. Roman Barták, Ph.D. Abstract The e ciency of all types of planning systems is strongly dependent on the in- put formulation, the structure of which must be exploited in order to provide an improved e ciency. Hence, the state-variable representation (SAS+ ) has be- come the input of choice for many modern planners. As majority of planning problems is encoded using a classical representation, several techniques for trans- lation into SAS+ have been developed in the past. These techniques, however, ignore the instance-specific information of planning problems. Therefore, we in- troduce a novel algorithm for constructing SAS+ that fully utilizes the information from the goal and the initial state. By performing an exhaustive experimental evaluation we demonstrate that for many planning problems the novel approach generates a more e cient encoding, providing thus an improved solving time. Finally, we present an overview and performance evaluation of several constraint models based on SAS+ and finite-state automata, showing that they represent a competitive alternative in the category of constraint-based planners. Keywords...

Feature extraction is an extremely important pre-processing step to pattern recognition, and machine learning problems. This thesis highlights how one can best extract features from the data in an exhaustively online and purely adaptive manner. The solution to this problem is given for both labeled and unlabeled datasets, by presenting a number of novel on-line learning approaches. Specifically, the differential equation method for solving the generalized eigenvalue problem is used to derive a number of novel machine learning and feature extraction algorithms. The incremental eigen-solution method is used to derive a novel incremental extension of linear discriminant analysis (LDA). Further the proposed incremental version is combined with extreme learning machine (ELM) in which the ELM is used as a preprocessor before learning. In this first key contribution, the dynamic random expansion characteristic of ELM is combined with the proposed incremental LDA technique, and shown to offer a significant improvement in maximizing the discrimination between points in two different classes, while minimizing the distance within each class, in comparison with other standard state-of-the-art incremental and batch techniques. In the second contribution, the differential equation method for solving the generalized eigenvalue problem is used to derive a novel state-of-the-art purely incremental version of slow feature analysis (SLA) algorithm, termed the generalized eigenvalue based slow feature analysis (GENEIGSFA) technique. Further the time series expansion of echo state network (ESN) and radial basis functions (EBF) are used as a pre-processor before learning. In addition, the higher order derivatives are used as a smoothing constraint in the output signal. Finally, an online extension of the generalized eigenvalue problem, derived from James Stone’s criterion, is tested, evaluated and compared with the standard batch version of the slow feature analysis technique, to demonstrate its comparative effectiveness. In the third contribution, light-weight extensions of the statistical technique known as canonical correlation analysis (CCA) for both twinned and multiple data streams, are derived by using the same existing method of solving the generalized eigenvalue problem. Further the proposed method is enhanced by maximizing the covariance between data streams while simultaneously maximizing the rate of change of variances within each data stream. A recurrent set of connections used by ESN are used as a pre-processor between the inputs and the canonical projections in order to capture shared temporal information in two or more data streams. A solution to the problem of identifying a low dimensional manifold on a high dimensional dataspace is then presented in an incremental and adaptive manner. Finally, an online locally optimized extension of Laplacian Eigenmaps is derived termed the generalized incremental laplacian eigenmaps technique (GENILE). Apart from exploiting the benefit of the incremental nature of the proposed manifold based dimensionality reduction technique, most of the time the projections produced by this method are shown to produce a better classification accuracy in comparison with standard batch versions of these techniques - on both artificial and real datasets.

Semantic interoperability (SI) of a product model refers to automatic exchange of meaning associated with the product data, among applications/domains throughout the product development cycle. In the product development cycle, several applications (engineering design, industrial design, manufacturing, supply chain, marketing, maintenance etc.) and different engineering domains (mechanical, electrical, electronic etc.) come into play making the ability to exchange product data with semantics very significant. With product development happening in multiple locations with multiple tools/systems, SI between these systems/domains becomes important. The thesis presents a feature-based framework for shape model to address these SI issues when exchanging shape models. Problem of exchanging semantics associated with shape model to support the product lifecycle has been identified and explained. Different types of semantic interoperability issues pertaining to the shape model have been identified and classified. Features in a shape model can be associated with volume addition/subtraction to/from base-solid, deformation/modification of base-sheet/base surface, forming of material of constant thickness. The DIFF model has been extended to represent, classify and extract Free-Form Surface Features (FFSFs) and deformation features in a part model. FFSFs refer to features that modify a free-form surface. Deformation features are created in constant thickness part models, for example, deformation of material (as in sheet-metal parts) or forming of material (as in injection molded parts with constant thickness), also referred to as constant thickness features. Volumetric features covered in the DIFF model have been extended to classify and represent volumetric features based on relative variations of cross-section and PathCurve. Shape feature ontology is described based on unified feature taxonomy with definitions and labels of features as defined in the extended DIFF model. Features definitions are used as intermediate and unambiguous representation for shape features. The feature ontology is used to capture semantics of shape features. The proposed ontology enables reasoning to handle semantic equivalences between feature labels, and is used to map shape features from a source to target applications. Reasoning framework for identification of semantically equivalent feature labels and representations for the feature being exchanged across multiple applications is presented and discussed. This reasoning framework is used to associate multiple construction paths for a feature and associate applicable meanings from the ontology. Interface is provided to select feature label for a target application from the list of labels which are semantically equivalent for the feature being exchanged/mapped. Parameters for the selected feature label can be mapped from the DIFF representation; the feature can then be represented/constructed in the target application using the feature label and mapped parameters. This work shows that product model with feature information (feature labels and representations), as understood by the target application, can be exchanged and maintained in such a way that multiple applications can use the product information as their understandable labels and representations. Finally, the thesis concludes by summarizing the main contributions and outlining the scope for future work.