The aim of this Master's thesis is to design methods of active learning and to experiment with datasets of historical documents. A large and diverse dataset IMPACT of more than one million lines is used for experiments. I am using neural networks to check the readability of lines and correctness of their annotations. Firstly, I compare architectures of convolutional and recurrent neural networks with bidirectional LSTM layer. Next, I study different ways of learning neural networks using methods of active learning. Mainly I use active learning to adapt neural networks to documents that the neural networks do not have in the original training dataset. Active learning is thus used for picking appropriate adaptation data. Convolutional neural networks achieve 98.6\% accuracy, recurrent neural networks achieve 99.5\% accuracy. Active learning decreases error by 26\% compared to random pick of adaptations data.

This work deals with automatic chord recognition using neural networks. The problem was separated into two subproblems. The first subproblem aims to experimental finding of most suitable solution for a acoustic model and the second one aims to experimental finding of most suitable solution for a language model. The problem was solved by iterative method. First a suboptimal solution of the first subproblem was found and then the second one. A total of 19 acoustic and 12 language models were made. Ten training datasets was created for acoustic models and three for language models. In total, over 200 models were trained. The best results were achieved on acoustic models represented by convolutional networks together with language models represented by recurent networks with LSTM modules.

Tato práce se zabývá automatickým generovaním popisů obrázků s využitím několika druhů neuronových sítí. Práce je založena na článcích z MS COCO Captioning Challenge 2015 a znakových jazykových modelech, popularizovaných A. Karpathym. Navržený model je kombinací konvoluční a rekurentní neuronové sítě s architekturou kodér--dekodér. Vektor reprezentující zakódovaný obrázek je předáván jazykovému modelu jako hodnoty paměti LSTM vrstev v síti. Práce zkoumá, na jaké úrovni je model s takto jednoduchou architekturou schopen popisovat obrázky a jak si stojí v porovnání s ostatními současnými modely. Jedním ze závěrů práce je, že navržená architektura není dostatečná pro jakýkoli popis obrázků.

In the current deep learning era, convolutional neural networks are commonly used as a backbone of systems that process images or videos. A lot of existing neural network architectures are however needlessly overparameterized and their performance can be closely matched by an alternative that uses much smaller amount of parameters. Our aim is to design a method that is able to find such alternative(s) for a given convolutional architecture. We propose a general scheme for architecture reduction and evaluate three algorithms that search for the op- timal smaller architecture. We do multiple experiments with ResNet and Wide ResNet architectures as the base using CIFAR-10 dataset. The best method is able to reduce the number of parameters by 75-85% without any loss in accuracy even in these already quite efficient architectures. 1

No description available.

Image classification in Machine Learning encompasses the task of identification of objects in an image. The technique has applications in various areas such as e-commerce, social media and security surveillance. In this report the author explores the impact of using t-Distributed Stochastic Neighbor Embedding (t-SNE) on data as a preprocessing step when classifying multiple classes of clothing with a state-of-the-art Deep Convolutional Neural Network (DCNN). The t-SNE algorithm uses dimensionality reduction and groups similar objects close to each other in three-dimensional space. Extracting this information in the form of a positional coordinate gives us a new parameter which could help with the classification process since the features it uses can be different from that of the DCNN. Therefore, three slightly different DCNN models receives different input and are compared. The first benchmark model only receives pixel values, the second and third receive pixel values together with the positional coordinates from the t-SNE preprocessing for each data point, but with different hyperparameter values in the preprocessing step. The Fashion-MNIST dataset used contains 10 different clothing classes which are normalized and gray-scaled for easeof-use. The dataset contains 70.000 images in total. Results show minimum change in classification accuracy in the case of using a low-density map with higher learning rate as the data size increases, while a more dense map and lower learning rate performs a significant increase in accuracy of 4.4% when using a small data set. This is evidence for the fact that the method can be used to boost results when data is limited. / Bildklassificering i maskinlärning innefattar uppgiften att identifiera objekt i en bild. Tekniken har applikationer inom olika områden så som e-handel, sociala medier och säkerhetsövervakning. I denna rapport undersöker författaren effekten av att användat-Distributed Stochastic Neighbour Embedding (t-SNE) på data som ett förbehandlingssteg vid klassificering av flera klasser av kläder med ett state-of-the-art Deep Convolutio-nal Neural Network (DCNN). t-SNE-algoritmen använder dimensioneringsreduktion och grupperar liknande objekt nära varandra i tredimensionellt utrymme. Att extrahera denna information i form av en positionskoordinat ger oss en ny parameter som kan hjälpa till med klassificeringsprocessen eftersom funktionerna som den använder kan skilja sig från DCNN-modelen. Tre olika DCNN-modeller får olika in-data och jämförs därefter. Den första referensmodellen mottar endast pixelvärden, det andra och det tredje motar pixelvärden tillsammans med positionskoordinaterna från t-SNE-förbehandlingen för varje datapunkt men med olika hyperparametervärden i förbehandlingssteget. I studien används Fashion-MNIST datasetet som innehåller 10 olika klädklasser som är normaliserade och gråskalade för enkel användning. Datasetet innehåller totalt 70.000 bilder. Resultaten visar minst förändring i klassificeringsnoggrannheten vid användning av en låg densitets karta med högre inlärningsgrad allt eftersom datastorleken ökar, medan en mer tät karta och lägre inlärningsgrad uppnår en signifikant ökad noggrannhet på 4.4% när man använder en liten datamängd. Detta är bevis på att metoden kan användas för att öka klassificeringsresultaten när datamängden är begränsad.

Against the theoretical background of evolutionary behavioral economics, this project analyzes bounded rationality and adaptive behaviour in organizational settings characterized by complexity and persistent uncertainty. In particular, drawing upon the standard NK model, two laboratory experiments investigate individual and collective decision-making in combinatorial problems of resource allocation featuring multiple dimensions and various levels of complexity. In the first study, investment horizons of different length are employed to induce a near or distant future temporal orientation, in order to assess the effects of complexity and time horizon on performance and search behaviour, examine the presence of a temporal midpoint heuristic, and inspect the moderating effects of deadline proximity on the performance-risk relationship. This is relevant for organizational science because the passage of time is essential to articulate many strategic practices, such as assessing progress, scheduling and coordinating task-related activities, discerning the processual dynamics of how these activities emerge, develop, and terminate, or interpreting retrospected, current, and anticipated events. A greater or lesser amount of time reflects then a greater or lesser provision of resources, thereby representing a constraint that can greatly affect the ability to maintain a competitive advantage or ensure organizational survival. In the second study, the accuracy of the imitative process is varied to induce a flawless or flawed information diffusion system and, congruently, an efficient or inefficient communication network, in order to assess the effects of complexity and parallel problem-solving on autonomous search behaviour, clarify the core drivers of imitative behaviour, control for the degree of strategic diversity under different communication networks, and evaluate individual as well as collective performance conditional to the interaction between the levels of complexity and the modalities of parallel problem-solving. This is relevant for organizational science because imitating the practices of high-performing actors is one of the key strategies employed by organizations to solve complex problems and improve their performance, thereby representing a major part of the competitive process. The project is intended to contribute grounding individual and collective behaviour in a more psychologically and socially informed decision-making, with a view to further the research agenda of behavioral strategy and sustain the paradigm shift towards an evolutionary-complexity approach to real economic structures.

Language is the principal medium for ideas, while dialogue is the most natural and effective way for humans to interact with and access information from machines. Natural language generation (NLG) is a critical component of spoken dialogue and it has a significant impact on usability and perceived quality. Many commonly used NLG systems employ rules and heuristics, which tend to generate inflexible and stylised responses without the natural variation of human language. However, the frequent repetition of identical output forms can quickly make dialogue become tedious for most real-world users. Additionally, these rules and heuristics are not scalable and hence not trivially extensible to other domains or languages. A statistical approach to language generation can learn language decisions directly from data without relying on hand-coded rules or heuristics, which brings scalability and flexibility to NLG. Statistical models also provide an opportunity to learn in-domain human colloquialisms and cross-domain model adaptations. A robust and quasi-supervised NLG model is proposed in this thesis. The model leverages a Recurrent Neural Network (RNN)-based surface realiser and a gating mechanism applied to input semantics. The model is motivated by the Long-Short Term Memory (LSTM) network. The RNN-based surface realiser and gating mechanism use a neural network to learn end-to-end language generation decisions from input dialogue act and sentence pairs; it also integrates sentence planning and surface realisation into a single optimisation problem. The single optimisation not only bypasses the costly intermediate linguistic annotations but also generates more natural and human-like responses. Furthermore, a domain adaptation study shows that the proposed model can be readily adapted and extended to new dialogue domains via a proposed recipe. Continuing the success of end-to-end learning, the second part of the thesis speculates on building an end-to-end dialogue system by framing it as a conditional generation problem. The proposed model encapsulates a belief tracker with a minimal state representation and a generator that takes the dialogue context to produce responses. These features suggest comprehension and fast learning. The proposed model is capable of understanding requests and accomplishing tasks after training on only a few hundred human-human dialogues. A complementary Wizard-of-Oz data collection method is also introduced to facilitate the collection of human-human conversations from online workers. The results demonstrate that the proposed model can talk to human judges naturally, without any difficulty, for a sample application domain. In addition, the results also suggest that the introduction of a stochastic latent variable can help the system model intrinsic variation in communicative intention much better.

In the past few years, complex neural networks have achieved state of the art results in image classification. However, training these models requires large amounts of labelled data. Whereas unlabelled images are often readily available in large quantities, obtaining l abels takes considerable human effort. Active learning reduces the required labelling effort by selecting the most informative instances to label. The most popular active learning query strategy framework, uncertainty sampling, uses uncertainty estimates of the model being trained to select instances for labelling. However, modern classification neural networks often do not provide good uncertainty estimates. Baye sian neural networks model uncertainties over model parameters, which can be used to obtain uncertainties over model predictions. Exact Bayesian inference is intractable for neural networks, however several approximate methods have been proposed. We experiment with three such methods using various uncertainty sampling active learning query strategies.

Inverse, multidimensional input-output flow mapping is very important for use of valves in precision motion control applications. Due to the highly nonlinear characteristic and uncertain model structure of the cartridge valves, it is hard to formulate the modelling of their flow mappings into simple parameter estimation problems. This contribution conducts a comprehensive analysis and validation of three- and four-dimensional input-output-mapping approaches for a proportional pilot operated seat valves. Therefore, a virtual and a physical test-rig setup are utilized for initial measurement, implementation and assessment. After modeling and validating the valve under consideration, as a function of flow, pressure and temperature different mapping methods are investigated. More specifically, state of the art approaches, deep-learning methods and a newly developed approach (extPoly) are examined. Especially ANNs and Polynomials show reasonable approximation results even for more than two inputs. However, the results are strongly dependent on the structure and distribution of the input data points. Besides identification effort, the invertibility was investigated.