Deep learning for computer vision: a comparison between convolutional neural networks and hierarchical temporal memories on object recognition tasks

In recent years, Deep Learning techniques have shown to perform well on a large variety of problems both in Computer Vision and Natural Language Processing, reaching and often surpassing the state of the art on many tasks. The rise of deep learning is also revolutionizing the entire field of Machine Learning and Pattern Recognition pushing forward the concepts of automatic feature extraction and unsupervised learning in general. However, despite the strong success both in science and business, deep learning has its own limitations. It is often questioned if such techniques are only some kind of brute-force statistical approaches and if they can only work in the context of High Performance Computing with tons of data. Another important question is whether they are really biologically inspired, as claimed in certain cases, and if they can scale well in terms of "intelligence". The dissertation is focused on trying to answer these key questions in the context of Computer Vision and, in particular, Object Recognition, a task that has been heavily revolutionized by recent advances in the field. Practically speaking, these answers are based on an exhaustive comparison between two, very different, deep learning techniques on the aforementioned task: Convolutional Neural Network (CNN) and Hierarchical Temporal memory (HTM). They stand for two different approaches and points of view within the big hat of deep learning and are the best choices to understand and point out strengths and weaknesses of each of them. CNN is considered one of the most classic and powerful supervised methods used today in machine learning and pattern recognition, especially in object recognition. CNNs are well received and accepted by the scientific community and are already deployed in large corporation like Google and Facebook for solving face recognition and image auto-tagging problems. HTM, on the other hand, is known as a new emerging paradigm and a new meanly-unsupervised method, that is more biologically inspired. It tries to gain more insights from the computational neuroscience community in order to incorporate concepts like time, context and attention during the learning process which are typical of the human brain. In the end, the thesis is supposed to prove that in certain cases, with a lower quantity of data, HTM can outperform CNN.

Engineering Concurrent and Event-driven Web Apps: An Agent-Oriented Approach based on the simpAL Language

Web is constantly evolving, thanks to the 2.0 transition, HTML5 new features and the coming of cloud-computing, the gap between Web and traditional desktop applications is tailing off. Web-apps are more and more widespread and bring several benefits compared to traditional ones. On the other hand reference technologies, JavaScript primarly, are not keeping pace, so a paradim shift is taking place in Web programming, and so many new languages and technologies are coming out. First objective of this thesis is to survey the reference and state-of-art technologies for client-side Web programming focusing in particular on what concerns concurrency and asynchronous programming. Taking into account the problems that affect existing technologies, we finally design simpAL-web, an innovative approach to tackle Web-apps development, based on the Agent-oriented programming abstraction and the simpAL language. == Versione in italiano: Il Web è in continua evoluzione, grazie alla transizione verso il 2.0, alle nuove funzionalità introdotte con HTML5 ed all’avvento del cloud-computing, il divario tra le applicazioni Web e quelle desktop tradizionali va assottigliandosi. Le Web-apps sono sempre più diffuse e presentano diversi vantaggi rispetto a quelle tradizionali. D’altra parte le tecnologie di riferimento, JavaScript in primis, non stanno tenendo il passo, motivo per cui la programmazione Web sta andando incontro ad un cambio di paradigma e nuovi linguaggi e tecnologie stanno spuntando sempre più numerosi. Primo obiettivo di questa tesi è di passare al vaglio le tecnologie di riferimento ed allo stato dell’arte per quel che riguarda la programmmazione Web client-side, porgendo particolare attenzione agli aspetti inerenti la concorrenza e la programmazione asincrona. Considerando i principali problemi di cui soffrono le attuali tecnologie passeremo infine alla progettazione di simpAL-web, un approccio innovativo con cui affrontare lo sviluppo di Web-apps basato sulla programmazione orientata agli Agenti e sul linguaggio simpAL.

Classificazione di oggetti in immagini attraverso il modello Bag of Visual Words

Generic object recognition is an important function of the human visual system and everybody finds it highly useful in their everyday life. For an artificial vision system it is a really hard, complex and challenging task because instances of the same object category can generate very different images, depending of different variables such as illumination conditions, the pose of an object, the viewpoint of the camera, partial occlusions, and unrelated background clutter. The purpose of this thesis is to develop a system that is able to classify objects in 2D images based on the context, and identify to which category the object belongs to. Given an image, the system can classify it and decide the correct categorie of the object. Furthermore the objective of this thesis is also to test the performance and the precision of different supervised Machine Learning algorithms in this specific task of object image categorization. Through different experiments the implemented application reveals good categorization performances despite the difficulty of the problem. However this project is open to future improvement; it is possible to implement new algorithms that has not been invented yet or using other techniques to extract features to make the system more reliable. This application can be installed inside an embedded system and after trained (performed outside the system), so it can become able to classify objects in a real-time. The information given from a 3D stereocamera, developed inside the department of Computer Engineering of the University of Bologna, can be used to improve the accuracy of the classification task. The idea is to segment a single object in a scene using the depth given from a stereocamera and in this way make the classification more accurate.

The Presence of Words in Visual Art from Surrealism to Conceptualism

Without a doubt, one of the biggest changes that affected XXth century art is the introduction of words into paintings and, in more recent years, in installations. For centuries, if words were part of a visual composition, they functioned as reference; strictly speaking, they were used as a guideline for a better perception of the subject represented. With the developments of the XXth century, words became a very important part of the visual composition, and sometimes embodied the composition itself. About this topic, American art critic and collector Russell Bowman wrote an interesting article called Words and images: A persistent paradox, in which he examines the American and the European art of the XXth century in almost its entirety, dividing it up in six “categories of intention”. The aforementioned categories are not based on the art history timeline, but on the role that language played for specific artists or movements. Taking inspiration from Bowman's article, this paper is structured in three chapters, respectively: words in juxtaposition and free association, words as means of exploration of language structures, and words as means for political and personal messages. The purpose of this paper is therefore to reflect on the role of language in contemporary art and on the way it has changed from artist to artist.