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.

Assessment of the resilience of coral reefs to natural and human disturbances by means of recruitment panels in Indo-Pacific

Indo-Pacific region encompasses about 75% of world's coral reefs, but hard coral cover in this region experienced a 32% region-wide decline since 1970s. This great change is primarily ascribable to natural and anthropogenic pressures, including climate change and human activities effects. Coral reef conservation requires management strategies oriented to maintain their diversity and the capacity to provide ecosystem goods and services. Coral reef resilience, i.e. the capacity to recover after disturbances, is critical to their long-term persistence. The aims of the present study were to design and to test field experiments intended to measure changes in recruitment processes, as a fundamental aspect of the coral reef resilience. Recruitment experiments, using artificial panels suspended in the water column, were carried out in two Indo-Pacific locations affected by different disturbances: a new mine in Bangka Island (Indonesia), and the increased sedimentation due to coastal dynamics in Vavvaru Island (Maldives). One (or more) putatively disturbed site(s) was selected to be tested against 3 randomly selected control sites. Panels’ arrangement simulates 2 proximities to living corals, i.e. the sources of propagules: few centimetres and 2 meters over. Panels were deployed simultaneously at each site and left submerged for about five months. Recruits were identified to the lowest possible taxonomic level and recruited assemblages were analysed in terms of percent cover. In general it was not possible to detect significant differences between the benthic assemblages recruited in disturbed and control sites. The high variability observed in recruits assemblages structure among control sites may be so large to mask the possible disturbance effects. Only few taxa showed possible effects of the disturb they undergo. The field tests have highlighted strengths and weaknesses of the proposed approach and, based on these results, some possible improvements were suggested.