Nonparametric discovery of learning patterns and autism subgroups from therapeutic data

Autism Spectrum Disorder (ASD) is growing at a staggering rate, but, little is known about the cause of this condition. Inferring learning patterns from therapeutic performance data, and subsequently clustering ASD children into subgroups, is important to understand this domain, and more importantly to inform evidence-based intervention. However, this data-driven task was difficult in the past due to insufficiency of data to perform reliable analysis. For the first time, using data from a recent application for early intervention in autism (TOBY Play pad), whose download count is now exceeding 4500, we present in this paper the automatic discovery of learning patterns across 32 skills in sensory, imitation and language. We use unsupervised learning methods for this task, but a notorious problem with existing methods is the correct specification of number of patterns in advance, which in our case is even more difficult due to complexity of the data. To this end, we appeal to recent Bayesian nonparametric methods, in particular the use of Bayesian Nonparametric Factor Analysis. This model uses Indian Buffet Process (IBP) as prior on a binary matrix of infinite columns to allocate groups of intervention skills to children. The optimal number of learning patterns as well as subgroup assignments are inferred automatically from data. Our experimental results follow an exploratory approach, present different newly discovered learning patterns. To provide quantitative results, we also report the clustering evaluation against K-means and Nonnegative matrix factorization (NMF). In addition to the novelty of this new problem, we were able to demonstrate the suitability of Bayesian nonparametric models over parametric rivals.

Semi-supervised learning guided by the modularity measure in complex networks

Semi-supervised learning techniques have gained increasing attention in the machine learning community, as a result of two main factors: (1) the available data is exponentially increasing; (2) the task of data labeling is cumbersome and expensive, involving human experts in the process. In this paper, we propose a network-based semi-supervised learning method inspired by the modularity greedy algorithm, which was originally applied for unsupervised learning. Changes have been made in the process of modularity maximization in a way to adapt the model to propagate labels throughout the network. Furthermore, a network reduction technique is introduced, as well as an extensive analysis of its impact on the network. Computer simulations are performed for artificial and real-world databases, providing a numerical quantitative basis for the performance of the proposed method.

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.

Learning by the Dendritic Prediction of Somatic Spiking

Recent modeling of spike-timing-dependent plasticity indicates that plasticity involves as a third factor a local dendritic potential, besides pre- and postsynaptic firing times. We present a simple compartmental neuron model together with a non-Hebbian, biologically plausible learning rule for dendritic synapses where plasticity is modulated by these three factors. In functional terms, the rule seeks to minimize discrepancies between somatic firings and a local dendritic potential. Such prediction errors can arise in our model from stochastic fluctuations as well as from synaptic input, which directly targets the soma. Depending on the nature of this direct input, our plasticity rule subserves supervised or unsupervised learning. When a reward signal modulates the learning rate, reinforcement learning results. Hence a single plasticity rule supports diverse learning paradigms.

Stability criteria for unsupervised temporal association networks

A biologically realizable, unsupervised learning rule is described for the online extraction of object features, suitable for solving a range of object recognition tasks. Alterations to the basic learning rule are proposed which allow the rule to better suit the parameters of a given input space. One negative consequence of such modifications is the potential for learning instability. The criteria for such instability are modeled using digital filtering techniques and predicted regions of stability and instability tested. The result is a family of learning rules which can be tailored to the specific environment, improving both convergence times and accuracy over the standard learning rule, while simultaneously insuring learning stability.

A new unsupervised convolutional neural network model for Chinese scene text detection

As one of the most popular deep learning models, convolution neural network (CNN) has achieved huge success in image information extraction. Traditionally CNN is trained by supervised learning method with labeled data and used as a classifier by adding a classification layer in the end. Its capability of extracting image features is largely limited due to the difficulty of setting up a large training dataset. In this paper, we propose a new unsupervised learning CNN model, which uses a so-called convolutional sparse auto-encoder (CSAE) algorithm pre-Train the CNN. Instead of using labeled natural images for CNN training, the CSAE algorithm can be used to train the CNN with unlabeled artificial images, which enables easy expansion of training data and unsupervised learning. The CSAE algorithm is especially designed for extracting complex features from specific objects such as Chinese characters. After the features of articficial images are extracted by the CSAE algorithm, the learned parameters are used to initialize the first CNN convolutional layer, and then the CNN model is fine-Trained by scene image patches with a linear classifier. The new CNN model is applied to Chinese scene text detection and is evaluated with a multilingual image dataset, which labels Chinese, English and numerals texts separately. More than 10% detection precision gain is observed over two CNN models.

Overview of the INEX 2010 XML mining track : clustering and classification of XML documents

The XML Document Mining track was launched for exploring two main ideas: (1) identifying key problems and new challenges of the emerging field of mining semi-structured documents, and (2) studying and assessing the potential of Machine Learning (ML) techniques for dealing with generic ML tasks in the structured domain, i.e., classification and clustering of semi-structured documents. This track has run for six editions during INEX 2005, 2006, 2007, 2008, 2009 and 2010. The first five editions have been summarized in previous editions and we focus here on the 2010 edition. INEX 2010 included two tasks in the XML Mining track: (1) unsupervised clustering task and (2) semi-supervised classification task where documents are organized in a graph. The clustering task requires the participants to group the documents into clusters without any knowledge of category labels using an unsupervised learning algorithm. On the other hand, the classification task requires the participants to label the documents in the dataset into known categories using a supervised learning algorithm and a training set. This report gives the details of clustering and classification tasks.

Modelling word meaning using efficient tensor representations

Models of word meaning, built from a corpus of text, have demonstrated success in emulating human performance on a number of cognitive tasks. Many of these models use geometric representations of words to store semantic associations between words. Often word order information is not captured in these models. The lack of structural information used by these models has been raised as a weakness when performing cognitive tasks. This paper presents an efficient tensor based approach to modelling word meaning that builds on recent attempts to encode word order information, while providing flexible methods for extracting task specific semantic information.

Robots move : bootstrapping the development of object representations using sensorimotor coordination

This paper is concerned with the unsupervised learning of object representations by fusing visual and motor information. The problem is posed for a mobile robot that develops its representations as it incrementally gathers data. The scenario is problematic as the robot only has limited information at each time step with which it must generate and update its representations. Object representations are refined as multiple instances of sensory data are presented; however, it is uncertain whether two data instances are synonymous with the same object. This process can easily diverge from stability. The premise of the presented work is that a robot's motor information instigates successful generation of visual representations. An understanding of self-motion enables a prediction to be made before performing an action, resulting in a stronger belief of data association. The system is implemented as a data-driven partially observable semi-Markov decision process. Object representations are formed as the process's hidden states and are coordinated with motor commands through state transitions. Experiments show the prediction process is essential in enabling the unsupervised learning method to converge to a solution - improving precision and recall over using sensory data alone.

Supervised latent dirichlet allocation models for efficient activity representation

Local spatio-temporal features with a Bag-of-visual words model is a popular approach used in human action recognition. Bag-of-features methods suffer from several challenges such as extracting appropriate appearance and motion features from videos, converting extracted features appropriate for classification and designing a suitable classification framework. In this paper we address the problem of efficiently representing the extracted features for classification to improve the overall performance. We introduce two generative supervised topic models, maximum entropy discrimination LDA (MedLDA) and class- specific simplex LDA (css-LDA), to encode the raw features suitable for discriminative SVM based classification. Unsupervised LDA models disconnect topic discovery from the classification task, hence yield poor results compared to the baseline Bag-of-words framework. On the other hand supervised LDA techniques learn the topic structure by considering the class labels and improve the recognition accuracy significantly. MedLDA maximizes likelihood and within class margins using max-margin techniques and yields a sparse highly discriminative topic structure; while in css-LDA separate class specific topics are learned instead of common set of topics across the entire dataset. In our representation first topics are learned and then each video is represented as a topic proportion vector, i.e. it can be comparable to a histogram of topics. Finally SVM classification is done on the learned topic proportion vector. We demonstrate the efficiency of the above two representation techniques through the experiments carried out in two popular datasets. Experimental results demonstrate significantly improved performance compared to the baseline Bag-of-features framework which uses kmeans to construct histogram of words from the feature vectors.

Feature selection and the concept of immediate neighborhood in the context of clustering techniques

The concept of feature selection in a nonparametric unsupervised learning environment is practically undeveloped because no true measure for the effectiveness of a feature exists in such an environment. The lack of a feature selection phase preceding the clustering process seriously affects the reliability of such learning. New concepts such as significant features, level of significance of features, and immediate neighborhood are introduced which result in meeting implicitly the need for feature slection in the context of clustering techniques.

Feature selection and the concept of immediate neighborhood in the context of clustering techniques

The concept of feature selection in a nonparametric unsupervised learning environment is practically undeveloped because no true measure for the effectiveness of a feature exists in such an environment. The lack of a feature selection phase preceding the clustering process seriously affects the reliability of such learning. New concepts such as significant features, level of significance of features, and immediate neighborhood are introduced which result in meeting implicitly the need for feature slection in the context of clustering techniques.

Characterization of Failure Modes in CFRP Composites - An ANN Approach

Carbon fiber reinforced polymer (CFRP) composite specimens with different thickness, geometry, and stacking sequences were subjected to fatigue spectrum loading in stages. Another set of specimens was subjected to static compression load. On-line acoustic Emission (AE) monitoring was carried out during these tests. Two artificial neural networks, Kohonen-self organizing feature map (KSOM), and multi-layer perceptron (MLP) have been developed for AE signal analysis. AE signals from specimens were clustered using the unsupervised learning KSOM. These clusters were correlated to the failure modes using available a priori information such as AE signal amplitude distributions, time of occurrence of signals, ultrasonic imaging, design of the laminates (stacking sequences, orientation of fibers), and AE parametric plots. Thereafter, AE signals generated from the rest of the specimens were classified by supervised learning MLP. The network developed is made suitable for on-line monitoring of AE signals in the presence of noise, which can be used for detection and identification of failure modes and their growth. The results indicate that the characteristics of AE signals from different failure modes in CFRP remain largely unaffected by the type of load, fiber orientation, and stacking sequences, they being representatives of the type of failure phenomena. The type of loading can have effect only on the extent of damage allowed before the specimens fail and hence on the number of AE signals during the test. The artificial neural networks (ANN) developed and the methods and procedures adopted show significant success in AE signal characterization under noisy environment (detection and identification of failure modes and their growth).

Wetlands of Greater Bangalore, India: Automatic Delineation through Pattern Classifiers

Wetlands are the most productive and biologically diverse but very fragile ecosystems. They are vulnerable to even small changes in their biotic and abiotic factors. In recent years, there has been concern over the continuous degradation of wetlands due to unplanned developmental activities. This necessitates inventorying, mapping, and monitoring of wetlands to implement sustainable management approaches. The principal objective of this work is to evolve a strategy to identify and monitor wetlands using temporal remote sensing (RS) data. Pattern classifiers were used to extract wetlands automatically from NIR bands of MODIS, Landsat MSS and Landsat TM remote sensing data. MODIS provided data for 2002 to 2007, while for 1973 and 1992 IR Bands of Landsat MSS and TM (79m and 30m spatial resolution) data were used. Principal components of IR bands of MODIS (250 m) were fused with IRS LISS-3 NIR (23.5 m). To extract wetlands, statistical unsupervised learning of IR bands for the respective temporal data was performed using Bayesian approach based on prior probability, mean and covariance. Temporal analysis of wetlands indicates a sharp decline of 58% in Greater Bangalore attributing to intense urbanization processes, evident from a 466% increase in built-up area from 1973 to 2007.

Experiments on three-dimensional wire-frame object recognition

Some experimental results on the recognition of three-dimensional wire-frame objects are presented. In order to overcome the limitations of a recent model, which employs radial basis functions-based neural networks, we have proposed a hybrid learning system for object recognition, featuring: an optimization strategy (simulated annealing) in order to avoid local minima of an energy functional; and an appropriate choice of centers of the units. Further, in an attempt to achieve improved generalization ability, and to reduce the time for training, we invoke the principle of self-organization which utilises an unsupervised learning algorithm.