Exploiting Learning Technologies for World Wide Web Agents

Postprint

Learning Mechanisms for Information Filtering Agents

Postprint

Intelligent Agents = Learning Agents

Postprint

Learning Meta-Descriptions of the FOAF Network

Postprint

Distributed Learning: An Agent-Based Approach to Data-Mining

Peer reviewed

An Empirical Investigation of Learning From the Semantic Web

Peer reviewed

Detection and Classification of Whale Acoustic Signals

This dissertation focuses on two vital challenges in relation to whale acoustic signals: detection and classification.

In detection, we evaluated the influence of the uncertain ocean environment on the spectrogram-based detector, and derived the likelihood ratio of the proposed Short Time Fourier Transform detector. Experimental results showed that the proposed detector outperforms detectors based on the spectrogram. The proposed detector is more sensitive to environmental changes because it includes phase information.

In classification, our focus is on finding a robust and sparse representation of whale vocalizations. Because whale vocalizations can be modeled as polynomial phase signals, we can represent the whale calls by their polynomial phase coefficients. In this dissertation, we used the Weyl transform to capture chirp rate information, and used a two dimensional feature set to represent whale vocalizations globally. Experimental results showed that our Weyl feature set outperforms chirplet coefficients and MFCC (Mel Frequency Cepstral Coefficients) when applied to our collected data.

Since whale vocalizations can be represented by polynomial phase coefficients, it is plausible that the signals lie on a manifold parameterized by these coefficients. We also studied the intrinsic structure of high dimensional whale data by exploiting its geometry. Experimental results showed that nonlinear mappings such as Laplacian Eigenmap and ISOMAP outperform linear mappings such as PCA and MDS, suggesting that the whale acoustic data is nonlinear.

We also explored deep learning algorithms on whale acoustic data. We built each layer as convolutions with either a PCA filter bank (PCANet) or a DCT filter bank (DCTNet). With the DCT filter bank, each layer has different a time-frequency scale representation, and from this, one can extract different physical information. Experimental results showed that our PCANet and DCTNet achieve high classification rate on the whale vocalization data set. The word error rate of the DCTNet feature is similar to the MFSC in speech recognition tasks, suggesting that the convolutional network is able to reveal acoustic content of speech signals.

Can a Broader Education Narrow the Gap? Evidence on Non-Academic Features of Schooling

Empirical studies of education programs and systems, by nature, rely upon use of student outcomes that are measurable. Often, these come in the form of test scores. However, in light of growing evidence about the long-run importance of other student skills and behaviors, the time has come for a broader approach to evaluating education. This dissertation undertakes experimental, quasi-experimental, and descriptive analyses to examine social, behavioral, and health-related mechanisms of the educational process. My overarching research question is simply, which inside- and outside-the-classroom features of schools and educational interventions are most beneficial to students in the long term? Furthermore, how can we apply this evidence toward informing policy that could effectively reduce stark social, educational, and economic inequalities?

The first study of three assesses mechanisms by which the Fast Track project, a randomized intervention in the early 1990s for high-risk children in four communities (Durham, NC; Nashville, TN; rural PA; and Seattle, WA), reduced delinquency, arrests, and health and mental health service utilization in adolescence through young adulthood (ages 12-20). A decomposition of treatment effects indicates that about a third of Fast Track’s impact on later crime outcomes can be accounted for by improvements in social and self-regulation skills during childhood (ages 6-11), such as prosocial behavior, emotion regulation and problem solving. These skills proved less valuable for the prevention of mental and physical health problems.

The second study contributes new evidence on how non-instructional investments – such as increased spending on school social workers, guidance counselors, and health services – affect multiple aspects of student performance and well-being. Merging several administrative data sources spanning the 1996-2013 school years in North Carolina, I use an instrumental variables approach to estimate the extent to which local expenditure shifts affect students’ academic and behavioral outcomes. My findings indicate that exogenous increases in spending on non-instructional services not only reduce student absenteeism and disciplinary problems (important predictors of long-term outcomes) but also significantly raise student achievement, in similar magnitude to corresponding increases in instructional spending. Furthermore, subgroup analyses suggest that investments in student support personnel such as social workers, health services, and guidance counselors, in schools with concentrated low-income student populations could go a long way toward closing socioeconomic achievement gaps.

The third study examines individual pathways that lead to high school graduation or dropout. It employs a variety of machine learning techniques, including decision trees, random forests with bagging and boosting, and support vector machines, to predict student dropout using longitudinal administrative data from North Carolina. I consider a large set of predictor measures from grades three through eight including academic achievement, behavioral indicators, and background characteristics. My findings indicate that the most important predictors include eighth grade absences, math scores, and age-for-grade as well as early reading scores. Support vector classification (with a high cost parameter and low gamma parameter) predicts high school dropout with the highest overall validity in the testing dataset at 90.1 percent followed by decision trees with boosting and interaction terms at 89.5 percent.

Bayesian Learning with Dependency Structures via Latent Factors, Mixtures, and Copulas

Bayesian methods offer a flexible and convenient probabilistic learning framework to extract interpretable knowledge from complex and structured data. Such methods can characterize dependencies among multiple levels of hidden variables and share statistical strength across heterogeneous sources. In the first part of this dissertation, we develop two dependent variational inference methods for full posterior approximation in non-conjugate Bayesian models through hierarchical mixture- and copula-based variational proposals, respectively. The proposed methods move beyond the widely used factorized approximation to the posterior and provide generic applicability to a broad class of probabilistic models with minimal model-specific derivations. In the second part of this dissertation, we design probabilistic graphical models to accommodate multimodal data, describe dynamical behaviors and account for task heterogeneity. In particular, the sparse latent factor model is able to reveal common low-dimensional structures from high-dimensional data. We demonstrate the effectiveness of the proposed statistical learning methods on both synthetic and real-world data.

Computational Methods for Identifying Hypertrophic Cardiomyopathy using 12-lead ECG

Hypertrophic cardiomyopathy (HCM) is a cardiovascular disease where the heart muscle is partially thickened and blood flow is - potentially fatally - obstructed. It is one of the leading causes of sudden cardiac death in young people. Electrocardiography (ECG) and Echocardiography (Echo) are the standard tests for identifying HCM and other cardiac abnormalities. The American Heart Association has recommended using a pre-participation questionnaire for young athletes instead of ECG or Echo tests due to considerations of cost and time involved in interpreting the results of these tests by an expert cardiologist. Initially we set out to develop a classifier for automated prediction of young athletes’ heart conditions based on the answers to the questionnaire. Classification results and further in-depth analysis using computational and statistical methods indicated significant shortcomings of the questionnaire in predicting cardiac abnormalities. Automated methods for analyzing ECG signals can help reduce cost and save time in the pre-participation screening process by detecting HCM and other cardiac abnormalities. Therefore, the main goal of this dissertation work is to identify HCM through computational analysis of 12-lead ECG. ECG signals recorded on one or two leads have been analyzed in the past for classifying individual heartbeats into different types of arrhythmia as annotated primarily in the MIT-BIH database. In contrast, we classify complete sequences of 12-lead ECGs to assign patients into two groups: HCM vs. non-HCM. The challenges and issues we address include missing ECG waves in one or more leads and the dimensionality of a large feature-set. We address these by proposing imputation and feature-selection methods. We develop heartbeat-classifiers by employing Random Forests and Support Vector Machines, and propose a method to classify full 12-lead ECGs based on the proportion of heartbeats classified as HCM. The results from our experiments show that the classifiers developed using our methods perform well in identifying HCM. Thus the two contributions of this thesis are the utilization of computational and statistical methods for discovering shortcomings in a current screening procedure and the development of methods to identify HCM through computational analysis of 12-lead ECG signals.

Data Mining for Network Intrusion Detection : A comparison of data mining algorithms and an analysis of relevant features for detecting cyber-attacks

Data mining can be defined as the extraction of implicit, previously un-known, and potentially useful information from data. Numerous re-searchers have been developing security technology and exploring new methods to detect cyber-attacks with the DARPA 1998 dataset for Intrusion Detection and the modified versions of this dataset KDDCup99 and NSL-KDD, but until now no one have examined the performance of the Top 10 data mining algorithms selected by experts in data mining. The compared classification learning algorithms in this thesis are: C4.5, CART, k-NN and Naïve Bayes. The performance of these algorithms are compared with accuracy, error rate and average cost on modified versions of NSL-KDD train and test dataset where the instances are classified into normal and four cyber-attack categories: DoS, Probing, R2L and U2R. Additionally the most important features to detect cyber-attacks in all categories and in each category are evaluated with Weka’s Attribute Evaluator and ranked according to Information Gain. The results show that the classification algorithm with best performance on the dataset is the k-NN algorithm. The most important features to detect cyber-attacks are basic features such as the number of seconds of a network connection, the protocol used for the connection, the network service used, normal or error status of the connection and the number of data bytes sent. The most important features to detect DoS, Probing and R2L attacks are basic features and the least important features are content features. Unlike U2R attacks, where the content features are the most important features to detect attacks.

Bewegungsklassifikation mithilfe mobiler Sensoren

In diesem Beitrag wird eine neue Methode zur Analyse des manuellen Kommissionierprozesses vorgestellt, mit der u. a. die Kommissionierzeitanteile automatisch erfasst werden können. Diese Methode basiert auf einer sensorgestützten Bewegungsklassifikation, wie sie bspw. im Sport oder in der Medizin Anwendung findet. Dabei werden mobile Sensoren genutzt, die fortlaufend Messwerte wie z. B. die Beschleunigung oder die Drehgeschwindigkeit des Kommissionierers aufzeichnen. Auf Basis dieser Daten können Informationen über die ausgeführten Bewegungen und insbesondere über die durchlaufenen Bewegungszustände gewonnen werden. Dieser Ansatz wird im vorliegenden Beitrag auf die Kommissionierung übertragen. Dazu werden zunächst Klassen relevanter Bewegungen identifiziert und anschließend mit Verfahren aus dem maschinellen Lernen verarbeitet. Die Klassifikation erfolgt nach dem Prinzip des überwachten Lernens. Dabei werden durchschnittliche Erkennungsraten von bis zu 78,94 Prozent erzielt.

Justified assessments of service provider reputation

Abstract Reputation, influenced by ratings from past clients, is crucial for providers competing for custom. For new providers with less track record, a few negative ratings can harm their chances of growing. In the JASPR project, we aim to look at how to ensure automated reputation assessments are justified and informative. Even an honest balanced review of a service provision may still be an unreliable predictor of future performance if the circumstances differ. For example, a service may have previously relied on different sub-providers to now, or been affected by season-specific weather events. A common way to ameliorate the ratings that may not reflect future performance is by weighting by recency. We argue that better results are obtained by querying provenance records on how services are provided for the circumstances of provision, to determine the significance of past interactions. Informed by case studies in global logistics, taxi hire, and courtesy car leasing, we are going on to explore the generation of explanations for reputation assessments, which can be valuable both for clients and for providers wishing to improve their match to the market, and applying machine learning to predict aspects of service provision which may influence decisions on the appropriateness of a provider. In this talk, I will give an overview of the research conducted and planned on JASPR. Speaker Biography Dr Simon Miles Simon Miles is a Reader in Computer Science at King's College London, UK, and head of the Agents and Intelligent Systems group. He conducts research in the areas of normative systems, data provenance, and medical informatics at King's, and has published widely and manages a number of research projects in these areas. He was previously a researcher at the University of Southampton after graduating from his PhD at Warwick. He has twice been an organising committee member for the Autonomous Agents and Multi-Agent Systems conference series, and was a member of the W3C working group which published standards on interoperable provenance data in 2013.

ImageLearn - Decoding Britain's Landscape

Abstract Ordnance Survey, our national mapping organisation, collects vast amounts of high-resolution aerial imagery covering the entirety of the country. Currently, photogrammetrists and surveyors use this to manually capture real-world objects and characteristics for a relatively small number of features. Arguably, the vast archive of imagery that we have obtained portraying the whole of Great Britain is highly underutilised and could be ‘mined’ for much more information. Over the last year the ImageLearn project has investigated the potential of "representation learning" to automatically extract relevant features from aerial imagery. Representation learning is a form of data-mining in which the feature-extractors are learned using machine-learning techniques, rather than being manually defined. At the beginning of the project we conjectured that representations learned could help with processes such as object detection and identification, change detection and social landscape regionalisation of Britain. This seminar will give an overview of the project and highlight some of our research results.

Automatically Evolving Rotation-invariant Texture Image Descriptors by Genetic Programming

In computer vision, training a model that performs classification effectively is highly dependent on the extracted features, and the number of training instances. Conventionally, feature detection and extraction are performed by a domain-expert who, in many cases, is expensive to employ and hard to find. Therefore, image descriptors have emerged to automate these tasks. However, designing an image descriptor still requires domain-expert intervention. Moreover, the majority of machine learning algorithms require a large number of training examples to perform well. However, labelled data is not always available or easy to acquire, and dealing with a large dataset can dramatically slow down the training process. In this paper, we propose a novel Genetic Programming based method that automatically synthesises a descriptor using only two training instances per class. The proposed method combines arithmetic operators to evolve a model that takes an image and generates a feature vector. The performance of the proposed method is assessed using six datasets for texture classification with different degrees of rotation, and is compared with seven domain-expert designed descriptors. The results show that the proposed method is robust to rotation, and has significantly outperformed, or achieved a comparable performance to, the baseline methods.