Massive Open Online Courses (MOOCs) have received widespread attention for their potential to scale higher education, with multiple platforms such as Coursera, edX and Udacity recently appearing. Despite their successes, a major problem faced by MOOCs is low completion rates. In this paper, we explore the accurate early identification of students who are at risk of not completing courses. We build predictive models weekly, over multiple offerings of a course. Furthermore, we envision student interventions that present meaningful probabilities of failure, enacted only for marginal students.To be effective, predicted probabilities must be both well-calibrated and smoothed across weeks.Based on logistic regression, we propose two transfer learning algorithms to trade-off smoothness and accuracy by adding a regularization term to minimize the difference of failure probabilities between consecutive weeks. Experimental results on two offerings of a Coursera MOOC establish the effectiveness of our algorithms.

Management of Parkinson's Disease (PD) could be improved significantly if reliable, objective information about fluctuations in disease severity can be obtained in ecologically valid surroundings such as the private home. Although automatic assessment in PD has been studied extensively, so far no approach has been devised that is useful for clinical practice. Analysis approaches common for the field lack the capability of exploiting data from realistic environments, which represents a major barrier towards practical assessment systems. The very unreliable and infrequent labelling of ambiguous, low resolution movement data collected in such environments represents a very challenging analysis setting, where advances would have significant societal impact in our ageing population. In this work we propose an assessment system that abides practical usability constraints and applies deep learning to differentiate disease state in data collected in naturalistic settings. Based on a large data-set collected from 34 people with PD we illustrate that deep learning outperforms other approaches in generalisation performance, despite the unreliable labelling characteristic for this problem setting, and how such systems could improve current clinical practice.

Mass digitization of historical documents is a challenging problem for optical character recognition (OCR) tools. Issues include noisy backgrounds and faded text due to aging, border/marginal noise, bleed-through, skewing, warping, as well as irregular fonts and page layouts. As a result, OCR tools often produce a large number of spurious bounding boxes (BBs) in addition to those that correspond to words in the document. This paper presents an iterative classification algorithm to automatically label BBs (i.e., as text or noise) based on their spatial distribution and geometry. The approach uses a rule-base classifier to generate initial text/noise labels for each BB, followed by an iterative classifier that refines the initial labels by incorporating local information to each BB, its spatial location, shape and size. When evaluated on a dataset containing over 72,000 manually-labeled BBs from 159 historical documents, the algorithm can classify BBs with 0.95 precision and 0.96 recall. Further evaluation on a collection of 6,775 documents with ground-truth transcriptions shows that the algorithm can also be used to predict document quality (0.7 correlation) and improve OCR transcriptions in 85% of the cases.

A user-adaptive information visualization system capable of learning models of users and the visualization tasks they perform could provide interventions optimized for helping specific users in specific task contexts. In this paper, we investigate the accuracy of predicting visualization tasks, user performance on tasks, and user traits from gaze data. We show that predictions made with a logistic regression model are significantly better than a baseline classifier, with particularly strong results for predicting task type and user performance. Furthermore, we compare classifiers built with interface-independent and interface-dependent features, and show that the interface-independent features are comparable or superior to interface-dependent ones. Finally, we discuss how the accuracy of predictive models is affected if they are trained with data from trials that had highlighting interventions added to the visualization.

Link prediction and multi-label learning on graphs are two important but challenging machine learning problems that have broad applications in diverse fields. Not only are the two problems inherently correlated and often appear concurrently, they are also exacerbated by incomplete data. We develop a novel algorithm to solve these two problems jointly under a unified framework, which helps reduce the impact of graph noise and benefits both tasks individually. We reduce multi-label learning problem into an additional link prediction task and solve both problems with marginalized denoising, which we co-regularize with Laplacian smoothing. This approach combines both learning tasks into a single convex objective function, which we optimize efficiently with iterative closed-form updates. The resulting approach performs significantly better than prior work on several important real-world applications with great consistency.

Clinical trial adaptation refers to any adjustment of the trial protocol after the onset of the trial. The main goal is to make the process of introducing new medical interventions to patients more efficient by reducing the cost and the time associated with evaluating their safety and efficacy. The principal question is how should adaptation be performed so as to minimize the chance of distorting the outcome of the trial. We propose a novel method for achieving this. Unlike previous work our approach focuses on trial adaptation by sample size adjustment. We adopt a recently proposed stratification framework based on collected auxiliary data and show that this information together with the primary measured variables can be used to make a probabilistically informed choice of the particular sub-group a sample should be removed from. Experiments on simulated data are used to illustrate the effectiveness of our method and its application in practice.

We consider the problem of embedding entities and relations of knowledge bases into low-dimensional continuous vector spaces (distributed representations). Unlike most existing approaches, which are primarily efficient for modelling pairwise relations between entities, we attempt to explicitly model both pairwise relations and long-range interactions between entities, by interpreting them as linear operators on the low-dimensional embeddings of the entities. Therefore, in this paper we introduces Path-Ranking to capture the long-range interactions of knowledge graph and at the same time preserve the pairwise relations of knowledge graph; we call it 'structured embedding via pairwise relation and long-range interactions' (referred to as SePLi). Comparing with the-state-of-the-art models, SePLi achieves better performances of embeddings.

A known limitation of many diagnosis algorithms is that the number of diagnoses they return can be very large. This raises the question of how to use such a large set of diagnoses. For example, presenting hundreds of diagnoses to a human operator (charged with repairing the system) is meaningless. In various settings, including decision support for a human operator and automated troubleshooting processes, it is sufficient to be able to answer a basic diagnostic question: is a given component faulty? We propose a way to aggregate an arbitrarily large set of diagnoses to return an estimate of the likelihood of a given component to be faulty. The resulting mapping of components to their likelihood of being faulty is called the system's health state. We propose two metrics for evaluating the accuracy of a health state and show that an accurate health state can be found without finding all diagnoses. An empirical study explores the question of how many diagnoses are needed to obtain an accurate enough health state, and a simple online stopping criterion is proposed.

Datalog-based systems often materialise all consequences of a datalog program and the data, allowing users' queries to be evaluated directly in the materialisation. This process, however, can be computationally intensive, so most systems update the materialisation incrementally when input data changes. We argue that existing solutions, such as the well-known Delete/Rederive (DRed) algorithm, can be inefficient in cases when facts have many alternate derivations. As a possible remedy, we propose a novel Backward/Forward (B/F) algorithm that tries to reduce the amount of work by a combination of backward and forward chaining. In our evaluation, the B/F algorithm was several orders of magnitude more efficient than the DRed algorithm on some inputs, and it was never significantly less efficient.

Analogies are a fundamental human reasoning pattern that relies on relational similarity. Understanding how analogies are formed facilitates the transfer of knowledge between contexts. The approach presented in this work focuses on obtaining precise interpretations of analogies. We leverage noisy semantic networks to answer and explain a wide spectrum of analogy questions. The core of our contribution, the Semantic Similarity Engine, consists of methods for extracting and comparing graph-contexts that reveal the relational parallelism that analogies are based on, while mitigating uncertainty in the semantic network. We demonstrate these methods in two tasks: answering multiple choice analogy questions and generating human readable analogy explanations. We evaluate our approach on two datasets totaling 600 analogy questions. Our results show reliable performance and low false-positive rate in question answering; human evaluators agreed with 96% of our analogy explanations.