For nearly 25 years my research group has investigated the use of domain knowledge, expressed in some version of mathematical logic, that is refined or exploited by numeric-based learning algorithms. These include what we called knowledge-based neural networks and knowledge-based support vector machines. I will cover the key ideas of these methods, as well as the behind-the-scenes motivations that lead to them. I will also describe why we switched from using the phrase 'prior knowledge' to using 'advice.' Finally, I will cover some of our recent work on fast learning and inference for Markov Logic Networks (which can be viewed as a knowledge-based graphical model).

We present a new weighted session similarity measure to capture the browsing interests of users in web usage profiles discovered from web log data. We base our similarity measure on the reasonable assumption that when users spend longer times on pages or revisit pages in the same session, then very likely, such pages are of greater interest to the user. The proposed similarity measure combines structural similarity with session-wise page significance. The latter, representing the degree of user interest, is computed using frequency and duration of a page access. Web usage profiles are generated using this similarity measure by applying a fuzzy clustering algorithm to web log data. For evaluating the effectiveness of the proposed measure, we adapt two model-based collaborative filtering algorithms for recommending pages. Experimental results show considerable improvement in overall performance of recommender systems as compared to use of other existing similarity measures.

Computational systems can use sociocultural knowledge to understand human behavior and interact with humans in more natural ways. However, such systems are limited by their reliance on hand-authored sociocultural knowledge and models. We introduce an approach to automatically learn robust, script-like sociocultural knowledge from crowdsourced narratives. Crowdsourcing, the use of anonymous human workers, provides an opportunity for rapidly acquir­ing a corpus of examples of situations that are highly specialized for our purpose yet sufficiently varied, from which we can learn a versatile script. We describe a semi-automated process by which we query human workers to write natural language narrative examples of a given situation and learn the set of events that can occur and the typical even ordering.

Recognizing social interests plays an important role of aiding human-computer interaction and human collaborative works. The recognition of social interest could be of great help to determine the smoothness of the interaction, which could be an indicator for group work performance and relationship. From socio-psychological theories, social engagement is the observable form of inner social interest, and represented as patterns of turn-taking and speech emotion during a face-to-face conversation. With these two kinds of features, a multi-layer learning structure is proposed to model the continuous trend of engagement. The level of engagement is classified into “high” and “low” two levels according to human-annotated score. In the result of assessing two-level engagemet, the highest accuracy of our model can reach 79.1%.

Public sensing is a new research area in the fields of wireless sensor networks and mobile computing. It leverages the mobile sensors and system resources readily available in mobile phones to execute sensing tasks. In order to plan, execute and adapt large-scale sensing tasks, applications need to query for the available resources, e.g. the density of certain sensors. We investigate how such information can be provided, and we propose a resource manager for public sensing. Our primary goal is to minimize the energy consumed by the mobile devices to make public sensing feasible without disturbing users. We propose a cluster-based protocol for collecting local views of the resource state using local ad-hoc communication since this is much more energy-efficient than long-range (e.g. cellular) communication. We compare our solution to a standard approach where mobile devices communicate their resource states using the cellular phone network. We show that 65% of the energy is saved and the communication load on the infrastructure is reduced by 90% while an average delivery ratio of 93% is retained.

Clustering ensemble mainly relies on the pairwise similarity to capture the consensus function. However, it usually considers each base clustering independently, and treats the similarity measure roughly with either 0 or 1. To address these two issues, we propose a coupled framework of clustering ensembles CCE, and exemplify it with the coupled version CCSPA for CSPA. Experiments demonstrate the superiority of CCSPA over baseline approaches in terms of the clustering accuracy.

Sound source localization and separation with permutation resolution are essential for achieving a computational auditory scene analysis system that can extract useful information from a mixture of various sounds. Because existing methods cope separately with these problems despite their mutual dependence, the overall result with these approaches can be degraded by any failure in one of these components. This paper presents a unified Bayesian framework to solve these problems simultaneously where localization and separation are regarded as a clustering problem. Experimental results confirm that our method outperforms state-of-the-art methods in terms of the separation quality with various setups including practical reverberant environments.

A key challenge in reducing the burden of cardiovascular disease is matching patients to treatments that are most appropriate for them. Different cardiac assessment tools have been developed to address this goal. Recent research has focused on heart rate motifs, i.e., short-term heart rate sequences that are over- or under-represented in long-term electrocardiogram (ECG) recordings of patients experiencing cardiovascular outcomes, which provide novel and valuable information for risk stratification. However, this approach can leverage only a small number of motifs for prediction and results in difficult to interpret models. We address these limitations by identifying latent structure in the large numbers of motifs found in long-term ECG recordings. In particular, we explore the application of topic models to heart rate time series to identify functional sets of heart rate sequences and to concisely describe patients using task-independent features for various cardiovascular outcomes. We evaluate the approach on a large collection of real-world ECG data, and investigate the performance of topic mixture features for the prediction of cardiovascular mortality. The topics provided an interpretable representation of the recordings and maintained valuable information for clinical assessment when compared with motif frequencies, even after accounting for commonly used clinical risk scores.

In many real applications, the input data are naturally expressed as tensors, such as virtual metrology in semiconductor manufacturing, face recognition and gait recognition in computer vision, etc. In this paper, we propose a general optimization framework for dealing with tensor inputs. Most existing methods for supervised tensor learning use only rank-one weight tensors in the linear model and cannot readily incorporate domain knowledge. In our framework, we obtain the weight tensor in a hierarchical way — we first approximate it by a low-rank tensor, and then estimate the low-rank approximation using the prior knowledge from various sources, e.g., different domain experts. This is motivated by wafer quality prediction in semiconductor manufacturing. Furthermore, we propose an effective algorithm named H-MOTE for solving this framework, which is guaranteed to converge. The time complexity of H-MOTE is linear with respect to the number of examples as well as the size of the weight tensor. Experimental results show the superiority of H-MOTE over state-of-the-art techniques on both synthetic and real data sets.

In this paper, we study the problem of large-scale Kernel Logistic Regression (KLR). A straightforward approach is to apply stochastic approximation to KLR. We refer to this approach as non-conservative online learning algorithm because it updates the kernel classifier after every received training example, leading to a dense classifier. To improve the sparsity of the KLR classifier, we propose two conservative online learning algorithms that update the classifier in a stochastic manner and generate sparse solutions. With appropriately designed updating strategies, our analysis shows that the two conservative algorithms enjoy similar theoretical guarantee as that of the non-conservative algorithm. Empirical studies on several benchmark data sets demonstrate that compared to batch-mode algorithms for KLR, the proposed conservative online learning algorithms are able to produce sparse KLR classifiers, and achieve similar classification accuracy but with significantly shorter training time. Furthermore, both the sparsity and classification accuracy of our methods are comparable to those of the online kernel SVM.