William Pentney

AAAI Conference 2008 Conference Paper

Structure Learning on Large Scale Common Sense Statistical Models of Human State

• William Pentney

Research has shown promise in the design of large scale common sense probabilistic models to infer human state from environmental sensor data. These models have made use of mined and preexisting common sense data and traditional probabilistic machine learning techniques to improve recognition of the state of everyday human life. In this paper, we demonstrate effective techniques for structure learning on graphical models designed for this domain, improving the SRCS system of (Pentney et al. 2006) by learning additional dependencies between variables. Because the models used for common sense reasoning typically involve a large number of variables, issues of scale arise in searching for additional dependencies; we discuss how we use data mining techniques to address this problem. We show experimentally that these techniques improve the accuracy of state prediction, and that, with a good prior model, the use of a common sense model with structure learning provides better prediction of unlabeled variables as well as labeled variables. The results also demonstrate that it is possible to collect new common sense information about daily life using such a statistical model and labeled data.

IJCAI Conference 2007 Conference Paper

• Shiaokai Wang
• William Pentney
• Ana-Maria Popescu
• Tanzeem Choudhury
• Matthai Philipose

Given sensors to detect object use, commonsense priors of object usage in activities can reduce the need for labeled data in learning activity models. It is often useful, however, to understand how an object is being used, i. e. , the action performed on it. We show how to add personal sensor data (e. g. , accelerometers) to obtain this detail, with little labeling and feature selection overhead. By synchronizing the personal sensor data with object-use data, it is possible to use easily specified commonsense models to minimize labeling overhead. Further, combining a generative common sense model of activity with a discriminative model of actions can automate feature selection. On observed activity data, automatically trained action classifiers give 40/85% precision/recall on 10 actions. Adding actions to pure object-use improves precision/recall from 76/85% to 81/90% over 12 activities.

AAAI Conference 2007 Conference Paper

Learning Large Scale Common Sense Models of Everyday Life

• William Pentney
• Jeff Bilmes

Recent work has shown promise in using large, publicly available, hand-contributed commonsense databases as joint models that can be used to infer human state from day-to-day sensor data. The parameters of these models are mined from the web. We show in this paper that learning these parameters using sensor data (with the mined parameters as priors) can improve performance of the models significantly. The primary challenge in learning is scale. Since the model comprises roughly 50, 000 irregularly connected nodes in each time slice, it is intractable either to completely label observed data manually or to compute the expected likelihood of even a single time slice. We show how to solve the resulting semisupervised learning problem by combining a variety of conventional approximation techniques and a novel technique for simplifying the model called context-based pruning. We show empirically that the learned model is substantially better at interpreting sensor data and an detailed analysis of how various techniques contribute to the performance.

AAAI Conference 2006 Conference Paper

Sensor-Based Understanding of Daily Life via Large-Scale Use of Common Sense

• William Pentney
• Shiaokai Wang

The use of large quantities of common sense has long been thought to be critical to the automated understanding of the world. To this end, various groups have collected repositories of common sense in machinereadable form. However, efforts to apply these large bodies of knowledge to enable correspondingly largescale sensor-based understanding of the world have been few. Challenges have included semantic gaps between facts in the repositories and phenomena detected by sensors, fragility of reasoning in the face of noise, incompleteness of repositories, and slowness of reasoning with these large repositories. We show how to address these problems with a combination of novel sensors, probabilistic representation, web-scale information retrieval and approximate reasoning. In particular, we show how to use the 50, 000-fact hand-entered Open- Mind Indoor Common Sense database to interpret sensor traces of day-to-day activities with 88% accuracy (which is easy) and 32/53% precision/recall (which is not).

AAAI Conference 2005 Conference Paper

Spectral Clustering of Biological Sequence Data

• William Pentney

In this paper, we apply spectral techniques to clustering biological sequence data that has proved more difficult to cluster effectively. For this purpose, we have to (1) extend spectral clustering algorithms to deal with asymmetric affinities, like the alignment scores used in the comparison of biological sequences, and (2) devise a hierarchical algorithm that can handle many clusters with imbalanced sizes robustly. We present an algorithm for clustering asymmetric affinity data, and demonstrate the performance of this algorithm at recovering the higher levels of the Structural Classification of Proteins (SCOP) on a data base of highly conserved subsequences.