Fear of increasing prices and concern about climate change are motivating residential power conservation efforts. We investigate the effectiveness of several unsupervised disaggregation methods on low frequency power measurements collected in real homes. Specifically, we consider variants of the factorial hidden Markov model. Our results indicate that a conditional factorial hidden semi-Markov model, which integrates additional features related to when and how appliances are used in the home and more accurately represents the power use of individual appliances, outperforms the other unsupervised disaggregation methods. Our results show that unsupervised techniques can provide perappliance power usage information in a non-invasive manner, which is ideal for enabling power conservation efforts.

Topic models, which factor each document into different topics and represent each topic as a distribution of terms, have been widely and successfully used to better understand collections of text documents. However, documents are also associated with further information, such as the set of real-world entities mentioned in them. For example, news articles are usually related to several people, organizations, countries or locations. Since those associated entities carry rich information, it is highly desirable to build more expressive, entity-based topic models, which can capture the term distributions for each topic, each entity, as well as each topic-entity pair. In this paper, we therefore introduce a novel Entity Topic Model (ETM) for documents that are associated with a set of entities. ETM not only models the generative process of a term given its topic and entity information, but also models the correlation of entity term distributions and topic term distributions. A Gibbs sampling-based algorithm is proposed to learn the model. Experiments on real datasets demonstrate the effectiveness of our approach over several state-of-the-art baselines.

Machine learning techniques are widely employed for automatic assertion generation in hardware verification. Our previous method [16] uses a decision tree based approach for mining assertions and does not have design coverage related feedback. The assertions are unaware of the design, over-constrained and have low expressiveness. We introduce a coverage guided mining approach for mining assertions from simulation traces. Our approach combines association rule learning, greedy set covering and formal verification. It circumvents the exhaustive rule generation of association mining using coverage feedback. The algorithm has been implemented as one part of GoldMine 1 . Experiments using a variety of designs, including USB, PCI and OpenRisc, show that the assertions generated by coverage guided association mining cover an average of 6.14 times input space than those generated by decision tree based mining algorithms. We also show that the coverage guided association mining produces an average of 2.75 times fewer propositions per assertion than decision tree based mining. All these mean the assertions generated by coverage guided association mining are more succinct, and of higher value for hardware design verification.

Abstract-Document representation is a fundamental problem for text mining. Many efforts have been done to generate concise yet semantic representation, such as bag-of-words, phrase, sentence and topic-level descriptions. Nevertheless, most existing techniques counter difficulties in handling monolingual comparable corpus, which is a collection of monolingual documents conveying the same topic. In this paper, we propose the use of frame, a high-level semantic unit, and construct frame-based representations to semantically describe documents by bags of frames, using an information network approach. One major challenge in this representation is that semantically similar frames may be of different forms. For example, "radiation leaked" in one news article can appear as "the level of radiation increased" in another article. To tackle the problem, a text-based information network is constructed among frames and words, and a link-based similarity measure called SynRank is proposed to calculate similarity between frames. As a result, different variations of the semantically similar frames are merged into a single descriptive frame using clustering, and a document can then be represented as a bag of representative frames. It turns out that frame-based document representation not only is more interpretable, but also can facilitate other text analysis tasks such as event tracking effectively. We conduct both qualitative and quantitative experiments on three comparable news corpora, to study the effectiveness of frame-based document representation and the similarity measure SynRank, respectively, and demonstrate that the superior performance of frame-based document representation on different real-world applications.

Abstract.Pattern-based classification has demonstrated its power in recent studies, but because the cost of mining discriminative patterns as features in classification is very expensive, several efficient algorithms have been proposed to rectify this problem. These algorithms assume that feature values of the mined patterns are binary, i.e., a pattern either exists or not. In some problems, however, the number of times a pattern appears is more informative than whether a pattern appears or not. To resolve these deficiencies, we propose a mathematical programming method that directly mines discriminative patterns as numerical features for classification. We also propose a novel search space shrinking technique which addresses the inefficiencies in iterative pattern mining algorithms. Finally, we show that our method is an order of magnitude faster, significantly more memory efficient and more accurate than current approaches.

No abstract

In the past, there have been dozens of studies on automatic authorship classification, and many of these studies concluded that the writing style is one of the best indicators for original authorship. From among the hundreds of features which were developed, syntactic features were best able to reflect an author's writing style. However, due to the high computational complexity for extracting and computing syntactic features, only simple variations of basic syntactic features such as function words, POS (Part of Speech) tags, and rewrite rules were considered. In this paper, we propose a new feature set of k-embedded-edge subtree patterns that holds more syntactic information than previous feature sets. We also propose a novel approach to directly mining them from a given set of syntactic trees. We show that this approach reduces the computational burden of using complex syntactic structures as the feature set. Comprehensive experiments on real-world datasets demonstrate that our approach is reliable and more accurate than previous studies.

No abstract