The prominence of weakly labeled data gives rise to a growing demand for object detection methods that can cope with minimal supervision. We propose an approach that automatically identifies discriminative configurations of visual patterns that are characteristic of a given object class. We formulate the problem as a constrained submodular optimization problem and demonstrate the benefits of th...

The landscape of online learning has evolved in a synchronous fashion with the development of the every-growing repertoire of technologies, especially with the recent addition of Massive Online Open Courses (MOOCs). Since MOOC platforms allow thousands of students to participate at the same time, MOOC participants can have fairly varied motivation. Meanwhile, a low course completion rate has be...

Business processes have tremendously changed the way large companies conduct their business: The integration of information systems into the workflows of their employees ensures a high service level and thus high customer satisfaction. One core aspect of business process engineering are events that steer the workflows and trigger internal processes. Strict requirements on interval-scaled tempor...

We discuss the frequent pattern mining problem in a general setting. From an analysis of abstract representations, summarization and frequent pattern mining, we arrive at a generalization of the problem. Then, we show how the problem can be cast into the powerful language of algorithmic information theory. This allows us to formulate a simple algorithm to mine for all frequent patterns. 1

Pattern mining from the web log data leads to discovery of usage patterns of the user who navigate the web. Patterns which appear frequently in the web log data are item-sets and sequences. In this paper, a novel algorithm Intelligent Generalized Sequential pattern (IGSP) is designed which shows better results than the Generalized Sequential Pattern (GSP) algorithm. Experiment is conducted with...

We consider the problem of finding pairs of short patterns such that, in a given input sequence of length n, the distance between each pair’s patterns is at least α. The problem was introduced in [1] and is motivated by the optimization of multiplexed nested PCR. We study algorithms for the following two cases; the special case when the two patterns in the pair are required to have the same len...

Transformation-based learning (TBL) and frequent pattern discovery (FPD) are two popular research paradigms, one from the domain of empirical natural language processing , the second from the eld of data mining. This paper describes how Eric Brill's original TBL algorithm can be improved via incorporation of FPD techniques. The algorithm B-Warmr is presented that upgrades TBL to rst-order logic...

This paper presents how to mine formulation rules from a named organization corpus. The TEIRESIAS algorithm, which is widely used in bioinformatics domain, is adopted. The experimental results based on MET2 test bed show that the approach of regarding the morpheme of a keyword as a cluster is the best, the approach of regarding all the keywords as the same cluster is the next, and the approach ...

____________________________________ vi List of Figures ___________________________________ x List of Tables __________________________________ xiv

Many tasks of contemporary Molecular Biology rely increasingly on automated techniques for the discovery of interesting patterns and associations among them, both in individual sequences and across sequence families. A number of computational models and tools have been set up in recent years in response to these needs. This paper concentrates on approaches based on discrete combinatorial algori...