Hierarchical probabilistic models, such as Gaussian mixture models, are widely used for unsupervised learning tasks. These models consist of observable and latent variables, which represent the observable data and the underlying data-generation process, respectively. Unsupervised learning tasks, such as cluster analysis, are regarded as estimations of latent variables based on the observable on...

In this paper we present a semi-automatic approach for acqusition of lexico-syntactic knowledge for event extraction in two Slavic languages, namely Bulgarian and Czech. The method uses several weaklysupervised and unsupervised algorithms, based on distributional semantics. Moreover, an intervention from a language expert is envisaged on different steps in the learning procedure, which increase...

This paper gives a brief description of a hierarchical architecture (HONG) that has been described elsewhere. The learning algorithm it uses is a mixed unsupervised/supervised method with most of the learning being unsupervised. The architecture generates multiple classifications for every data pattern presented, and combines them to obtain the final classification. The main purpose of this pap...

(Supervised and Unsupervised STratified Adaptive IncrementalNetwork) is a network model of human category learning. SUSTAIN initially assumes a simple category structure. If simple solutions prove inadequate and SUSTAIN is confronted with a surprising event (e.g. it is told that a bat is a mammal instead of a bird), SUSTAIN recruits an additional cluster to represent the surprising event. Newly...

While there is a need for business process management professionals in economy, enrolment in IS programs is declining. The resulting gap is even widened by early terminations of students due to various reasons. Missing motivation, as one of these reasons, is tackled by the Active Lab course concept presented in this paper. Such Active Labs are inspired by Social BPM Labs, which represent course...

We handle design problems with modeling and optimization. Sometimes this procedure is under question due to the lack of knowledge about the design constraints, especially when human evaluation is involved. In this work we look into the problem of indentifying constraints when they are not physically available and can only be identified by trial. The solution is indeed an application of active l...

We develop a fully discriminative learning approach for supervised Latent Dirichlet Allocation (LDA) model, which maximizes the posterior probability of the prediction variable given the input document. Different from traditional variational learning or Gibbs sampling approaches, the proposed learning method applies (i) the mirror descent algorithm for exact maximum a posterior inference and (i...

We handle design problems with modeling and optimization. Sometimes this procedure is under question due to the lack of knowledge about the design constraints, especially when human evaluation is involved. In this work we look into the problem of identifying constraints when they are not physically available and can only be identified by trial. The solution is indeed an application of active le...

This paper describes our approach to tackle the Author Verification task at PAN 2015. Our method builds a representation of an author’s style by using the information contained in dependency trees. This information is represented as syntactic n-grams and used to conform a vector space. Using unsupervised machine learning approach, each instance is associated to the correponding author using the...