We describe and empirically evaluate machine learning methods for the prediction of zinc binding sites from protein sequences. We start by observing that a data set consisting of single residues as examples is affected by autocorrelation and we propose an ad-hoc remedy in which sequentially close pairs of candidate residues are classified as being jointly involved in the coordination of a zinc ...

In this paper we explore the possibilities to perform the search for results of non-deterministic computations in parallel. We present three different approaches to this problem: Dividing the search space statically, using a Bag of Tasks approach and using semi-explicit parallelism. Then we discuss their advantages and limitations. Finally, we discuss benchmarks using a monadic SAT solver as an...

The development of our team for RoboCup ’99 is mainly oriented towards a transparent way of transferring explicit knowledge into the agent control and its combination with learning algorithms capable of fine-tuning the acquired skills. The explicit knowledge is formulated in terms of rules, the non-explicit knowledge is to be realized as a set of parameters adapted by hierarchical reinforcement...

This paper focuses on the problem of aspect-specific sentiment analysis. The goal here is to not only extract aspects of a product or service, but also to identify specific sentiments being expressed about them. Most existing algorithms address this problem by treating aspect extraction and sentiment analysis as separate phases or by enforcing explicit modeling assumptions on how these two phas...

Dataless text classification [Chang et al., 2008] is a classification paradigm which maps documents into a given label space without requiring any annotated training data. This paper explores a crosslingual variant of this paradigm, where documents in multiple languages are classified into an English label space. We use CLESA (cross-lingual explicit semantic analysis) to embed both foreign lang...

In this paper we propose a method for logo recognition using deep learning. Our recognition pipeline is composed of a logo region proposal followed by a Convolutional Neural Network (CNN) specifically trained for logo classification, even if they are not precisely localized. Experiments are carried out on the FlickrLogos-32 database, and we evaluate the effect on recognition performance of synt...

On-line learning robotic systems have many desirable properties. This work contributes a reinforcement learning framework for learning a time-constrained closed-loop control policy. The task is to verge the two cameras of a stereo vision system to foveate on the same world feature, within a limited number of perception-action cycles. On-line learning is beneficial in at least the following ways...

Dual-process models of categorization posit dissociable implicit and explicit category learning systems. Evidence in favour of these accounts is typically obtained by examining how categorization responses differ over time, with differing category structures, and by changing task demands. If these two categorization systems are activated concurrently (e.g., COVIS) then both implicit and explici...

BACKGROUND Motor learning is central to domains such as sports and rehabilitation; however, often terminologies are insufficiently uniform to allow effective sharing of experience or translation of knowledge. A study using a Delphi technique was conducted to ascertain level of agreement between experts from different motor learning domains (i.e., therapists, coaches, researchers) with respect t...

The development of our team for RoboCup '99 is mainly oriented towards a transparent way of transferring explicit knowledge into the agent control and its combination with learning algorithms capable of ne-tuning the acquired skills. The explicit knowledge is formulated in terms of rules, the non-explicit knowledge is to be realized as a set of parameters adapted by hierarchical reinforcement l...