Scalable Heterogeneous Transfer Ranking
نویسندگان
چکیده
In this paper, we propose to study the problem of heterogeneous transfer ranking, a transfer learning problem with heterogeneous features in order to utilize the rich large-scale labeled data in popular languages to help the ranking task in less popular languages. We develop a large-margin algorithm, namely LM-HTR, to solve the problem by mapping the input features in both the source domain and target domain into a shared latent space and simultaneously minimizing the feature reconstruction loss and prediction loss. We analyze the theoretical bound of the prediction loss and develop fast algorithms via stochastic gradient descent so that our model can be scalable to large-scale applications. Experiment results on two application datasets demonstrate the advantages of our algorithms over other state-of-the-art methods.
منابع مشابه
Intelligent scalable image watermarking robust against progressive DWT-based compression using genetic algorithms
Image watermarking refers to the process of embedding an authentication message, called watermark, into the host image to uniquely identify the ownership. In this paper a novel, intelligent, scalable, robust wavelet-based watermarking approach is proposed. The proposed approach employs a genetic algorithm to find nearly optimal positions to insert watermark. The embedding positions coded as chr...
متن کاملProbabilistic information retrieval in a distributed heterogeneous environment
This thesis describes a probabilistic model for optimum information retrieval in a distributed heterogeneous environment. The model assumes the collection of documents o ered by the environment to be hierarchically partitioned into subcollections. Documents as well as subcollections have to be indexed. At this, indexing methods using di erent indexing vocabularies can be employed. A query provi...
متن کاملAvailability Analysis For Heterogeneous Nucleation In A Uniform Electric Field
Industrial demands for more compact heat exchangers are a motivation to find new technology features. Electrohydrodynamics (EHD) is introduced as a promising phenomenon for heat transfer enhancement mechanisms. Similar to any new technology, EHD has not been understood completely yet and requires more fundamental studies. In boiling phase change phenomena, nucleation is the dominant mechanism i...
متن کاملSoMap: Dynamic Clustering and Ranking of Geotagged Posts
This demo presents SoMap, a web-based platform that provides new scalable methods to aggregate, analyse and valorise large collections of heterogeneous social data in urban contexts. The platform relies on geotagged data extracted from social networks and microblogging applications such as Instagram, Flickr and Twitter and on Points Of Interest gathered from OpenStreetMap. It could be very insi...
متن کاملFlexible sample selection strategies for transfer learning in ranking
Ranking is a central component in information retrieval systems; as such, many machine learning methods for building rankers have been developed in recent years. An open problem is transfer learning, i.e. how labeled training data from one domain/market can be used to build rankers for another. We propose a flexible transfer learning strategy based on sample selection. source domain training sa...
متن کامل