One of the nice properties of kernel classifiers such as SVMs is that they often produce sparse solutions. However, the decision functions of these classifiers cannot always be used to estimate the conditional probability of the class label. We investigate the relationship between these two properties and show that these are intimately related: sparseness does not occur when the conditional pro...

Data sparseness is one of the factors that degrade statistical machine translation (SMT). Existing work has shown that using morphosyntactic information is an effective solution to data sparseness. However, fewer efforts have been made for Chinese-to-English SMT with using English morpho-syntactic analysis. We found that while English is a language with less inflection, using English lemmas in ...

With the development of personalized services, collaborative filtering techniques have been successfully applied to the network recommendation system. But sparse data seriously affect the performance of collaborative filtering algorithms. To alleviate the impact of data sparseness, using user interest information, an improved user-based clustering Collaborative Filtering (CF) algorithm is propo...

Nonnegative matrix factorization (NMF) has been widely studied in recent years due to its effectiveness representing nonnegative data with parts-based representations. For NMF, a sparser solution implies better representation. However, current NMF methods do not always generate sparse solutions. In this paper, we propose new method log-norm imposed on the factor matrices enhance sparseness. Mor...

<p style='text-indent:20px;'>Compressive speech enhancement makes use of the sparseness and non-sparseness noise in time-frequency representation to perform enhancement. However, reconstructing sparsest output may not necessarily translate a good enhanced signal as distortion be at risk. This paper proposes two level optimization approach incorporate objective quality measures compressive...

Non-negative matrix factorization (NMF) is a recently developed technique for finding parts-based, linear representations of non-negative data. Although it has successfully been applied in several applications, it does not always result in parts-based representations. In this paper, we show how explicitly incorporating the notion of ‘sparseness’ improves the found decompositions. Additionally, ...

Can we model speech recognition in noise by exploring higher order statistics of the combined signal? How will changes in these statistics affect speech perception in noise? This study addresses these questions in two experiments. One investigated the relationship between an established ”glimpsing” model and the fourth order statistic, kurtosis. The glimpsing model [1] proposes that listeners c...

Predictive multiple suppression methods consist of two main steps: a prediction step, in which multiples are predicted from the seismic data, and a subtraction step, in which the predicted multiples are matched with the true multiples in the data. The last step appears crucial in practice: an incorrect adaptive subtraction method will cause multiples to be sub-optimally subtracted or primaries ...

In hands-free telephony, the acoustic coupling between the loudspeaker and the microphone generates echoes that can seriously degrade user experience. For this reason, effective Acoustic echo cancellation (AEC) is important to maintaining and improving the perceived voice quality of a call. Traditionally, adaptive filters have been deployed in acoustic echo cancellers to estimate the Acoustic i...