Information-Theoretic Approaches in Speech Processing and Recognition

Dear Colleagues,

Information theoretic quantities, such as entropy, mutual information and transfer entropy, have been successfully utilized in many areas of science and engineering. Mutual Information has been frequently preferred as a key quantity to reveal statistical dependencies between random variables, especially in cases where widely utilized linear correlation analyses become insufficient. As an asymmetric quantity, Transfer Entropy has been applied to detect directional information flows, helping to better understand the cause and effect relationships between different variables. Despite many applications in signal processing and machine learning, information theoretic quantities have been seldomly used in the speech processing literature. Most applications involve different but more informative feature selection of speech signals by using Mutual Information and its variants. Similar quantities are also utilized to provide improvements in the speech recognition quality. Another research area includes multimodal applications, such as the analysis of coupled effects of visual lip movements and speech recognition. In another recent study, speech information is decomposed into four components, which are language content, timbre, pitch and rhythm, via a triple information bottleneck. Deep learning applications are also common where convolutional bottleneck features are analyzed for speech recognition from an information-theoretic point of view. In addition to these, Bayesian inference and prediction of speech signals is another featured topic of interest within the scope of this special issue.

In this Special Issue, we would like to collect papers focusing on the theory and applications of information-theoretic approaches in speech processing and recognition. Some application areas can be classified as hands-free computing, automatic emotion recognition, automatic translation, home automation, telematics, and robotics, but a broader list of topics in information theory and Bayesian statistics are encouraged. Of special interest are theoretical papers elucidating the state-of-the-art of multimodal signal processing approaches.

Dr. Deniz Gençağa
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Entropy is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.