Dear Colleagues,

The idea of using non-negative matrix factorization (NMF) for topic modeling was first introduced by Lee and Seung (1999) and popularized by Blei et al. (2003) with the publication of the latent Dirichlet allocation (LDA) algorithm. Both NMF and LDA are able to discover latent topics within a document collection, but used different mathematical approaches. NMF algorithms decompose a matrix into two or more reduced-dimension non-negative matrices. One of the advantages of NMF is that it can be used to identify meaningful subcomponents of high-dimensional data from a variety of modalities. For example, it can be used to decompose audio signals or images into their component parts, or to identify clusters of similar documents or related molecules. NMF is a powerful tool for data analysis and pattern recognition, and has found many applications across many domains, including signal processing, computer vision, image compression, gene expression data analysis, clustering, text mining, topic modelling, social network analysis, audio source separation, bioinformatics and chemoinformatics. Many different algorithms have been developed for NMF, including Lanczos, active sets methods and variational approaches.

NMF research has focussed on the design of novel NMF algorithms with improved initialization techniques, convergence properties, and the ability to take additional domain information and constraints as input. In recent years, there has been a growing interest in using NMF for data augmentation in the field of deep learning. Algorithms is a leading academic publication and seeks high-quality papers on the latest advances in NMF algorithms, a rapidly growing area of research.

Researchers are invited to submit original papers on all aspects of non-negative matrix factorization, including but not limited to:

• Novel NMF algorithms with improved convergence properties;
• Novel variants of LDA, NMF and neural-inspired algorithms for topic modeling;
• Initialization techniques utilizing but not limited to genetic algorithms;
• Numerical analysis and comparison of existing NMF algorithms (e.g., Lanczos, active sets methods and variational approaches);
• Deep NMF architectures for topic modeling;
• NMF algorithms for constraint clustering and the inclusion of domain information (e.g., vectors obtained from large language models such as BERT or transformer models);
• Neural network architectures augmented with NMF layers;
• The use of NMF to assist with the interpretation of large language models (using transformer architectures) and deep neural networks;
• Applications of NMF within varied domain areas, including but not limited to natural language processing, bioinformatics and computer vision;
• Semi-NMF and simultaneous NMF algorithms.
• Improving the scalability of neural-inspired algorithms for topic modeling.
• Algorithms to support cross-lingual topic modeling;
• Topic modeling algorithms that allow for the inclusion of domain information (e.g., vectors obtained from large language models such as BERT or transformer models);
• The use of topic modeling algorithms to assist with the interpretation of large language models (using transformer architectures) and deep neural networks;
• Applications of topic modeling algorithms within varied domain areas including but not limited to natural language processing, bioinformatics and computer vision;
• Smart user interfaces for user interaction and steering of topic modeling algorithm output.

Dr. Aneesha Bakharia
Dr. Khanh Luong
Guest Editors

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 100 words) can be sent to the Editorial Office for announcement on this website.

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. Algorithms 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 1600 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.