Many recent breakthroughs in areas ranging from speech and image recognition to game playing, autonomous vehicles, and medical diagnostics have been made possible by artificial intelligence techniques popularly known as deep learning. Deep learning is a branch of machine learning that usually makes use of multiple processing layers and hierarchical representations to drive the learning process. This issue of the IBM Journal of Research and Development emphasizes new applications, architectures, data sets, tools, and other technologies that advance the field in deep learning. In particular, this issue covers various practical applications (e.g., involving speech, language, and image recognition), deep learning accelerators, hyperparameter optimization and parameter servers, and deep learning capabilities provided as a service.
Nontopical papers: Three nontopical papers at the end of this issue concern machine learning and hardware design, natural-language classifiers for question-answering systems, and speech recognition approaches for U.S. Open Tennis Championships.
Analysis and prediction of brain dynamics, observed in vast new datasets and rendered in large-scale simulations, are beginning to facilitate applications in important societal areas such as assistive technology, education, and the treatment of disease. This journal issue emphasizes solutions, models, and technologies that advance the analysis of neuroscientific data and the application of predictive models to anticipate changes in, or better understand, brain function, cognition,and behavior.
This journal issue emphasizes new solutions, models, capabilities, and other technologies that represent important advancements toward cognitive and contextual analytics for IT services. Topics include end-to-end distributed solutions, problem extraction and search, predictive maintenance and upgrades, and much more.
This issue of the IBM Journal of Research and Development emphasizes solutions, models, analytics, and other technologies that form important building blocks for improving food and water safety, security, and defense. Topics span areas including genetics, microbiology, epidemiology, crop production, and traceability in the food supply chain.
Nontopical paper: At the end of this issue is a nontopical paper concerning predictive models in B2B (business-to-business) recommender systems.
IBM Journal digital library
The complete digital library (1957 - present) of the IBM Journal of Research and Development, including all IBM Systems Journal issues, is available on IEEE Xplore.
The IBM Journal of Research and Development welcomes submissions that relate to the themes described in the currently open Calls for Papers.
IBM Journal subscriptions
A subscription to the IBM Journal of Research and Development includes access to the current year of the journal, as well as the complete ~60 years for both the IBM Journal of Research and Development and the IBM Systems Journal. Subscriptions can be ordered through the IEEE Xplore platform.
Single article sales
Single papers and full issues can also be purchased through IEEE Xplore. Browse the entire digital library of the IBM Journal of Research and Development, as well as back issues of the IBM Systems Journal.