IBM POWER9 servers are designed specifically to handle both traditional and emerging artificial intelligence workloads, as well as the fast-evolving world of high-performance computing workloads. To mitigate the reduction in performance improvement from transistors in modern VLSI technology, the POWER9 servers embed a series of innovations in both hardware and system software architectures. The papers in this issue of the IBM Journal of Research and Development effectively describe these innovations.
This issue of the IBM Journal of Research and Development describes the innovative design and technology of the IBM z14, the latest IBM mainframe, with its significant new capabilities, including pervasive encryption, analytics and machine learning, platform simplification, enhanced system capacity and performance, significant reduction in I/O latency, data serving, reliability, robustness, and new semiconductor process technology. The topics covered in this issue also consider client value, solutions, and business motivation, along with additional new technologies and innovations across the stack.
User-generated health data includes data related to activity, diet, exercise, sleep, symptoms, treatments, and outcomes that are collected by the patient outside clinical settings. Beyond various data capture and processing challenges, it is essential to understand how insights generated from the data can be made useful for key stakeholders (e.g., payers, providers, and patients). This issue of the IBM Journal of Research and Development emphasizes new methods, models, capabilities, and technologies that focus on the collection, processing, privacy, curation, analysis, interpretation, and use of insights from user-generated health data.
Nontopical papers: At the end of this issue are two nontopical papers on 1) building a cognitive platform for the managed IT services lifecycle and 2) reengineering a server ecosystem for enhanced portability and performance.
Unprecedented growth of digital information provides special opportunities for having a positive impact on public policy and many other activities for societal benefit. Such activities may use analytics and include projects ranging from helping those in need and improving access to health care and education, to reducing pollution and creating collaboration approaches for effective utilization of various resources. This issue of the IBM Journal of Research and Development emphasizes new solutions, models, capabilities, and technologies that focus on addressing humanitarian challenges. Topics include health and well-being, improvements to cities and ecoscapes, the fostering of innovation and philanthropic projects, addressing the challenges of humanitarian crises and disasters, developing tools for collaboration, and more.
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 Journal issue 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.
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.