Johns Hopkins scientists develop and apply cutting edge technology in a wide variety of fields of inquiry. In keeping with JHU’s collaborative and interdisciplinary culture, ML@JHU gathers ML practitioners from a variety of schools, departments, institutions and centers in order to exchange ideas, coordinate curriculum, work closely with experimental scientists and domain experts, and promote machine learning on campus.
Schools
The Johns Hopkins University has been ranked the #1 U.S. academic institution for over 31 consecutive years in total R&D spending on science, medicine, and engineering. Its machine learning community is strong, multi-disciplinary, and collaborative, with theory and applications feeding off each other.
- The Whiting School of Engineering is home to the departments of Computer Science, Electrical and Computer Engineering, and Applied Mathematics & Statistics, among others.
- The Bloomberg School of Public Health was the first school of public health, and is currently the largest school of public health with around 600 full time faculty. Consistently ranked the number one school of public health, it receives around 20% of all federal research funding in public health. Many members of the Department of Biostatistics are ML practitioners. JHSPH has many interesting datasets, including international medicine and epidemiology.
- The School of Medicine is consistently rated one of the top three medical research schools in the country. Johns Hopkins sees nearly one million patients a year. Patient medical records comprising text, images, etc. present a wide range of interesting ML problems. The Department of Biomedical Engineering, a strong participant in ML@JHU, is shared between the Homewood Campus (home of Engineering, Arts, & Sciences) and the medical campus (home of the hospital and the schools of Medicine, Public Health, & Nursing).
- The Applied Physics Laboratory is a non-profit center for engineering, research, and development. With over 4,500 employees, over 67% of whom are engineers and scientists, and nearly $1 billion in annual funding, it is one of the world’s largest research centers. Data from APL is varied and complex, often relating to national security.
Centers
JHU is known for cross-departmental collaboration. It has built critical mass in several interdisciplinary research areas. Each “research center” draws its faculty and students from several traditional departments. The following major centers include significant work on developing and applying new ML techniques within their general research area.
- The Center for Language and Speech Processing (CLSP) and the associated Human Language Technology Center of Excellence (HLTCOE) are among the world leaders in speech and language research, specializing in information-theoretic approaches to machine translation, speech recognition, and natural language processing.
- The Institute for Data Intensive Engineering and Science (IDIES) facilitates and supports large-scale data processing, in fields as diverse as astrophysics, geological and environmental sciences, and neuroscience. ML and “big learning” play a prominent role in extracting value from these massive terabyte- and petabyte-scale data. IDIES is a CUDA Center of Excellence and is building massive computational resources to support such efforts, most recently the 5-petabyte Data-Scope cluster and one of the nation’s first 100-gigabit Internet connections.
- The Laboratory for Computational Sensing and Robotics (LCSR) and the Center for Computer-Integrated Surgical Systems and Technology (CISST) focus on developing systems that integrate novel computer and human/machine interface technologies, both to revolutionize surgical procedures and for more general vision and robotics applications. ML for sensing and control is fundamental to these goals.
- The Institute for Computational Medicine (ICM) and the Center for Imaging Science (CIS) work on computational models to transform medicine: bioinformatics, biological systems monitoring, and much more. Nonlinear registration, subspace factor analysis, and graphical models are a small subset of the kinds of ML tools used in these domains.
- The new Center for Personalized Cancer Medicine and Center for Population Health Information Technology (CPHIT) both aim to systematically improve patient care through learning algorithms that run on massive datasets of genetic data or electronic medical records.
- The Space Telescope Science Institute is deeply involved in all aspects of the Hubble Space Telescope. Similarly, the Sloan Digital Sky Survey @ JHU is home to the largest astronomy data set in the world, from a variety of space telescopes. Thus, massive machine learning problems abound, especially on hyper-spectral image data. With the advent of “big data” and bringing computation to the data, instead of the other way around, hosting these data brings unique big-data problems to JHU.
- The Mathematical Institute for Data Science (MINDS) consists of a multidisciplinary team of mathematicians, statisticians, computer scientists, and engineers whose goal is to develop the fundamental mathematical, statistical, and computational principles for the analysis and interpretation of massive amounts of complex high-dimensional data.
Natural Learning
JHU also has a significant presence in the study of learning and information processing in humans and other organisms. Theories of biological learning often serve as inspiration for machine learning algorithms, and vice-versa.
- The new Science of Learning Institute seeks to study human learning at all levels, from underlying genetics and biology up through effective educational practices. Its charter includes a focus on using ML to improve human learning, not merely to model it.
- The Krieger Mind-Brain Institute [past/upcoming talks] asks: How does neural activity in the brain give rise to mental phenomena? How does the brain process information about the world to generate perception, knowledge, decision, and action? The institute combines state-of-the-art experimental techniques for measuring neural activity with linear and nonlinear modeling on large compute clusters.
- The Cognitive Science Department [talks], ranked #1 by the NRC, studies the mind as a computational system, with a particular focus on language and spatial representation. The department’s research seeks to discover the formal structure of cognition at all levels of analysis.
- The department of Psychological and Brain Sciences [talks] includes research on learning, perception, attention, cognition, and cognitive neuroscience.
- The Vision Sciences Group is a community of researchers from several departments who work on biological vision as well as machine vision.
- The Neuroscience Department [talks] has a distinguished history and a large group of faculty and affiliated faculty who work on cognitive and systems neuroscience.
- The School of Education, whose graduate education program was ranked #1 by U.S. News and World Report in 2014, is becoming increasingly interested in machine learning through their Center for Technology in Education and the university-wide Science of Learning initiative.