Departmental laboratories and centers for instruction and research include:
Adnan Y. Darwiche, Director
The group focuses on research in automated reasoning (logical and probabilistic) and machine learning, including their application to problems in science and engi-neering. On the theoretical side, the group focuses on tractable circuit representations and models that combine logic and probability, in addition to new models for machine learning that can integrate background knowledge. On the practical side, the group builds scalable reasoning and learning systems that can scale to real-world problems.
Judea Pearl, Director
The laboratory targets research areas concerned with evidential reasoning, the distributed interpretation of multisource data in networks of partial beliefs; learning, the structuring and parameterizing of links in belief networks to form a representation consistent with a stream of observations; constraint processing, including intelligent backtracking, learning while searching, temporal reasoning, etc.; graphoids, the characterization of informational dependencies and their graph representations; and default reasoning, use of qualitative probabilistic reasoning to draw plausible and defeasible conclusions from incomplete information.
Guy Van den Broeck, Director
The laboratory performs research on machine learning (statistical relational learning, tractable learning), knowledge representation and reasoning (graphical models, lifted probabilistic inference, knowledge compilation), applications of probabilistic reasoning and learning (probabilistic programming, probabilistic databases), and artificial intelligence in general.
Quanquan Gu, Director
The laboratory conducts research on machine learning, optimization, and high-dimensional statistical inference. Its focus is on development and analysis of nonconvex optimization algorithms for machine learning to understand large-scale, dynamic, complex, and heterogeneous data; and on building the theoretical foundations of deep learning and deep reinforcement learning.
Eran Halperin, Director
The laboratory aims to improve understanding and treatment of human disease by analysis of big data collected in relation to diseases. The main focus of the laboratory has been development of methods for analysis of genomic data—including genetics, epigenetics, RNA, and microbiome data; as well as medical records, images, and waveforms of UCLA Health medical center patients. The methods developed are typically standalone tools, often used by other researchers for analysis of specific diseases. The methodology involves a combination of machine learning, optimization algorithms, combinatorial optimization, and classical and Bayesian statistics.
Joseph J. DiStefano III, Director
This interdisciplinary research typically involves integration of theory with real laboratory data, using biomodeling, computational, and biosystems approaches. Problem domains are physiological systems, disease processes, pharmacology, and some post-genomic bioinformatics. Laboratory pedagogy involves development and exploitation of the synergistic and methodologic interface between structural and computational biomodeling with laboratory data, or computational systems biology, with a focus on integrated approaches for solving complex biosystem problems from sparse biodata e.g., in physiology, medicine, and pharmacology), as well as voluminous biodata (e.g., from genomic libraries and DNA array data).
Eleazar Eskin, Director
The laboratory is comprised of a computational genetics group affiliated with both the Computer Science and Human Genetics departments. Research interests are in computational genetics, bioinformatics, computer science, and statistics. The laboratory focuses on developing techniques for solving the challenging computational problems that arise in attempting to understand the genetic basis of human disease.
Sriram Sankararman, Director
This interdisciplinary research group is affiliated with UCLA departments of Computer Science, Human Genetics, and Biomathematics. The laboratory is broadly interested in questions at the intersection of computer science, statistics, and biomedicine. It develops statistical and computational methods to make sense of complex, high-dimensional datasets generated in the fields of genomics and medicine, to answer questions ranging from how humans have evolved, to what the biological underpinnings of diseases are, to how we can improve the diagnosis and treatment of disease. A major focus of this research is understanding and interpreting human genomes. The biological questions of interest center around understanding how evolution shapes human genes, and how they modulate complex traits that include common diseases. The laboratory develops and extends tools from a diverse set of disciplines including machine learning, algorithms, optimization, high-dimensional statistics, and information theory. It also applies these tools to high-dimensional genomic and medical datasets that are publicly available or being generated by laboratory collaborators.
Anthony J. Nowatzki, Director
The laboratory studies how to redesign computer architectures and accelerators to continue improving performance and energy efficiency, even while technology scaling reaches its physical limits. Broadly, its approach is to consider how to reform traditional hardware/software abstractions to convey rich information that can make building efficient micro-architectures possible. These changes necessitate codesign of ISAs, architecture, execution models, and compilers.
Yuval Tamir, Director
The laboratory conducts research on the design, implementation, and evaluation of computer systems that use state-of-the-art technology to achieve high performance and high reliability. Projects involve both software and hardware, and often focus on parallel and distributed systems in the context of general-purpose as well as embedded applications.
Milos D. Ercegovac, Director
The laboratory is used for fast digital arithmetic (theory, algorithms, and design) and numerically intensive computing on reconfigurable hardware. Research includes floating-point arithmetic, online arithmetic, application-specific architectures, and design tools.
Majid Sarrafzadeh, Director
The ER Lab goal is to use technology in health care to reduce the cost of providing high-quality care to the chronically ill, estimated (by Milken Institute Center for Health Care Economics) to be over $1 trillion per year. The laboratory strives to improve global and local public health surveillance, with a resultant reduction in epidemics, increased control over infectious disease, and improved drug safety. Other goals are diminished rate of medical errors; ongoing preventive health, with attendant reductions in morbidity, mortality, and cost of care; and consumer engagement in health and self-management.
Jason Cong, Director
The laboratory investigates cutting-edge research topics at the intersection of VLSI technologies, design automation, architecture, and compiler optimization at multiple scales, from micro-architecture building blocks to heterogeneous compute nodes and scalable data centers. Currently, the laboratory is focused on architecture and design automation for emerging technologies; and customizable domain-specific computing with applications to multiple domains such as imaging processing, bioinformatics, data mining, and machine learning.
Song-Chun Zhu, Director
The laboratory is affiliated with the Computer Science and Statistics departments. Research begins with computer vision and expands to other disciplines. The objective is to pursue a unified framework for representation, learning, inference, and reasoning; and to build intelligent computer systems for real-world applications. Its projects span four directions: vision (object, scene, events, etc.); cognition (intentions, roles causality, etc.); learning (information projection, stochastic grammars, etc.); and art (abstraction, expression, aesthetics, etc.).
Demetri Terzopoulos, Director
The laboratory engages in a broad spectrum of visual computing research unifying computer graphics (image synthesis), computer vision (image analysis), and related fields; with emphasis on geometric, physics-based, learning-driven, and artificial intelligence/life modeling and simulation. Major research interests include biomimetic simulation of humans and other animals, from biomechanics to sensorimotor control to intelligence; and image/video analysis combining (deep) learning and modeling paradigms, especially for application to medicine and health care.
The collective brings together researchers from multiple departments at UCLA, including Mathematics, Statistics, Computer Science, Brain Mapping, Computational Biology, Neuroimaging, Image Informatics, Psychology, and Radiology.
Stefano Soatto, Director
Researchers investigate how images—i.e., measurements of light—can be used to infer properties of the physical world such as shape, motion, location, and material properties of objects. This is key to developing engineering systems that can “see” and interact intelligently with the world around them. For example, images captured by a car-mounted video camera can be processed by computers to infer a model of the car’s surroundings, e.g., other vehicles, pedestrians, etc. This technology can also be used to analyze images captured in the environment to understand the effects of climate change by monitoring the behavior of animals and plants. Analysis of images of the human body can be used both for diagnostic purposes and for planning interventions.
The group is a collaboration of all UCLA faculty from the information and data management field. It is interested in multiple research areas including big data, archival information systems, knowledge discovery and data mining, Earth Science Partners’ private network, genomics graph database development, multimedia information stream system technology, Smart Space middleware architecture, and technologically based assessment of language and literacy, to name just a few.
Kai-Wei Chang, Director
The group focuses on developing reliable machine learning solutions for processing natural languages. Specifically, it targets design of models, algorithms, and learning mechanisms to improve the generalization ability of natural-language processing models such that they can generalize across unseen tasks, unseen inputs, and low-resource languages.
Carlo A. Zaniolo, Director
This research group investigates Web-based information systems and seeks to develop enabling technology for such systems by integrating the Web with database systems. Current research efforts include the DeAL system, a next-generation datalog system; SemScape, an NLP-based framework for mining unstructured or free text; EARL (Early Accurate Result Library) for Hadoop; Panta Rei, a study of support for schema evolution in the context of snapshot databases and transaction-time databases; Stream Mill, a complete data stream management system; and ArchIS, a powerful archival information system.
Lixia Zhang, Principal Investigator
The laboratory mission is to help the Internet grow. Its research efforts focus on design and development of network architecture and protocols, and the challenges in building secure networks and systems. Its past work has turned to Internet standards and successful startups. Since 2010, the laboratory has been working on design and development of named data networking (NDN), a new Internet architecture.
George Varghese, Director
The laboratory focuses on research in network design automation, an effort to build a comprehensive set of design tools for networks inspired by electronic design automation for chips. A major focus is analysis and synthesis of router configuration files to avoid major outages that frequently cripple major service providers. This work involves development of new tools inspired by other fields such as programming languages, hardware design, and data mining; but targeted to incorporate the special structure and challenges of networks. It involves collaboration with multiple disciplines such as programming languages, systems, and network debugging; and includes other UCLA faculty.
Ravi Netravali, Director
The group is focused on building practical systems to improve the performance and ease of debugging large-scale distributed applications. Such applications include web pages, mobile apps, video streaming and analytics systems, data analytics platforms, and more. The group uses a cross-layer methodology that aims to understand the impact of decisions at different layers in the end-to-end system; and designs solutions that incorporate fundamental principles at the network, operating system, and application vantage points.
Leonard Kleinrock, Director
The laboratory offers an environment to support advanced research in technologies at the forefront of all things regarding networking and connectivity, and will deliver the benefits of that research to society globally. The laboratory’s broad-based agenda enables faculty, students, and visitors to pursue research challenges of their own choosing, without externally imposed constraints on scope or risk. It draws inspiration from the foundational role of UCLA as the birthplace of the Internet. With its open inclusive structure, the laboratory will help to realize the vision of creating high-leverage technologies, as was accomplished years ago with the Internet.
Songwu Lu, Director
The laboratory’s research areas include wireless networking, mobile systems, and cloud computing. Its focus is on design, implementation, and experimentation of protocols, algorithms, and systems for wireless data networks. The goal is to build high-performance and dependable networking solutions for the wireless Internet.
The laboratory is used for research into compilers, embedded systems, and programming languages.
Harry Xu, Director
The group builds systems to improve the efficiency, scalability, reliability, and security of modern applications and workloads. These include graph analytics, video analytics, machine learning, smart contracts, etc. The group’s solutions cross multiple layers of the compute stack, spanning the areas of programming languages, compilers, operating systems, runtime systems, distributed systems, networking, and computer architecture.
Miryung Kim, Director
The laboratory conducts research in software engineering, in particular debugging and testing for big data systems and automated tools for data science and ML-based systems. Its overall goal is to improve software engineering productivity and correctness. To achieve it, the laboratory designs scalable software systems, software analysis algorithms, and automated development tools. It also conducts user studies with software engineers, and carries out statistical analysis of open-source project data to allow data-driven decisions for designing novel software engineering tools. With expertise in software evolution, the laboratory is known for its research on code clones—code duplication detection, management, and removal solutions. The laboratory is a leader in creation and definition of the emerging area where software engineering and data science intersect. It has conducted the most comprehensive study of industry data scientists, and developed automated debugging and testing technologies for widely-used big data systems such as Apache Spark. Through tech-transfer, several companies have used SEAL research on interactive code clone search and big data analytics debugging technologies.
The group is a collaboration of faculty from the software systems and network systems fields. It conducts research on the design, implementation, and evaluation of operating systems, networked systems, programming languages, and software engineering tools.
The center was established in 2001 with researchers from several laboratories in the Computer Science, and Electrical and Computer Engineering, departments. It serves as a forum for intelligent-agent researchers and visionaries from academia, industry, and government, with an interdisciplinary focus on fields such as engineering, medicine, biology, and social sciences. Information and technology are exchanged through symposia, seminars, short courses, and collaboration in joint research projects sponsored by government and industry.
Research projects include use of unmanned autonomous vehicles, coordination of vehicles into computing clouds, and integration of body sensors and smart phones into m-health systems. Ongoing research encompasses personal and body networks, cognitive radios, ad hoc multihop networking, vehicular networks, dynamic unmanned backbone, underwater unmanned vehicles, mobile sensor platforms, and network coding.
Jason Cong, Director
The center looks beyond parallelization and focuses on domain-specific customization as a disruptive technology to bring orders-of-magnitude power-performance efficiency improvement to application domains. CDSC develops a general methodology for creating novel, customizable computing platforms, and associated compilation tools and runtime management environment to support domain-specific computing. Its recent focus is on design and implementation of accelerator-rich architectures, from single chips to data centers. It also includes highly automated compilation tools and runtime management software for customizable heterogeneous platforms, including multi-core CPUs, many-core GPUs, and FPGAs; and a general, reusable methodology for customizable computing applicable across domains. By combining these capabilities, the goal is to deliver a supercomputer-in-a-box or -in-a-cluster, customizable to an application domain to enable disruptive innovations therein. This approach has been successful in medical image processing, precision medicine, and machine learning. Originally funded by a $10 million National Science Foundation (NSF) Expeditions in Computing award, in 2014 CDSC received $3 million from Intel Corporation with matching NSF InTrans program support. CDSC research is also supported by SRC JUMP and several industrial partners.
Amit Sahai, Director
The center was established in 2014 through an NSF Secure and Trustworthy Cyberspace (SaTC) Frontier Award. The center tackles the deep and far-reaching problem of general-purpose software obfuscation. The goal of obfuscation is to enable software that can keep secrets: software that makes use of secrets, but such that they remain hidden even if an adversary can examine the software code in its entirety and analyze its behavior as it runs. The center is headquartered at UCLA with partners at Columbia, Johns Hopkins, and Stanford universities, and University of Texas at Austin.
Rafail Ostrovsky, Director
CICS was established in 2003 to promote all aspects of research and education in cryptography and computer security. It explores novel techniques for securing national and private-sector information infrastructures across various network-based and wireless platforms as well as wide-area networks. The inherent challenge is to provide guarantees of privacy and survivability under malicious and coordinated attacks.
The center has raised federal, state, and private-sector funding, including collaboration with Israel through multiple U.S.–Israel Binational Science Foundation grants. It has also attracted multiple international visiting scholars. ClCS explores and develops state-of-the-art cryptographic algorithms, definitions, and proofs of security; novel cryptographic applications such as new electronic voting protocols and identification, data-rights management schemes, and privacy-preserving data mining; security mechanisms underlying a clean-slate design for a next-generation secure Internet; biometric-based models and tools, such as encryption and identification schemes based on fingerprint scans; and the interplay of cryptography and security with other fields such as bioinformatics, machine learning, complexity theory, etc.
The institute was established in 2013 with a focus on the continuing growth of data and demand for smart analytics to mine that data. Such analytics are creating major transformative opportunities in science and industry. To fully capitalize on these opportunities, computing technology must solve the three-pronged challenge created by the exploding size of big data, the growing complexity of big data, and the increased sophistication of analytics that can be used to extract patterns and trends from the data.
Benjamin M. Wu, D.D.S, Ph.D. (Bioengineering), Director; Bruce Dobkin, M.D. (Medicine/Neurology), William Kaiser, Ph.D. (Electrical and Computer Engineering), Gregory J. Pottie, Ph.D. (Electrical and Computer Engineering), Co-Directors
WHI is leading initiatives in health care solutions in the fields of disease diagnosis, neurological rehabilitation, optimization of clinical outcomes for many disease conditions, geriatric care, and many others. WHI also promotes this new field in the international community through the founding and organization of the leading Wireless Health conference series.
WHI technology always serves the clinician community through jointly developed innovations and clinical trial validation. Each WHI program is focused on large-scale product delivery in cooperation with manufacturing partners. WHI collaborators include the UCLA schools of Medicine, Nursing, and Engineering and Applied Sciences; Clinical Translational Science Institute for medical research; Ronald Reagan UCLA Medical Center; and faculty from many departments across UCLA. WHI education programs span high school, undergraduate, and graduate students, and provide training in end-to-end product development and delivery for WHI program managers.
WHI develops innovative, wearable biomedical monitoring systems that collect, integrate, process, analyze, communicate, and present information so that individuals become engaged and empowered in their own health care, improve their quality of life, and reduce burdens on caregivers. WHI products appear in diverse areas including motion sensing, wound care, orthopaedics, digestive health and process monitoring, advancing athletic performance, and many others. Clinical trials validating WHI technology are underway at 10 institutions. WHI products developed by the UCLA team are now in the marketplace in the U.S. and Europe. Physicians, nurses, therapists, other providers, and families can apply these technologies in hospital and community practices. Academic and industry groups can leverage the organization of WHI to rapidly develop products in complete-care programs, and validate in trials. WHI welcomes new team members, and continuously forms new collaborations with colleagues and organizations in medical science and health care delivery.
In summarizing the resources now available to conduct experimentally based research in the UCLA Computer Science Department, it is useful to identify the major components of the research environment: the departmental computing facility, other hardware and software systems, administrative structure, and technical support staff.
Computing facilities range from large campus-operated supercomputers to a major local network of servers and workstations that are administered by the department computing facilities (DCF) or school network (SEASnet).
The departmental research network includes Oracle servers and shared workstations, on the school ethernet TCP/IP local network. A wide variety of peripheral equipment is also part of the facility, and many more research-group workstations share the network; the total number of machines exceeds 1000, the majority running the Linux operating system. The network consists of switched 10/100/1000 ethernet to the desktop with a gigabit backbone connection. The department LAN is connected to the campus gigabit backbone. An 802.11n wireless network is also available to faculty, staff, and graduate students.
The central facilities and wide-area networking are operated by the campuswide Information Technology Services. Access to the departmental and SEASnet machines is controlled so as to maximize the usefulness of these computers for education and research, but no direct charges are involved.
The support staff consists of hardware and software specialists. The hardware laboratory supports network connections, configures routers, switches, and network monitoring tools. The software group administers the department UNIX servers, providing storage space and backup for department users.