ICSI hosts basic, pre-competitive research of fundamental importance to computer science and engineering. Projects are chosen based on the interests of the Institute’s principal investigators and the strengths of its researchers and affiliated UC Berkeley faculty.

Recent projects are listed below; the full list of each group's projects is accessible via the links listed in the sidebar.

Exploring Internet Balkanization through the Lens of Regional Discrimination

One of the Internet’s greatest strengths is the degree to which it facilitates access to any of its resources from users anywhere in the world. Various forces, however, have arisen that restrict particular users from accessing particular destinations, resulting in a "balkanization" of the network. This project explores apt methodologies for understanding such balkanization, an effort we will undertake in the context of examining "regional discrimination," i.e., the degree to which certain web services deny or degrade access to users from particular geographic regions.

Networking and Security
Previous Work: Automatic Recognition of Camera Speech (ARCS)

In this ICSI project, researchers are working to improve speech recognition from noisy, often distorted audio taken from the body cameras of working police officers during traffic stops. This is part of a larger project at Stanford to extract information from these data. The Stanford project is focused on the analysis of the interactions between the officers and the communities they serve, in the hope that they could help to transform the relationship between the police and communities, produce solid data on officer-community interaction, and inform officer training programs.

Interoperability Challenges and Scenarios in Computational Design and Manufacturing

A common informal definition of interoperability refers to the ability of a system, whose interfaces are completely understood, to work with other products or systems, present or future, without any restricted access or implementation. In the context of computational design and manufacturing, interoperability subsumes the problems of data sharing, exchange, and translation, as well as the problems of systems integration.

Research Initiatives

The FrameNet project is building a semantically-rich lexicon of English and a corresponding set of annotated texts, based on more than 600 semantic frames and 130,000 sentences. Comparable FrameNet projects are underway for Spanish, German, and other languages. By providing a layered semantic representation of text, FrameNet delivers a key component of next-generation question answering, machine translation, and other natural language processing applications. Learn more on the FrameNet Web site.

Effective and Economical Protection for High-Performance Research and Education Networks

As scientific research requires free exchange of information and ideas among collaborators world-wide, scientists depend critically on full and open access to the Internet. Yet in today’s world, such open access also exposes sites to incessant network attacks. Some of the most powerful networks today remain particularly hard to defend: for the 100G environments and backbones that facilitate modern data-intensive sciences, classic inline firewalls remain infeasible options.

Networking and Security
Lumen Privacy Monitor

Your mobile phone hosts a rich array of information about you and your behavior. This includes a wide range of unique identifiers and sensitive personal information that enables online tracking, often times for delivering targeted advertisement. It is, however, striking how little insight and control we, as mobile users have into the operation and performance of our devices, into how (or whether) they protect information we entrust to them, and who they share it with.

Networking and Security
Cybermanufacturing: Abstractions and Architectures for Open Composable Services

A key ingredient of any scenario in cloud-based design and manufacturing  and, more generally, in cybermanufacturing is the notion of virtualization of software and hardware resources in terms of open, interoperable and composable services. Such virtualization has proved challenging because modern manufacturing information systems are conceived and implemented as a federated integration of proprietary software tools that communicate via customized translators or by exchanging variety of neutral file formats.

Research Initiatives
Tactile Internet Islands

Today’s wireless communications solutions provide infrastructure for delivering content. Content delivery occurs either within a constraint island of coverage, within the reach, for example, of a wireless local area network (WLAN), or ubiquitously from anywhere to anywhere using, for example, cellular infrastructure. Content delivered can be, for example, voice, videos, images, and web information. Unfortunately however, wireless communications infrastructure available to us today does not allow for the remote control of real or virtual objects, as end-to-end latency far exceeds 10ms.

Research Initiatives
Previous Work: Using Individual Differences to Personalize Security Mitigations

Researchers at ICSI are leveraging well-studied individual differences in the psychology literature in order to improve computer security outcomes. Specifically, they are looking at how people with different decision-making styles may be more or less receptive to different types of security messaging. Applying techniques from behavioral economics, the goal is to frame security mitigations for individual users so that they see the security messages that are most likely to have an effect on them.

Usable Security and Privacy
Previous Work: Mobile Contextual Privacy

This project is rethinking how smartphones grant third-party applications access to sensitive user data. Currently, mobile platforms ask the user for permission the first time an application attempts to access certain data types; when access is granted, the user is never asked to make this decision again, even if the context in which subsequent data requests occur are substantially different from the context of the first request. For example, no distinction is made between using location data for location-based features and user tracking.

Usable Security and Privacy
Previous Work: Streaming Algorithms for Fundamental Computations in Numerical Linear Algebra

Streaming algorithms that use every input datum once (single-pass) or scan the input a small number of times (multiple passes) are gaining importance due to the increasing volumes of data that are available for business, scientific, and security applications. Performing large-scale data analysis and machine learning often requires addressing numerical linear algebra primitives, such as L2 regression, singular value decompositions, L1 regression, and canonical correlations.

Big Data
Shining Light on Non-Public Data Flows

This project looks into the usage and collection of data by programs that operate behind the scenes. The collected data and its use by a network of sellers, brokers, and marketers represents a direct privacy threat as it can be used for marketing, profiling, crime, or government surveillance, and yet consumers have little knowledge about it and no legal means to access the data. ICSI researchers are conducting surveys and experiments to determine the current status of this data and observe its effects.

Audio and Multimedia, Networking and Security

The NTL (Neural Theory of Language) project of the AI Group works in collaboration with other units on the UC Berkeley campus and elsewhere. It combines basic research in several disciplines with applications to natural language understanding systems. Basic efforts include studies in the computational, linguistic, neurobiological, and cognitive bases for language and thought. This research continues to yield a variety of theoretical and practical findings.

Understanding the State of TLS Using Large-scale Passive Measurements

This project leverages and extends the data collection of the ICSI SSL Notary for an extensive study of the real-world TLS/X.509 ecosystem through measurement-centric research. The SSL/TLS protocol suite constitutes the key building block of today’s Internet security, providing encryption and authentication for end-to-end communication with the help of an associated global X.509 public key infrastructure. However, from its first version in 1994 until today, researchers and practioners keep discovering TLS deficiencies undermining the protocol’s security on a regular basis.

Networking and Security
Towards a Science of Censorship Resistance

This project focuses on establishing a science of censorship resistance. Recent years have seen significant efforts on the part of both practitioners and researchers in countering large-scale Internet censorship imposed by nation-states. Driven by an active arms race, much of the research work in the field has been reactive in nature, lacking solid and methodical foundations.

Networking and Security
Previous Work: Security and Privacy for Wearable and Continuous Sensing Platforms

In this collaborative project, researchers at ICSI, UC Berkeley, and University of Washington are systematically exploring the security and privacy issues brought up by the increasing popularity of wearable computers. The recent demand for devices like Google Glass, smart watches, and wearable fitness monitors suggests that wearable computers may become as ubiquitous as cellphones.

Networking and Security, Usable Security and Privacy
Internet-Wide Vulnerability Measurement, Assessment, and Notification

Vulnerable software costs the U.S. economy more than $180 billion a year, and large-scale, remotely exploitable vulnerabilities affecting millions of Internet hosts have become a regular occurrence. This project seeks to reduce the impact of software vulnerabilities in Internet-connected systems by developing measurement-driven techniques for global vulnerability detection, assessment, and mitigation.

Networking and Security
Undermining Political “Scratch” Effects with Technology (UPSET)

This project leverages technology to significantly reduce the cost of political campaigns via the sharing economy, and thus undercut the impact of money on politics ( While the technology is important, the resulting structural change is more so; designing an efficient “supply chain” of services between a skilled citizen force and the campaigns (see for the beginnings of a separate spinoff organization to implement these ideas in real campaigns).

Research Initiatives
Previous Work: Word Bug

ICSI Speech researchers are working with Versame to develop methods for the analysis of speech being directed at infants and toddlers, in order to provide better measures of the lexical stimulation they are getting. The initial project is focused on the counting of speech units from unrestricted audio, where the likely speech units are syllables or words.

Previous Work: Deep and Wide Learning for Automatic Speech Recognition

In this project, speech researchers are looking at trade-offs between two approaches to automatic speech recognition (ASR): signal processing of multiple acoustic features vs. using simpler features and relying on machine learning algorithms to replace feature engineering. The goal is not only to improve accuracy for difficult examples, but also to learn about the computational consequences for high performance computing.