Logo: University of Southern California

Events Calendar



Select a calendar:



Filter March Events by Event Type:



Events for March 23, 2017

  • CS Colloquium: Long Lu (Stony Brook University) - New OS and Programming Support for Securing Mobile and IoT Platforms

    Thu, Mar 23, 2017 @ 11:00 AM - 12:20 PM

    Computer Science

    Conferences, Lectures, & Seminars


    Speaker: Long Lu, Stony Brook University

    Talk Title: New OS and Programming Support for Securing Mobile and IoT Platforms

    Series: CS Colloquium

    Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.

    Software running on mobile and IoT platforms increasingly falls victim to new attacks, which cause device compromises and privacy leaks that are often more severe than their counterparts on conventional computers. My research finds that new attacks on these platforms are possible primarily due to a gap between the evolving security needs of software and the legacy security support provided by operating systems and programming tools.

    In this talk, I will first overview my recent works that aim to bridge this gap by rethinking the principles and designs of security mechanisms in operating systems, compilation toolchains, and TEEs (Trusted Execution Environments). I will then present two systems that address a critical yet previously unmet security need of today's apps, namely in-app isolation. The first system introduces a new OS-managed code execution unit, called shred, to compensate thread and process. A shred is a segment of a thread execution. Code inside a shred can access, in addition to the regular virtual memory, a private memory region. Using shreds, programmers can now protect sensitive in-memory code and data against untrusted code running in the same process or thread. The second system enables comprehensive security policy enforcement at the sub-app granularities, preventing mutually distrusting app modules from abusing each other's resources and privileges. In the final part of the talk, I will discuss my ongoing and future works on laying the system foundation for securing IoT platforms.

    Biography: Long Lu is an Assistant Professor of Computer Science and the director of RiS3 Lab at Stony Brook University. Long's research spans the broad area of systems and software security. His recent work is focused on application and operating system security for emerging platforms, such as mobile and IoT/CPS devices. He designs code and data protection mechanisms, program analysis techniques, and user-facing software tools to prevent real attacks. He constantly publishes in the top-tier computer security conferences and is frequently invited to serve on their program committees. His research outcomes have been adopted by IBM, Microsoft, NEC, and Samsung. His work is currently funded by NSF, ONR, ARO, and AFRL. Long is a recipient of the NSF CAREER Award and the Air Force Faculty Fellowship. He holds a Ph.D. in Computer Science from Georgia Tech.

    Host: CS Department

    Location: Ronald Tutor Hall of Engineering (RTH) - 217

    Audiences: Everyone Is Invited

    Posted By: Assistant to CS chair

    OutlookiCal
  • CS Colloquium: Matthew Brown (UCLA) -Typed Self-Applicable Meta-Programming

    Thu, Mar 23, 2017 @ 04:00 PM - 05:20 PM

    Computer Science

    Conferences, Lectures, & Seminars


    Speaker: Matthew Brown, UCLA

    Talk Title: Typed Self-Applicable Meta-Programming

    Series: CS Colloquium

    Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.

    Meta-programming is a fundamental technique in computer science. It allows high levels of abstraction to be utilized with low cost. Meta-programs like compilers, interpreters, and program optimizers make high-level programming languages efficient, providing increased programmer productivity and performance comparable to lower-level languages. Self-applicable meta-programming makes meta-programming first-class, enabling many powerful
    techniques. However, meta-programming and particularly self-applicable meta-programming is often complex, error-prone and difficult to debug. For these reasons it has untapped potential to provide benefits in many areas. Typed meta-programming uses modern techniques for type checking meta-programs to make them less error-prone and easier to understand and debug. It also brings the power of self-applicable meta-programming to statically-typed languages, ending a long-persisting trade-off between static and dynamic type checking. In this talk I discuss foundational results in typed self-applicable meta-programming.

    Biography: Matt Brown is PhD candidate at UCLA, working in the compilers lab under Jens Palsberg. He holds a Bachelor's degree from UC Santa Cruz and a Master's from UCLA. His research focus is typed self-applicable meta-programming, which uses typed program representation techniques to ensure correctness properties of self-applicable meta-programs like self-interpreters. Other research interests include type systems, program verification, concurrency, and functional programming languages. He was recently a part-time lecturer at Loyola Marymount University and has six years of industry experience.

    Host: CS Department

    Location: Henry Salvatori Computer Science Center (SAL) - 101

    Audiences: Everyone Is Invited

    Posted By: Assistant to CS chair

    OutlookiCal