Colloquia and Seminars

To join the email distribution list of the cs colloquia, please visit the list subscription page.


Computer Science events calendar in HTTP ICS format for of Google calendars, and for Outlook.

Academic Calendar at Technion site.

Upcoming Colloquia & Seminars

  • Performance Prediction of Programs on Heterogeneous and Massively Parallel Architectures

    Speaker:
    Uri Shomroni, M.Sc. Thesis Seminar
    Date:
    Thursday, 29.8.2019, 16:30
    Place:
    Room 601 Taub Bld.
    Advisor:
    Prof. A. Mendelson

    Massively parallel, throughput-oriented processors are becoming increasingly common. Maximizing the benefit of these processors requires algorithms to be implemented differently than the sequential algorithms that most software developers are familiar with. This change is often very time-consuming and is not guaranteed to give an increase in performance matching the amount of effort. This work proposes an approach to predict the performance gain from porting an algorithm from the CPU to the GPU, based only on the measurements performed on the CPU implementation. CPU performance counters are measured on the sequential implementation and fed to a simple machine learning model trained on a given set of benchmarks. The experiments performed in this work have shown that the predictions' accuracy is similar or better than existing methods, that usually require converting the algorithm as a starting point. (Talk will be given in Hebrew)

  • Thapl - A Theaterical Programming Language

    Speaker:
    Matan Peled, M.Sc. Thesis Seminar
    Date:
    Tuesday, 3.9.2019, 12:30
    Place:
    Room 601 Taub Bld.
    Advisor:
    Prof. Y. Gil and Prof. D. H. Lorenz

    The purpose of this research is to explore an innovative approach to the declarative and imperative paradigms of programming languages. To demonstrate this approach, we have developed a prototype domain-specific language inspired by the scripts of theatrical plays, hence dubbed Tʜᴀᴘʟ. Tʜᴀᴘʟ's intended use is in the context of animation generation in “slide-show” presentations. There are two different ways to integrate animations in a presentation. The first is to embed an animation within a single slide, so that the animation starts when that slide starts and stops when continuing to the next slide, as if it were a video. The other is to present the animation as a series of slides, where each slide is static, like a flip book or a kinetoscope; Tʜᴀᴘʟ embraces this latter approach. The theatrical play metaphor encompasses the concept of a classical theatrical play script (e.g., Shakespearean play), enumerating the “Dramatis Personæ,” scenery, text, and actions (e.g., “exit chased by a bear”). Tʜᴀᴘʟ expands the declarative and imperative paradigm by introducing a declarative vocabulary for describing actors and actions on those actors, as well as a behavior language with specialised constructs for describing sequential and concurrent actions. Although programming languages that create presentations and animations already exist, they either operate at unreasonably low levels of abstraction which are ill-suited for creating animations or are aimed at authoring other media types (e.g., films), and only support slide-show creation as an after-thought. Conversly, existing software tools for creating slide-shows are not designed for animations. Modifying an existing animation is hard, and for solutions that utilize binary formats, fundamental software-engineering tools such as source control and are impractical. Tʜᴀᴘʟ is useful whenever one would like to graphically elucidate concepts with on-screen visual objects that dynamically change. Examples include academic courses (especially those that deal with graphs and algorithms), product presentations, and fiscal reviews.

  • The 9th Annual International TCE Conference on Autonomous Systems

    The 9th Annual International TCE Conference on Autonomous Systems

    Date:
    Wednesday, 11.9.2019, 09:00
    Place:
    CS Taub Build. Auditorium 1

    On Wednesday, September 11, TCE center will host this year the 9th annual Henry Taub International Conference on Autonomous Systems and will focus on the advent of autonomous systems including self-driving cars, automatic delivery drones, and service robots considered by many a major revolution of modern times that holds a profound impact on the economy and society.

    The event will be a unique opportunity to hear and meet international experts from the academia and the industry working on various aspects of autonomous systems ranging from sensing and machine learning to cyber security. as follows:

    Conference Chairs:
    Alex Bronstein, Computer Science Department, Technion
    Guy Gilboa, Electrical Engineerin, Technion

    More details, program and registration.
    Early bird registration is open until August 15.

  • Online Linear Models for Edge Computing

    Speaker:
    Hadar Sivan, M.Sc. Thesis Seminar
    Date:
    Wednesday, 11.9.2019, 11:30
    Place:
    Room 601 Taub Bld.
    Advisor:
    Prof. A. Schuster

    Maintaining an accurate trained model on an infinite data stream is challenging due to concept drifts that render a learned model inaccurate. Updating the model periodically can be expensive, and so traditional approaches for computationally limited devices involve a variation of online or incremental learning, which tend to be less robust. The advent of heterogeneous architectures and Internet-connected devices gives rise to a new opportunity. A weak processor can call upon a stronger processor or a cloud server to perform a complete batch training pass once a concept drift is detected -- trading power or network bandwidth for increased accuracy. We capitalize on this opportunity in two steps. We first develop a computationally efficient bound for changes in any linear model with convex, differentiable loss. We then propose a sliding window-based algorithm that uses a small number of batch model computations to maintain an accurate model of the data stream. It uses the bound to continuously evaluate the difference between the parameters of the existing model and a hypothetical optimal model, triggering computation only as needed. Empirical evaluation on real and synthetic datasets shows that our proposed algorithm adapts well to concept drifts and provides a better tradeoff between the number of model computations and model accuracy than classic concept drift detectors. When predicting changes in electricity prices, for example, we achieve 6% better accuracy than the popular EDDM, using only 20 model computations.

  • DeepRED: Deep Image Prior Powered by RED

    Speaker:
    Gary Mataev, M.Sc. Thesis Seminar
    Date:
    Monday, 16.9.2019, 11:30
    Place:
    Taub 401 Taub Bld.
    Advisor:
    Prof. M. Elad

    Inverse problems in imaging are extensively studied, with a variety of strategies, tools, and theory that have been accumulated over the years. Recently, this field has been immensely influenced by the emergence of deep-learning techniques. One such contribution, which is the focus of this paper, is the Deep Image Prior (DIP) work by Ulyanov, Vedaldi, and Lempitsky (2018). DIP offers a new approach towards the regularization of inverse problems, obtained by forcing the recovered image to be synthesized from a given deep architecture. While DIP has been shown to be quite an effective unsupervised approach, its results still fall short when compared to state-of-the-art alternatives. In this work, we aim to boost DIP by adding an explicit prior, which enriches the overall regularization effect in order to lead to better-recovered images. More specifically, we propose to bring-in the concept of Regularization by Denoising (RED), which leverages existing denoisers for regularizing inverse problems. Our work shows how the two (DIP and RED) can be merged into a highly effective unsupervised recovery process while avoiding the need to differentiate the chosen denoiser, and leading to very effective results, demonstrated for several tested problems.