HDSI General

Events from this organizer

Today

  • Security and Privacy in an Everchanging System Landscape

    Special Seminar Series
    SDSC, The Auditorium 9836 Hopkins Dr, La Jolla, San Diego, CA, United States

    Abstract: From AI and IoT to AR/VR and Web 3.0, computer systems are evolving at an unprecedented rate. While this evolution has given rise to exciting applications and opportunities, it has also brought about novel security and privacy challenges within these systems and across their interactions with existing platforms. In this talk, I will discuss how system security researchers can keep up with this everchanging landscape and showcase some of my lab's recent work on understanding and detecting malicious web bots. I will explore how we can build and roll out research infrastructure to measure web bot activities and later use our newfound understanding to develop practical solutions to counter them. I will highlight how we can apply similar research principles to areas such as AI and IoT. Finally, I will conclude my talk by previewing some of my ongoing work and outlining my research roadmap toward achieving "security at inception" for emerging systems.

  • Scaling and Generalizing Approximate Bayesian Inference | David Blei

    Jeffrey Elman Distinguished Lecturer Series
    SDSC, The Auditorium 9836 Hopkins Dr, La Jolla, San Diego, CA, United States

    A core problem in statistics and machine learning is to approximate difficult-to-compute probability distributions. This problem is especially important in Bayesian statistics, which frames all inference about unknown quantities as a calculation about a conditional distribution. In this talk I review and discuss innovations in variational inference (VI), a method that approximates probability distributions through optimization. VI has been used in myriad applications in machine learning and Bayesian statistics.

  • Algorithms for multi-group learning

    Colloquia Lecture Series
    SDSC, The Synthesis Center 9500 Gilman Drive, La Jolla, CA, United States

    Abstract: Multi-group agnostic learning is a formal learning criterion that is concerned with the conditional risks of predictors within subgroups of a population. The criterion addresses recent practical concerns such as subgroup fairness and hidden stratification. I'll talk about the structure of solutions to the multi-group learning problem, as well as some simple and near-optimal algorithms for the learning problem. This is based on joint work with Christopher Tosh.

  • Spectral clustering in high-dimensional Gaussian mixture block models

    Colloquia Lecture Series
    SDSC, The Synthesis Center 9500 Gilman Drive, La Jolla, CA, United States

    The Gaussian mixture block model is a simple generative model for networks: to generate a sample, we associate each node with a latent feature vector sampled from a mixture of Gaussians, and we add an edge between nodes if and only if their feature vectors are sufficiently similar. The different components of the Gaussian mixture […]

  • Representation Learning: A Causal Perspective

    Colloquia Lecture Series
    SDSC, The Auditorium 9836 Hopkins Dr, La Jolla, San Diego, CA, United States

    Abstract: Representation learning constructs low-dimensional representations to summarize essential features of high-dimensional data like images and texts. Ideally, such a representation should efficiently capture non-spurious features of the data. It shall also be disentangled so that we can interpret what feature each of its dimensions captures. However, these desiderata are often intuitively defined and challenging to quantify or enforce.

  • On the complexity of Frank-Wolfe methods

    Colloquia Lecture Series
    SDSC, The Auditorium 9836 Hopkins Dr, La Jolla, San Diego, CA, United States

    Abstract: Frank-Wolfe methods are popular for optimization over a polytope. One of the reasons is because they do not need projection onto the polytope but only linear optimization over it. This talk has two parts. The first part will be about the complexity of Wolfe's method, an algorithm closely related to Frank-Wolfe methods. In 1974 […]

  • 5-Year Anniversary Celebration & Ribbon Cutting

    Featured

    On June 5th, 2023 the Halıcıoğlu Data Science Institute celebrated the 5-year anniversary since becoming its own academic department. In parallel to these celebrations, HDSI unveiled it's first dedication space, the Halıcıoğlu Data Science Institute building.

  • Deep Latent Variable Models for Compression and Natural Science | Stephan Mandt

    Computer Science & Engineering Building (CSE), Room 1202

    Latent variable models have been an integral part of probabilistic machine learning, ranging from simple mixture models to variational autoencoders to powerful diffusion probabilistic models at the center of recent media attention. Perhaps less well-appreciated is the intimate connection between latent variable models and data compression, and the potential of these models for advancing natural science. This talk will explore these topics.

  • Causal Inference symposium

    Causality is increasingly a part of AI, data science, robotics, and more, but it is not always clear how we can learn causality from data. This symposium will be featuring leading HDSI Faculty who will be providing an introductory overview on these methods, followed by domain-specific talks and open discussion.

  • Some new results for streaming principal component analysis

    Special Seminar Series

    Abstract: While streaming PCA (also known as Oja’s algorithm) was proposed about four decades ago and has roots going back to 1949, theoretical resolution in terms of obtaining optimal convergence rates has been obtained only in the last decade. However, we are not aware of any available distributional guarantees, which can help provide confidence intervals on the quality of the solution. In this talk, I will present the problem of quantifying uncertainty for the estimation error of the leading eigenvector using Oja's algorithm for streaming PCA, where the data are generated IID from some unknown distribution. Combining classical tools from the U-statistics literature with recent results on high-dimensional central limit theorems for quadratic forms of random vectors and concentration of matrix products, we establish a distributional approximation result for the error between the population eigenvector and the output of Oja's algorithm. We also propose an online multiplier bootstrap algorithm and establish conditions under which the bootstrap distribution is close to the corresponding sampling distribution with high probability. While there are optimal rates for the streaming PCA problem, they typically apply to the IID setting, whereas in many applications like distributed optimization, the data is generated from a Markov chain and the goal is to infer parameters of the limiting stationary distribution. If time permits, I will also present our near-optimal finite sample guarantees which remove the logarithmic dependence on the sample size in previous work, where Markovian data is downsampled to get a nearly independent data stream.