Project Guidelines

For the course project you can either choose a topic of your liking and read 2-3 related papers, or decide to work on an open problem or do an empirical evaluation. In all cases your project must involve a significant theoretical component. There are two deliverables, an online blog post and a $20$ minute in-class presentation. Below is a list of possible project topics. The list will be continuously updated.

Project Schedule

  • Dec 13: 7pm: Jingry Yi: Distribution-Specific Hardness of Learning Neural Networks
  • Dec 13: 7:30pm: Gang Qiao: On the Quality of the Initial Basin in Overspecified Neural Networks
  • Dec 13: 8pm: Pengxiang Wu: Follow the Leader with Dropout Perturbations.
  • Dec 13: 8:30pm: Zhiqiang Tang: Algorithms for stochastic multi armed bandits
  • Dec 13: 9pm: Behnam Babagholami: Domain Adaptation and Transfer Learning
  • Dec 13: 9:30pm: Alireza Naghizadeh: Subspace Clustering
  • Dec 13: 10:00pm: Hui Qu: On the expressive power of deep neural networks
  • Dec 13: 10:30pm: Yikai Zhang: Coresets, Sparse Greedy Approximation, and the Frank-Wolfe Algorithm
  • Dec 13: 11:00pm: Qiaoying Huang: Online dictionary learning for sparse coding
  • Dec 13: 11:30pm: Orsan Aytekin: Hypothesis testing
  • Dec 14: 7pm: Ana Uribe: Bandit Convex Optimization
  • Dec 14: 7:30pm: Malihe Alikhani: Topic Models
  • Dec 14: 8pm: He Chen: Second order regret bounds parametrized by variance across actions and top epsilon percentile
  • Dec 14: 8:30pm: Mohamed Ahmed: Privacy preserving correlation
  • Dec 14: 9pm: Michael Jiao: Membership query learning
  • Dec 14: 9:30pm: Faisal Mohammad: Restricted Boltzmann machines for collaborative filtering
  • Dec 14: 10:00pm: Kaitlin Poskaitis: Spectral Clustering and kernel k-means
  • Dec 14: 10:30pm: Charles Shvartsman: Cortical Learning
  • Dec 14: 11:00pm: Yiran Sun: Community Detection
  • Dec 14: 11:30pm: Dong Yang: Structure Learning for Ising Models
  • Dec 14: 12am: Timothy Yong: Adaptive data analysis
  • -- >


    Possible Project Topics

  • Do a survey on the chaning method
    [See Chapter 5 of this book] [See Chapter 8 of this book]
  • Algorithms for stochastic multi armed bandits
    [Survey by Bubek and Cesa-Bianchi] [UCB algorithm]
  • Algorithms for online convex optimization in the bandit setting
    [paper by Bubek, Eldan and Lee]
  • Attribute efficient learning
    [learning sparse parities]
  • Deep Nets
    [hardness result] [hardness result] [structure of local minima]
  • Computational lower bounds
    [hardness of DNFs] [hardness of halfspaces]
  • Tensor methods in ML
    [Tensor methods for latent variable models] [Tensors methods for HMMs]
  • Adaptive data analysis
    [On holdout resuse]
  • Membership query learning
    [Learning decision trees] [Learning DNFs]
  • Learning graphical models
    [Structure learning of ising models]
  • Hypothesis testing
    [Testing closeness of distributions]