Research

Highlighted Projects

  • Regularized Generative Adversarial Nets (GANs)

    • Theory and Principles for Regularization, Generalization and Semi-Supervised Learning of GANs

      • LS-GAN: Loss-Sensitive GANs (more...)
      • GLSAL: Generalized Loss-Sensitive Adversarial Learning with Manifold Margins (more...)
      • LGAN: Localized GANs for Semi-Supervised Learning with Laplace-Beltrami Operators (more...)

    • Applications

      • HTG: An Adversarial Approach to Hard Triplet Generation (more...)

  • Learning Transformation Equivariant Representations (TERs)

    • Theory, Principles and Models for TERs

      • AET: AutoEncoding Transformations rather than Data (more...)
      • AVT: Autoencoding Variational Transformations (more...)

  • Small Data Challenges in Big Data Era

    • Read Our survey of

      • Recent progress on unsupervised and semi-supervised learning (more...)

    • Small Data Methods

      • Unsupervised Learning & Semi-Supervised Learning
      • Few-Shot Learning & Zero-Shot Learning

  • Time-Series Data Analysis

    • SFM: State-Frequency Memory Recurrent Neural Networks
    • Stock Price Prediction via Discovering Multi-Frequency Trading Patterns
    • Mixture Factorized Ornstein-Uhlenbeck Processes
    • First-Take All: Hashing Time-Series Data

  • Perception, Control, Decision-Making and Optimization in Smart City

    • Machine Perception for Smart City (e.g., surveillance cameras and IoT)
    • AI for Traffic system diagnostics and traffic signal optimization
    • AI for scheduling and optimization of civil utilities