X
Xi Chen
Researcher at University of California, Berkeley
Publications - 53
Citations - 26834
Xi Chen is an academic researcher from University of California, Berkeley. The author has contributed to research in topics: Reinforcement learning & Autoregressive model. The author has an hindex of 39, co-authored 53 publications receiving 22393 citations.
Papers
More filters
Proceedings Article
Meta learning shared hierarchies
TL;DR: In this article, a set of primitives are shared within a distribution of tasks and are switched between by task-specific policies, leading to an optimization problem for quickly reaching high reward on unseen tasks.
InfoGAN: interpretable representation learning by information maximizing Generative Adversarial Nets
TL;DR: Experiments show that InfoGAN learns interpretable representations that are competitive with representations learned by existing fully supervised methods.
Proceedings Article
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
TL;DR: In this article, the authors investigated and improved upon three limiting design choices employed by flow-based models in prior work: the use of uniform noise for dequantization, use of inexpressive affine flows, and use of purely convolutional conditioning networks in coupling layers.
Proceedings Article
PixelSNAIL: An Improved Autoregressive Generative Model
TL;DR: In this paper, a new generative model architecture that combines causal convolutions with self-attention is proposed, which achieves state-of-the-art results on CIFAR-10 (2.85 bits per dim) and ImageNet (3.80 bits per degree).
Journal ArticleDOI
Deep unsupervised cardinality estimation
Zongheng Yang,Eric Liang,Amog Kamsetty,Chenggang Wu,Yan Duan,Xi Chen,Pieter Abbeel,Joseph M. Hellerstein,Sanjay Krishnan,Ion Stoica +9 more
TL;DR: This paper proposed a Monte Carlo integration scheme on top of autoregressive models that can efficiently handle range queries with dozens of dimensions or more, achieving up to 90x accuracy improvement over the second best method.