Zongheng Yang

I'm a PhD candidate in the RISELab at UC Berkeley, where I'm advised by Ion Stoica. My research focuses on learning and optimizing the capabilities of data systems using deep learning.

In my previous life, I worked at Google Brain (TensorFlow & deep learning research; 2016-2017), and interned at Microsoft Research (2019), Databricks (2014), and Twitter (2013). I received my Bachelors in computer science and math from UC Berkeley. During my undergrad, I worked on various big data systems in the AMPLab.

Publications

• NeuroCard: One Cardinality Estimator for All Tables.
  Zongheng Yang, Amog Kamsetty*, Sifei Luan*, Eric Liang, Yan Duan, Xi Chen, and Ion Stoica. (*equal contribution)
  VLDB 2021 / PVLDB 14. Arxiv Talk Slides Star
• Variable Skipping for Autoregressive Range Density Estimation.
  Eric Liang*, Zongheng Yang*, Ion Stoica, Pieter Abbeel, Yan Duan, Xi Chen. (*equal contribution)
  ICML 2020. Website Code
• Qd-tree: Learning Data Layouts for Big Data Analytics.
  Zongheng Yang, Badrish Chandramouli, Chi Wang, Johannes Gehrke, Yinan Li, Umar Farooq Minhas, Per-Ake Larson, Donald Kossmann, Rajeev Acharya.
  SIGMOD 2020. Website Talk Slides

• 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.
  VLDB 2020 / PLVDB 13. Arxiv Talk Slides Star

• Learning to Optimize Join Queries With Deep Reinforcement Learning.
  Sanjay Krishnan, Zongheng Yang, Ken Goldberg, Joseph Hellerstein, Ion Stoica.
  arXiv, 2018. Blog
• Ray: A Distributed Framework for Emerging AI Applications.
  Philipp Moritz, Robert Nishihara, Stephanie Wang, Alexey Tumanov, Richard Liaw, Eric Liang, Melih Elibol, Zongheng Yang, William Paul, Michael I. Jordan, Ion Stoica.
  OSDI 2018.

Earlier Work

• Natural TTS Synthesis by Conditioning WaveNet on Mel Spectrogram Predictions.
  Jonathan Shen, Ruoming Pang, Ron J. Weiss, Mike Schuster, Navdeep Jaitly, Zongheng Yang, Zhifeng Chen, Yu Zhang, Yuxuan Wang, RJ Skerry-Ryan, Rif A. Saurous, Yannis Agiomyrgiannakis, Yonghui Wu.
  ICASSP 2018.
• Tacotron: Towards End-to-End Speech Synthesis.
  Yuxuan Wang, RJ Skerry-Ryan, Daisy Stanton, Yonghui Wu, Ron J. Weiss, Navdeep Jaitly, Zongheng Yang, Ying Xiao, Zhifeng Chen, Samy Bengio, Quoc Le, Yannis Agiomyrgiannakis, Rob Clark, Rif A. Saurous.
  INTERSPEECH 2017.
• ZipG: A Memory-efficient Graph Store for Interactive Queries.
  Anurag Khandelwal, Zongheng Yang, Evan Ye, Rachit Agarwal, Ion Stoica.
  SIGMOD 2017.
• Ernest: Efficient Performance Prediction for Large-Scale Advanced Analytics.
  Shivaram Venkataraman, Zongheng Yang, Michael Franklin, Benjamin Recht, Ion Stoica.
  NSDI 2016.
• SparkR: Scaling R Programs with Spark.
  Shivaram Venkataraman, Zongheng Yang, Davies Liu, Eric Liang, Hossein Falaki, Xiangrui Meng, Reynold Xin, Ali Ghodsi, Michael Franklin, Ion Stoica, Matei Zaharia.
  SIGMOD 2016 (industrial track).

Earlier Projects

• Ray: contributed a fault-tolerant and highly available control-state storage (chain-replicated redis).
• TensorFlow: my public work at Google Brain included working on TensorFlow. One of my contributions is a new checkpoint format with much better efficiency.
• Succinct/ZipG: a data store that queries directly on compressed files. I worked on ZipG, a distributed graph store on top that focuses on memory-efficiency.
• SparkR: an R front-end for Apache Spark. Talk: SparkR: Enabling Interactive Data Science at Scale. Spark Summit, June 2014, San Francisco.
• Apache Spark: I was a Spark contributor, mostly on Spark SQL.