Tim Shi   施天麟

PhD Student,   Stanford University

Co-founder,   AI+ Club

Email   tianlins [at] cs [dot] stanford [dot] edu
Curriculum Vitae   [PDF]

  About Me

I love to build creative software.
I am optimisitic about AI and singularity.


I've done lots of research in Bayesian machine learning during my undergrad. The following are the papers I've written. Now, I am working on a more ambitious project in the intersecion of AI and human-computer interaction. Stay tuned! Part of it has exposed in OpenAI Universe.

Online Bayesian Passive-Aggressive Learning
Tianlin Shi and Jun Zhu.

Journal of Machine Learning Research 2017 (in press)
Documents: [Abstract] [PDF] [Slides at ICML] [ICML Version] [Supp] [Slides at TNList Lab] [Talk]
Improving Survey Aggregation with Sparsely Represented Signals
Tianlin Shi, Forest Agostinelli, Matthew Staib, David Wipf, Thomas Moscibroda

KDD 2016: The 22nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining.
Documents: [Abstract] [PDF] [Video]
Errata: In theorem 4.3, M should be f, the global mean.

Max-Margin Deep Generative Models
Chongxuan Li, Jun Zhu, Tianlin Shi and Bo Zhang.

NIPS 2015: The Twenty-ninth Annual Conference on Neural Information Processing Systems (NIPS).
Documents: [Abstract] [PDF]

Learning Where to Sample in Structured Prediction
Tianlin Shi, Jacob Steinhardt and Percy Liang

AISTATS 2015: Accepted as Oral (6.1%).
The 18th International Conference on Artificial Intelligence and Statistics, San Diego, California, USA.
Documents: [Abstract] [PDF] [Slides]

Correlated Compressive Sensing for Networked Data
Tianlin Shi, Da Tang, Liwen Xu and Thomas Moscibroda

UAI 2014: 30th Conference on Uncertainty in Artificial Intelligence, Quebec City, Canada, July 2014.
Documents: [Abstract] [PDF]
This paper originates from the course project of Network Science, taught by Prof. Thomas Moscibroda . See our project report..
A Reverse Hierarchy Model for Predicting Eye Fixations
Tianlin Shi, Liang Ming and Xiaolin Hu.

CVPR 2014: IEEE Conference on Computer Vision and Pattern Recognition, Columbus, Ohio, USA., 2014
Documents: [Abstract] [Arxiv]
A Fully Polynomial-Time Approximation Scheme for Approximating a Sum of Random Variables
Jian Li*, and Tianlin Shi*.

ORL: Operations Research Letters 42.3 (2014): 197-202.
Documents: [Abstract] [PDF]  [Elsevier]   [Arxiv]   [Software] .
*Author names follow alphabetic order.
This paper is based on a pset problem from the course Algorithm Design, taught by Prof. Jian Li
Gradient-based inference for higher-order probabilistic programming languages
Tianlin Shi, Alexey Radul and Vikash Mansinghka

NEML 2014: Workshop at New England Machine Learning Day
Documents: [Abstract] [Poster] .

  Some Side Projects I Have Worked on

Hummer: Learn Singing through Games
Uber Automator
RayX: A C++ Ray Tracer
MicroReader on Pebble