PhD Scholarships in Bayesian Statistical Probabilistic Programming

We invite motivated and well-qualified candidates to work on Bayesian Statistical Probabilistic Programming, as part of a research programme funded by the National Research Foundation, Singapore.

The appointees will be hosted by the School of Computer Science and Engineering, Nanyang Technological University (NTU), Singapore. They will work in the Probabilistic Programming Lab, where research is carried out on a wide range of topics, especially in the interface of programming languages, machine learning, and Bayesian statistics, but also in allied areas in semantics of computation, formal methods and verification, and in logic and algorithms.

Bayesian Statistical Probabilistic Programming lies in the interface of programming languages, machine learning, and Bayesian statistics. These positions will suit researchers with expertise in one (or more) of the three areas, and are interested and committed to collaborating with experts in the other areas.

Background

The ultimate goal of our research programme is to develop provably correct and scalable general-purpose inference algorithms for programs in universal probabilistic programming languages (PPL).

Where possible, we seek results and insights that are generally applicable to Bayesian inference of nonparametric models, in a manner that is programming language neutral.

Our general objectives are:

1. Design new inference algorithms for nonparametric models
2. Devise new methods to verify and diagnose Bayesian inference methods
3. Develop optimising techniques for inference procedures and statistical tests

The appointees will work in one or more of the following research directions:

• Nonparametric inference algorithms for universal probabilistic programming: ICML21, ICML22, NeurIPS22
• Static analysis, verification and diagnostics of Bayesian inference: PLDI21, PLDI22a, LICS22
• Estimator efficiency: variance reduction and cost analysis: UAI22
• Optimisation techniques for inference engines
• Foundations, reasoning techniques and enabling methods for Bayesian statistic probabilistic programming: LICS21a, PLDI22b, ESOP21, LICS21b, LICS19

(For references, see Luke Ong's dblp page.)

The appointees will work closely with the PI, Prof. Luke Ong, and other team members.

Selection Criteria

• Outstanding academic record
• Awards and prizes
• Evidence of research abilities, such as research reports or publications
• References

Desirable Skills

The following skills are not essential, but desirable:

• Experience in functional programming
• Experience in algorithmic studies and experiments

Practical

Four PhD scholarships are available. This call was posted in October 2022. It will remain open until the two positions are filled.

Our admin colleagues will provide support for visa application, relocation, etc.

Starting date: start of next semester.

Location: Singapore (remote work from abroad is not allowed).

Applications and Enquiries

To apply, email a copy of your application to the PI, Luke Ong (Email: luke.ong @ ntu . edu . sg), with the subject: PhD Scholarship Application (Bayesian Statistical Probabilistic Programming), including the following:

• your academic transcripts and CV
• brief description of research interests
• dblp or Google scholar link to your publications (if any)
• representative papers or reports (if available)
• contacts of three references.