Formal methods in Machine learning

Open Ph.D. positions

One Ph.D. position (Residency: IIT Goa)

  • Stipend: 31,000Rs + HRA for first two years and then 35,000Rs + HRA
  • The position also allows for visiting LaBRI, France for few months. The funding will cover travel, accommodation and daily allowance.

One Ph.D. position (Residency: LaBRI, France (for 3 years), IIT Goa (remaining))

  • Stipend in France: 1,500 euro / month
  • Stipend in India: 35,000Rs + HRA

Eligibility criteria

  • BTech or MSc in CSE, Mathematics or relevant areas + qualified a national level examination like GATE, JEST etc
  • or, MTech in CSE, Mathematics or relevant areas
  • Applicant should also have interest in discrete mathematics, algorithms, automata theory, machine learning, and programming.

Queries

Selection process

The selection process is through online test and interview. The candidates are expected to show strong proficiency in one of the following areas:
  • Basic mathematics: Discrete mathematics, probability, linear algebra, calculus
  • Computer science fundamentals: Data structures, algorithms, programming, automata theory, machine learning

Important Dates

  • Application Deadline: 10th June 2023
  • Shortlisting of Candidates: 12th June 2023
  • Online Test: 18th June 2023
  • Online Interview: 19th - 23rd June 2023

Summary

As intelligent algorithms continue to gain wider adoption, they are beginning to replace humans in certain contexts. Ensuring the reliability and security of these algorithms in high-risk areas such as self-driving cars or unmanned planes/trains is critical. However, conventional testing methods are often insufficient to address the increased level of risk involved. As a result, formal verification of these algorithms becomes necessary. Even though there are significant challenges to formally verify machine learning algorithms, some recent developments offer hope in this rapidly evolving field of research.

The research is conducted by Sreejith A V (IIT Goa), Meenakshi D'Souza (IIIT Bangalore), and Vincent Penelle (LABRI, University of Bordeaux, France).

The research is funded by CEFIPRA: Indo-French centre for promotion of advanced research.

Resources
  • Supratik's course on Formal methods in Machine learning