mrinal

Profile picture for user mrinal
Dr. Mrinal Kanti Das
Assistant Professor
Email Me
Biosketch
  • * Post Doctoral Fellow
    • Department of Computer Science, University of Massachusetts Amherst, USA
    • Advisor: Prof. Andrew McCallum
  • * Post Doctoral Fellow
    • Department of Computer Science, Aalto University, Finland
    • Advisor: Prof. Samuel Kaski
  • * PhD. in Machine Learning
    • Department of Computer Science, Indian Institute of Science Bangalore
    • Advisor: Prof. Chiranjib Bhattacharyya
Research

I work broadly in principles and algorithms related to several machine learning areas, and more precisely on privacy aware learning, learning from big data, and Bayesian models. I enjoy the harmony of mathematics, programming, and real world problems that I encounter in my research. 

  • Privacy Aware Learning
    • With the increase of machine learning applications around every corner of the society, preserving privacy of individuals and institutions have become a concern. Privacy preservation requires perturbation of data which contradicts with learning principles and requirements. Privacy aware learning has been gaining popularity rapidly due to the potential of learning from data despite preserving privacy. However, there is hardly any practically applicable algorithm going beyond theories. I along with my collaborators in Aalto university, Finland have discovered a simple yet powerful technique to achieve privacy aware learning with reasonable accuracy. I have conceived the novel concept of projecting outliers to tighter bounds without affecting non-outliers which has been the key in our method. Our publication in Biology Direct, 2018 shows that it is possible to predict sensitivity of drugs even without revealing private information.
  • Learning from Big Data
    • I have also worked on Bayesian nonparametric models for learning very large scale (more than 8 million documents and 700 million tokens) datasets. There is NO method known using MCMC for such scale without using expensive parallel hardware. The work has been accepted at ICML, 2015.
  • Learning subtle information in text documents with MTV
    • Topic models are popular mathematical tools for analysing text datasets, where a corpus is a collection of documents. The state of art notion in topic models was to use single topic vector per document. I conceived the novel yet simple idea of using multiple topic vectors (MTV). We have observed phenomenal ability of MTV in (i) discovering subtle topics, (ii) modeling specific correspondence, (iii) modeling multi-glyphic topical correspondence, (iv) content driven user profiling for comment-worthy recommendations, (v) discovering taste of users in e-commerce portals. All of them helped in inventing novel models (i) subtle topic models (STM, in ICML, 2013), (ii) specific correspondence topic models (SCTM, in WSDM, 2014), (iii) multi-glyphic correspondence topic model (AAAI, 2015), (iv) collaborative correspondence topic models (CCTM in RecSys, 2015), (v) SOPER (CIKM, 2017).
Teaching
  • Theory and Practice of Data Science

    • This course is to teach under-graduate students to gain theoretical and practical maturity and ability to be able to solve problems encountered in data science today whether be related to industries or academic research.
      Spring semester at Indian Institute of Technology Palakkad, 2018.

  • Database Management Systems Practical

    • To provide hands-on experience to students on database management systems through complementing theory classes.
      Spring semester at Indian Institute of Technology Palakkad, 2018.

  • Paradigms of programming

    • This course covers various programming paradigms like imperative, object oriented, functional, logic etc.
      Fall semester at Indian Institute of Technology Palakkad, 2017.

  • Principles of machine learning

    • I had been a guest lecturer covering generative models, EM algorithm, Gibbs sampling, Gaussian mixture models, topic models.
      Fall semester at Indian Institute of Technology Palakkad, 2017.

  • Privacy aware learning

    • This course covers topic in basic machine learning, differential privacy fundamentals and techniques. Later part it covers recent research papers and also mini-projects.
      Fall semester at Aalto university, 2015.
      Course page at Aalto University

Research Group
Research Area
Machine Learning
Data Science
Privacyaware Learning
Bayesian models
Additional Information
Title
Organizing Events
Description

Organized ScienceIE – SemEval 2017 task 10: “Extracting Keyphrases and Relations from Scientific Publications” with Isabelle Augenstein, Sebastian Riedel, and Andrew McCallum. See ScienceIE website for details.

Our work got into WIRED.

Title
Reviewing and Program Committee
Description
  • 2019: SemEval (steering committee)

  • 2018: IJCAI, ICML, SemEval (steering committee)

  • 2017: IEEE Transaction on Pattern Analysis and Machine Intelligence (PAMI)

  • 2016: Neural Information P{rocessing Systems (NIPS), Neural Processing Letters

  • 2014: IEEE Transactions on Evolutionary Computation, IEEE CONECCT

Recent Publications

Antti Honkela(1), Mrinal Das(1), Arttu Nieminen, Onur Dikmen, Samuel Kaski.
Biology Direct (2018)
Isabelle Augenstein, Mrinal Das, Sebastian Riedel, Lakshmi Vikraman, Andrew McCallum.
International Workshop on Semantic Evaluation, 2017. (2017)
Lucky Dhakad, Mrinal Das, Chiranjib Bhattacharyya, Samik Datta, Mihir Kale, Vivek Mehta.
International Conference on Information and Knowledge Management (CIKM) (2017)
Antti Honkela(1), Mrinal Das(1), Onur Dikmen, Samuel Kaski.
Theory and Practice of Differential Privacy WS at ICML 2016. (2016)
Goutham Tholpadi, Mrinal Das, Trapit Bansal, Chiranjib Bhattacharyya.
Association for the Advancement of Artificial Intelligence Conference (AAAI) (2015)