My Professional Career

[March, 2024 - Now] Senior AI Researcher at Dolby

[April, 2021 - Now] Adjunct Member of Machine Learning Research Group, Electronics and Communication Sciences Unit, Indian Statistical Institute

[April, 2023 - February, 2024] Senior Software Engineer - Machine Learning at LinkedIn

[April, 2021 - March, 2023] Applied Reseacrh Engineer - Natural Language Processing at LinkedIn

[Januray, 2021 - March, 2021] Visiting Scientist at Electronics and Communication Sciences Unit, Indian Statistical Institute

My Education

[Ph.D. in Computer Science, 2014-2021] from Indian Statistical Institute under the supervision of Prof. Swagatam Das

[M.Tech. in Computer Science, 2012-2014] from Indian Statistical Institute

[B.Tech. in Computer Science and Engineering, 2008-2012] from West Bengal University of Technology

My Current Research Interests

• Few-shot Learning especially Few-task Problems
• Class Imbalanced Learning
• Adversarial Attacks and Robust Models
• Generative and Language Models
• Evolutionary Optimization specifically Differential Evolution

My Publications

Please visit my Google Scholar page for more information and preprints. You can also find implementations in MatLab or Python for many of these works in my GitHub profile.

2023

1. Shounak Datta*, Sankha Subhra Mullick*, Anish Chakrabarty, and Swagatam Das. "Interval bound interpolation for few-shot learning with few tasks." In International Conference on Machine Learning (ICML23), pp. 7141-7166. PMLR, 2023. [*equal contribution]
2. Sankha Subhra Mullick, Mohan Bhambhani, Suhit Sinha, Akshat Mathur, Somya Gupta, and Jidnya Shah. "Content moderation for evolving policies using binary question answering." In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics, Volume 5: Industry Track, (ACL23), pp. 561-573. 2023.

2022

1. Swagatam Das*, Sankha Subhra Mullick*, and Ivan Zelinka. "On supervised class-imbalanced learning: An updated perspective and some key challenges." IEEE Transactions on Artificial Intelligence, volume 3, no. 6 (2022): 973-993. [*equal contribution]
2. Arkaprabha Basu, Sourav Das, Sankha Subhra Mullick, and Swagatam Das. "Do Preprocessing and Class Imbalance Matter to the Deep Image Classifiers for COVID-19 Detection? An Explainable Analysis." IEEE Transactions on Artificial Intelligence, volume 4, number 2 (2022): 229-241.

2021

1. Arka Ghosh, Sankha Subhra Mullick, Shounak Datta, Swagatam Das, Asit K. Das, and R. Mallipeddi. "A Black-box Adversarial Attack Strategy with Adjustable Sparsity and Generalizability for Deep Image Classifiers." Pattern Recognition, volume 122 (2021): 108279

2020

1. Sankha Subhra Mullick, Shounak Datta, Sourish Gunesh Dhekane, and Swagatam Das. “Appropriateness of Performance Indices for Imbalanced Data Classification: An Analysis.” Pattern Recognition, volume 102 (2020): 107197

2019

1. Sankha Subhra Mullick, Shounak Datta and Swagatam Das. “Generative adversarial minority oversampling.” In IEEE International Conference on Computer Vision (ICCV19), pp: 1695-1704. 2019.
2. Imon Banerjee, Sankha Subhra Mullick, and Swagatam Das. “On convergence of the class membership estimator in fuzzy k-nearest neighbour classifier.” IEEE Transactions on Fuzzy Systems, volume 27 number 6, (2019): 1226-1236.

2018

1. Sankha Subhra Mullick, Shounak Datta and Swagatam Das. "Adaptive learning-based k-nearest neighbour classifiers with resilience to class imbalance," IEEE Transactions on Neural Networks and Learning Systems, volume 29 number 11, (2018): 5713-5725.

2017

1. Arka Ghosh, Swagatam Das, Sankha Subhra Mullick, R Mallipeddi, and AK Das. “A switched parameter differential evolution with optional blending crossover for scalable numerical optimization”. Applied Soft Computing, volume 57 (2017): 329-352.
2. Nimagna Biswas, Saurajit Chakraborty, Sankha Subhra Mullick, and Swagatam Das. “A parameter independent fuzzy weighted k-Nearest neighbour classifier.” Pattern Recognition Letters, volume 101 (2017): 80-87.
3. Shounak Datta, Sankha Subhra Mullick, and Swagatam Das. "Generalized mean based back-propagation of errors for ambiguity resolution." Pattern Recognition Letters, volume 94 (2017): 22-29.

2016

1. Swagatam Das, Sankha Subhra Mullick, and Ponnuthurai N. Suganthan. "Recent advances in differential evolution: an updated survey." Swarm and Evolutionary Computation, volume 27 (2016): 1-30.

2015

1. Swagatam Das, Arka Ghosh, and Sankha Subhra Mullick. "A switched parameter differential evolution for large scale global optimization: simpler may be better." Mendel 2015, pp. 103-125. 2015. Springer, Cham.

My Professional Activities

• My Patents
  1. Suhit Sinha, Sankha Subhra Mullick, Akshat Mathur, Somya Gputa, and Jidnya Samir Shah. "Zero-Shot Training for Multimodal Content Classifier". US Patent App. 18/423,799 dated 01/26/2024, Status: pending..
• Reveiw Duties: I act as an active reviewer for IEEE Transactions on Evolutionary Computations (since 2019), IEEE Transactions on Fuzzy Systems (since 2019), IEEE Transactions on Neural Networks and Learning Systems (since 2018), Swarm and Evolutionary Computation (since 2018), Neurocomputing (since 2018), Pattern Recognition (since 2017), Information Sciences (since 2017), etc.
• Teaching Duties: I served as a teaching Assistant, Pattern Recognition and Image Processing, M. Tech. in Computer Science, Indian Statistical Institute, Fall 2020. I am a regular speaker at Summer School on Computer Vision, Graphics and Image Processing, an Electronics and Communication Sciences Unit, Indian Statistical Institute initiative since 2016. I am also a regular speaker at Winter School on Deep Learning, another yearly initiative by the Electronics and Communication Sciences Unit, Indian Statistical Institute, since 2021.
• Leadership: Co-conceptualized and currently managing the Winter School on Deep Learning (WSDL) since 2022. WSDL is a fully online, two-month long, weekends-only, certificate course, that aims to regularly bring the best minds in the world, in the field of the theory and application of Computer Science, to an audience of widely varying expertise. Check out the previous sessions of WSDL [2022, 2023, and 2024] to know more about the school and its success story.
• Supervision: I have supervised more than 15 undergraduate and graduate research interns from notable academic institutes such as the Indian Statistical Institute, Jadavpur University, IIT Guwahati, etc. on projects leading to journal papers, technical reports, and dissertations since 2015.

Beyond Research

Sankha is an avid reader who never discriminates between genres and gobbles up everything he can find close to his hand (he mostly prefers high fantasy and fiction though). He loves video games, especially RPGs (that may not always require his trusty PC and PS). Sankha has a tendency to go off the grid at least once a year, usually for high-altitude trekking. He also likes to write a lot, either in his mother tongue, Bengali or if the audience demands then, English. His literary works are mainly non-fiction, boring, post-editorial articles in newspapers, but at times, he also may feel fancy to venture into his imaginative (and romantic) side. Strongly believes in feminism, equity, and diversity. Oh, did you notice that Sankha has a knack for speaking about himself in the third person? Say, like Captain Terrence “Terry” Vincent Jeffords?

This website is developed by Sankha Subhra Mullick, over the Hydeout project, on 9th of September, 2019.