Current research : 1) Domain: Distributed Machine Learning and Computation under Constraints of Privacy, Security, Communication and Computational efficiency 2) Foundations: Randomized algorithms, Non-Asymptotic Statistics, Learning Augmented Algorithms.

Recent preprints/papers

(2020) New! DAMS: Meta-estimation of private sketch data structures for differentially private COVID-19 contact tracing, P.Vepakomma, S.N.Pushpita, R.Raskar (PDF)

(2020) New! Splintering with distributions: A stochastic decoy scheme for private computation, P.Vepakomma, J.Balla, R.Raskar (PDF)

(2020) New! FedML: A research library and benchmark for federated machine learning C He, S Li, J So, M Zhang, H Wang, X Wang, P Vepakomma, A Singh, arXiv preprint arXiv:2007.13518 (PDF)

(2020) New! NoPeek: Information leakage reduction to share activations in distributed deep learning P Vepakomma, A Singh, O Gupta, R Raskar arXiv preprint arXiv:2008.09161 (PDF)

(2020) New! PPContactTracing: A Privacy-Preserving Contact Tracing Protocol for COVID-19 Pandemic P Singh, A Singh, G Cojocaru, P Vepakomma, R Raskar arXiv preprint arXiv:2008.06648 (PDF)

(2020) New! Privacy in Deep Learning: A Survey F Mirshghallah, M Taram, P Vepakomma, A Singh, R Raskar (PDF)

(2020) New! Assessing Disease Exposure Risk with Location Histories and Protecting Privacy: A Cryptographic Approach in Response To A Global Pandemic A Berke, M Bakker, P Vepakomma, R Raskar, K Larson, AS Pentland (PDF)

(2020) New! SplitNN-driven Vertical Partitioning I Ceballos, V Sharma, E Mugica, A Singh, A Roman, P Vepakomma, arXiv preprint arXiv:2008.04137 (PDF)

(2020) COVID-19 Contact-Tracing Mobile Apps: Evaluation and Assessment for Decision Makers R Raskar, G Nadeau, J Werner, R Barbar, A Mehra, G Harp, P.vepakomma, et.al arXiv preprint arXiv:2006.05812 

(2020) Book Chapter "Privacy-preserving distributed deep learning methods for multi-institution training without sharing patient data" in book titled "Artificial Intelligence in Medicine: Technical Basis and Clinical Applications", Authors: Ken Chang, Praveer Singh, Praneeth Vepakomma, Maarten G. Poirot, Ramesh Raskar, DanielL.Rubin, Jayashree Kalpathy-Cramer, Editor: Lei Xing/Stanford

(2020) Towards split learning at scale: System design. Iker Rodríguez, Eduardo Muñagorri, Alberto Roman, Abhishek Singh, Praneeth Vepakomma and Ramesh Raskar, Workshop on MLOps Systems To be held along with Third Conference on Machine Learning and Systems (MLSys)

(2019) Advances and open problems in federated learning, with 58 authors from 25 institutions.

(2019) Split learning for health: Distributed deep learning without sharing raw patient data, Praneeth Vepakomma, Otkrist Gupta (LendBuzz/MIT), Tristan Swedish (MIT), Ramesh Raskar (MIT),  Accepted to ICLR 2019 Workshop on AI for social good. (Project Page for Split Learning

Press Coverage: MIT Technology Review https://www.technologyreview.com/the-download/612567/a-new-ai-method-can-train-on-medical-records-without-revealing-patient-data/ 

Project Page: https://splitlearning.github.io/

Invited Talk: Workshop on Federated Learning and Analytics (FL-IBM’20), IBM Thomas J Watson Research Center, Title: Split Learning: A new resource efficient alternative for distributed machine learning, https://federated-learning.bitbucket.io/ibm2020/    

(2019) Detailed comparison of communication efficiency of split learning and federated learning Abhishek Singh, Praneeth Vepakomma, Otkrist Gupta, Ramesh Raskar

(2019) ExpertMatcher: Automating ML Model Selection for Clients using Hidden Representations, Vivek Sharma, Praneeth Vepakomma, Tristan Swedish, Ken Chang, Jayashree Kalpathy-Cramer, Ramesh Raskar

(2019) ExpertMatcher: Automating ML Model Selection for Users in Resource Constrained Countries, Vivek Sharma, Praneeth Vepakomma, Tristan Swedish, Ken Chang, Jayashree Kalpathy-Cramer, Ramesh Raskar

Recent talk on Split Learning at Datacouncil.ai SF 2019 (Slides)

(2019) Praneeth Vepakomma and Yulia Kempner (HIT, Israel), “Diverse data selection via combinatorial quasi-concavity of distance covariance: A polynomial time global minimax algorithm”, Discrete Applied Mathematics


(2018) Praneeth Vepakomma, Chetan Tonde (Rutgers, Amazon) and Ahmed Elgammal (Rutgers) "Supervised Dimensionality Reduction via Distance Correlation Maximization”, Electronic Journal of Statistics

(2018) Sai Sri Sathya , Praneeth Vepakomma, Ramesh Raskar, Ranjan Ramachandra, and Santanu Bhattacharya (Rutgers) “A review of homomorphic encryption libraries for secure computation”, Electronic Journal of Statistics

(2016) Susovan Pal (UCLA/Ecole Polytechnique, France) and Praneeth Vepakomma, "Optimal bandwidth estimation for a fast manifold learning algorithm to detect circular structure in high-dimensional data".

(2016) Praneeth Vepakomma and Ahmed Elgammal (Rutgers) "A Fast Algorithm for Manifold Learning by Posing it as a Symmetric Diagonally Dominant Linear System", Applied and Computational Harmonic Analysis.

My Math talk at Rutgers University Math Dept: Part 1

Part 2 (Contd)                                               



Other Smaller Conference/Workshop Papers:

Iterative Embedding with Robust Correction using Feedback of Error Observed, Praneeth Vepakomma & Ahmed Elgammal at  International Conference on Machine Learning, Machine Learning for Interactive Systems Workshop, at Lille, France (ICML Workshop)

Distance Correlation Maximization using Graph Laplacians, Praneeth Vepakomma, Chetan Tonde & Ahmed Elgammal , New England Machine Learning 2014 at Microsoft Research, New England.

A-Wristocracy: Deep Learning on Wrist-worn Sensing for Recognition of User Complex Activities, Praneeth Vepakomma, Debraj De, Sajal K Das, Shekhar Bhansali 2015, IEEE Body Sensor Networks Conference, MIT Media Lab

Embedding Super-Symmetric Tensors of Higher-Order Similarities of High-Dimensional Data, Praneeth Vepakomma, Ahmed Elgammal, Tensor Methods for Machine Learning at European Conference on Machine Learning (ECML Workshop)

Tech report: 

 "Scoring Practices for Remote Sensing of Land Mines", Mathematical Problems in Industry, Duke University, Slides.

Other:

Mentored 65+ students on Springboard in data science

(Interviewed in book:) Data scientist: the definitive guide to becoming a data scientist

Consulting provided in data science to various corporate organizations

Member: Data Driven Justice Tech Consortium, Chicago 2016. 

Training/Participation in Summer Schools, training and workshops:

i) NSF-CBMS 2016 Regional Conference (Summer School) on Topological Data Analysis at University of Texas, Austin, May 31 to June 4 2016. 

ii) 2016 Mathematical Problems in Industry (MPI) interactive workshop, to be held at Duke University on June 13-17, 2016 with travel/accommodation grant. (WPI/Duke Travel Grant) 

iii) Machine learning and physical models at IPAM, UCLA, Los Angeles, 2016.  

iv) Data Science for Social Good, University of Chicago, 2016 (Motorola Solutions, Travel Sponsor) 

v) Joint Statistical Meetings, JSM 2016, Chicago. (Motorola Solutions, Travel Sponsor) 

vi) Workshop on Interface of Statistics and Optimization, SAMSI, Duke University, Feb 8-10 2017 (SAMSI Travel Grant).

vii) Functional Near-Infrared Spectroscopy Symposium, Connectivity Course: Structural and Functional Brain Connectivity via MRI and fMRI (Numerdox Travel Fund), Boston Univ. & Harvard Univ.

© Vepakomma, Praneeth 2017