Publications

[Preprints] [Journals] [Conferences] [Workshops] [Thesis] [Talks] [Patents]

The list of publications are also available on Google scholar and Semantic scholar.

Preprints

P. Jawanpuria, N T V Satya Dev, A. Kunchukuttan, and B. Mishra. Learning Geometric Word Meta-Embeddings. Technical report, arXiv preprint, arXiv:2004.09219, 2020.
[arXiv:2004.09219].

P. Jawanpuria, M. Meghwanshi, and B. Mishra. A Simple Approach to Learning Unsupervised Multilingual Embeddings. Technical report, arXiv preprint, arXiv:2004.05991, 2020.
[arXiv:2004.05991].

J. Saketha Nath and P. Jawanpuria. Statistical Optimal Transport posed as Learning Kernel Embedding. Technical report, arXiv preprint, arXiv:2002.03179, 2020.
[arXiv:2002.03179].

B. Mishra, H. Kasai, and P. Jawanpuria. Riemannian optimization on the simplex of positive definite matrices. Technical report, arXiv preprint, arXiv:1906.10436, 2019.
[arXiv:1906.10436].

Journal articles

B. Mishra, H. Kasai, P. Jawanpuria, and A. Saroop. A Riemannian gossip approach to decentralized subspace learning on Grassmann manifold. Machine Learning Journal (MLJ), 2019.
[Local copy][Online html version][Matrix completion code][Multitask code].

P. Jawanpuria, A. Balgovind, A. Kunchukuttan, and B. Mishra. Learning multilingual word embeddings in a latent metric space: a geometric approach. Transactions of the Association for Computational Linguistics (TACL), 7:107-120, 2019.
[Local copy][Publisher’s copy][Codes@Github].

P. Jawanpuria, J. Saketha Nath, and Ganesh Ramakrishnan. Generalized hierarchical kernel learning. Journal of Machine Learning Research (JMLR), 16(Mar):617−652, 2015.
[Local copy][Publisher’s copy][Matlab codes].

Conference proceedings

P. Jawanpuria, M. Meghwanshi, and B. Mishra. Geometry-aware domain adaptation for unsupervised alignment of word embeddings. In Annual Meeting of the Association for Computational Linguistics (Short Papers), 2020.
[arXiv:2004.08243].

P. Jawanpuria, M. Meghwanshi, and B. Mishra. Low-rank approximations of hyperbolic embeddings. In IEEE Conference on Decision and Control (IEEE CDC), 2019.
[Local copy][arXiv:1903.07307][Matlab codes].

H. Kasai, P. Jawanpuria, and B. Mishra. Riemannian adaptive stochastic gradient algorithms on matrix manifolds. In International Conference on Machine Learning (ICML), 2019.
[Local copy][Publisher’s copy][Matlab codes].
Longer version: [arXiv:1902.01144]

M. Nimishakavi, P. Jawanpuria and B. Mishra. A dual framework for low-rank tensor completion. In Conference on Neural Information Processing Systems (NeurIPS), 2018 (formerly called Advances in Neural Information Processing Systems).
[Local copy][Publisher’s copy][Codes@Github].
Longer version: [Local copy][arXiv:1712.01193]

P. Jawanpuria and B. Mishra. A unified framework for structured low-rank matrix learning. In International Conference on Machine Learning (ICML), 2018.
[Local copy][Publisher’s copy][Matlab codes][Slides][Poster].
Longer version: [Local copy][arXiv:1704.07352].

P. Jawanpuria, Maksim Lapin, Matthias Hein and Bernt Schiele. Efficient output kernel learning for multiple tasks. In Advances in Neural Information Processing Systems (NIPS), 2015.
[Local copy][Publisher’s copy][Long version][Matlab codes].

P. Jawanpuria, Manik Varma and J. Saketha Nath. On p-norm path following in multiple kernel learning for non-linear feature selection. In International Conference on Machine Learning (ICML), 2014.
[Local copy][Publisher’s copy].

P. Jawanpuria and J. Saketha Nath. A convex feature learning formulation for latent task structure discovery. In International Conference on Machine Learning (ICML), 2012.
[Local copy][Publisher’s copy][Supplementary material][Matlab codes].

P. Jawanpuria, J Saketha Nath and Ganesh Ramakrishnan. Efficient rule ensemble learning using hierarchical kernels. In International Conference on Machine Learning (ICML), 2011.
[Local copy][Publisher’s copy][Technical report].

P. Jawanpuria and J. Saketha Nath. Multi-task multiple kernel learning. In SIAM International Conference on Data Mining (SDM), 2011.
[Local copy][Matlab codes].

Workshop papers

M. Meghawanshi, P. Jawanpuria, A. Kunchukuttan, H. Kasai, and B. Mishra. McTorch, a manifold optimization library for deep learning. In NeurIPS workshop on Machine Learning Open Source Software (MLOSS), 2018.
[arXiv:1810.01811][Project page].

M. Bhutani, P. Jawanpuria, H. Kasai, and B. Mishra. Low-rank geometric mean metric learning. In ICML workshop on Geometry in Machine Learning (GiMLi), 2018.
[Local copy][arXiv:1806.05454][Poster].

M. Nimishakavi, P. Jawanpuria, and B. Mishra. A dual framework for low rank tensor completion. In NIPS workshop on Synergies in Geometric Data Analysis, 2017.
[Local copy][Slides][Poster]. An extended version published in NeurIPS’18.

P. Jawanpuria and B. Mishra. A unified framework for structured low-rank matrix learning. In NIPS workshop on Optimization for Machine Learning, 2017.
[Local copy][Poster]. An extended version published in ICML’18.

Ph.D. Thesis

P. Jawanpuria. Learning kernels for multiple predictive tasks.
Ph.D. thesis, IIT Bombay, 2015 [Local copy]. The jury included Kilian Q. Weinberger, Harish Karnick, Anirban Dasgupta, Ganesh Ramakrishnan, Sunita Sarawagi, Pushpak Bhattacharya, and Saketh Nath.

Presentations

Talk at IIT Delhi and IIT Roorkee (2019) on optimization on manifolds and its application to learning multilingual word embeddings.

Talk at machine learning summer school (2019) on learning cross-lingual mappings of word embeddings.

Talk at Microsoft’s MLADS-SYNAPSE (2019) on our TACL work “Learning multilingual word embeddings in a latent metric space: a geometric approach”.

Talk at DiDi AI Lab in Los Angeles (2019) on optimization on manifolds and its application to learning multilingual word embeddings.

Talk at IIIT Hyderabad (2019) on low-rank matrix learning from machine learning a optimization perspective.

In ICML 2018, presented our work “A unified framework for structured low-rank matrix learning” [Slides].

Talks at IIT Madras and IIT Bombay (2018) on large-scale structured low-rank matrix learning [IITM slides][IITB slides].

In ICML 2014, presented our work “On p-norm path following in multiple kernel learning for non-linear feature selection” [Slides].

In ICML 2012, presented our work “A convex feature learning formulation for latent task structure discovery” [Slides].

Patents

A patent filed on chat group recommendations for chat applications.

A patent filed on form conversion.

A patent filed on camera application of mobile phones.