Welcome to my page

sarva-dharmān parityajya mām ekaḿ śaraṇaḿ vraja ahaḿ tvāḿ sarva-pāpebhyo mokṣayiṣyāmi mā śucaḥ

Adepu Ravi Sankar

PhD

I am a research scientist at Amazon, India. I work towards giving Machine Learning solutions to industry problems. Prior to Amazon, I did my post-doc at Verisk Analytics Inc. I obtained my Ph.D in Computer Science & Engineering department at the Indian Institute of Technlogy Hyderabad advised by Dr. Vineeth N Balasubramanian. My thesis is on 'Understanding the Loss Surface of Deep Neural Networks'. My research is focused on the design of algorithms for deep learning models with inferences from loss landscape geometry. My Ph.D. work brings together tools from convex optimization, numerical linear algebra, learning theory.

[Old] For more on Deep Learning follow my blog http://mydeeplearning.blogspot.com/("Awarded Feedspot Top 40 Neural Network Blogs")

I organized a reading group on Convex Optimization in the Summer of 2017, you can find the recorded lectures here.

Scribe for the course Convex Optimization: Theory, Click here for the scribe

Education

  1. PhD: Indian Institute of Technology Hyderabad, CSE, August 2014 - July 2020
    Thesis title: "Understanding the loss surface of deep neural networks", advised by Dr Vineeth NB
  2. M.Tech: Indian Institute of Technology Hyderabad, CSE, 2011 - 2014
    My masters thesis is on "Temporal Coherence in Energy-based Deep Learning Machines for Action Recognition", advised by Dr Vineeth NB

Recognitions

  1. Recipient of ICML 2018 volunteer award
  2. Recipient of Microsoft Research travel grant to attend ICML 2018
  3. Recipient of Intel India PhD fellowship 2015 awarded by Intel India
  4. Recipient of Microsoft Research travel grant to attend ACML-2015 at HongKong
  5. NVIDIA Award for Best Paper using GPU Technologies, IEEE International Conference on High Performance Computing (HiPC), Dec 2014
  6. FlySky Incubation Award, 2012 (IIT-H Startup Award)

Publications

  1. Adepu Ravi Sankar, Yash Khasbage, Rahul Vigneswaran, Vineeth NB, A Deeper Look at the Hessian Eigenspectrum of Deep Neural Networks and its Applications to Regularization, Accepted at AAAI 2021
  2. Sneha Kudugunta, Vaibhav B Sinha, Adepu Ravi Sankar, Surya Teja Chavali, Purushottam Kar, Vineeth N Balasubramanian, DANTE: Deep AlterNations for Training nEural networks, Elsevier Neural Networks 2020 (Impact Factor 5.5) PDF
  3. A. Ravi Sankar, Vishwak S., V. Balasubramanian, On the Analysis of Trajectories of Gradient Descent in the Optimization of Deep Neural Networks, Theory of Deep learning workshop and also in Modern Trends in Nonconvex Optimization for Machine Learning , ICML 2018 PDF
  4. A. Ravi Sankar, V. Balasubramanian, Are Saddles Good Enough for Deep Learning, arXiv:1706.02052, Proceedings of ACM IKDD Joint International Conference on Data Science & Management of Data (CoDS-COMAD’18), Jan 2018 .
  5. Vishwak S., A. Ravi Sankar, V. Balasubramanian, ADINE: An Adaptive Momentum Method for Stochastic Gradient Descent, Proceedings of ACM IKDD Joint International Conference on Data Science & Management of Data (CoDS-COMAD’18), Jan 2018 PDF
  6. Adepu Ravi Sankar, Vineeth N Balasubramanian, Similarity-based Contrastive Divergence Methods for Energy-based Deep Learning Models, The 7th Asian Conference on Machine Learning (ACML-2015), Hongkong 20-22 November 2015.(29% acceptance rate) PDF
  7. S. Chakraborty, V. Balasubramanian, Adepu Ravi Sankar, S. Panchanathan, J. Ye, BatchRank: A Novel Batch Mode Active Learning Framework for Hierarchical Classification, In Proceedings of the 21st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2015. (19.4% acceptance rate)
  8. Sai Rajeshwar, A Ravi Sankar, V. Balasubramanian, C.D. Sudheer, Scaling up the training of Deep CNNs for Human Action Recognition, 4th International Workshop on Parallel and Distributed Computing for Large Scale Machine Learning and Big Data Analytics (in conjunction with IPDPS 2015), Hyderabad, India, 2015.
  9. Sai Rajeshwar, A Ravi Sankar, V. Balasubramanian, C.D. Sudheer, Parallel Learning of Deep Convolutional Neural networks and its Application to Action Recognition, IEEE International Conference on High Performance Computing - Student Research Symposium, Goa, India, 2014 (Best paper award)

Other Technical Reports

  1. Sahil Manocha, Adepu Ravi Sankar, Vineeth N Balasubramanian, Dissimilarity Based Contrastive Divergence for Anomaly Detection, Appearing in Proceedings of the 2nd Indian Workshop on Machine Learning, IIT Kanpur, India, 2016.
    PDF, Source code
  2. Adepu Ravi Sankar, Vineeth N Balasubramanian, Similarity-based Contrastive Divergence Methods for Energy-based Deep Learning Model, In Xerox Research Centre India Research Symposium (XRCI Open 2016), Jan. 2016

Professional Service

  1. Reviewer: ACML-22, CVPR-22, AAAI-21, IJCAI-20, ADCOM-18, NCC-18, IEEE-TVT, IEEE-TNNLS
  2. Sub-reviewer: NeurIPS-20, SDM-20, ECML-PKDD-19,AAAI-19, ADCOM-19, IJCAI-18, WACV-17
  3. Head Volunteer, 2nd Department CSE Day-IIT Hyderabad held on 22nd October 2016.
  4. Local Organizing Committee, Conformal Prediction for Reliable Machine Learning workshop, 15th - 17th December 2015, IIT Hyderabad.

Teaching Assistant

  1. Optimization Methods in Machine Learning (CS6230)-Fall 2018 (Co-Instructor, Link to video lectures)
  2. Optimization Methods in Machine Learning (CS6230), Fall 2016 & 2017
  3. Guest lectures in Deep Learning (CS5480) Spring 2018 & Applied Machine Learning (CS6510) Fall 2016
  4. Optimization Methods in Machine Learning (CS6230), Fall 2016 & 2017
  5. Computer Vision (CS5290), Fall 2015
  6. Numerical Linear Algebra for Data Analysis (CS5270), Spring 2015
  7. Introduction to Database Management Systems (CS3010), Spring 2014

Programming Tools

  1. Programming Languages: C, C++, Matlab, Python
  2. Deep Learning Tools: Caffe, Torch, Pylearn2, Theano
  3. Big data tools / misc: Hive, Spark sql, hadoop, amazon aws tools

Internships

  1. Amazon Machine Learning Lab, Bangalore: I worked on MyFit project during summer of 2016 for automatic size recommendation of Apparels

Summer School / Workshops

  1. Participated in discussion meeting on "The Theoretical Basis of Machine Learning (ML)" held at ICTS Bangalore, Dec 27-29, 2019.
  2. Invited to participate in "ACM-MSR: Academic Research Summit 2018: A Future with AI", held at IIIT Hyderabad, Jan 24-25 2018.
  3. Invited to participate in Amazon India Artificial Intelligence Summit held at Bangalore, 2017
  4. Invited and participated in ACM-MSR Academic Research Summit on29th-30th of January, 2016 in Pune.
  5. Selected and participated in XRCI Open 2016 in Bangalore.
  6. Invited and participated in Intel India Academic Forum 2015 at J W Marriot hotel Delhi.
  7. Selected and participated in Microsoft Research India Summer School on Machine Learning 2015 at IISc Bangalore.
  8. Selected to participate in Summer School on Non-Convex Optimization for Machine Learning 2015 at IIT Bombay.
  9. Selected and participated in Symposium on Learning, Algorithms and Complexity 2015 at IISc Bangalore.
  10. Participated in WiSSAP 2014 on Deep Learning for Multi-Lingual Speech Processing at IIIT Hyd.
  11. Selected and participated in the IMPECS School on Advanced Algorithms-2013 at IIITDM, Jabalpur.
  12. Selected and participated in IBM I-Care Winter School-2012 on Big Data Analytics at IISc Bangalore.

Professional Activities

  1. Local organizing committee for Conformal Prediction for Reliable Machine Learning workshop,15th-17th December 2015, IIT Hyderabad

Other relevant projects

  1. e-Drishti(Jan-2015 - Jun 2015)
  2. HMM-MLP-HMM Hybrid Architecture for Action Recognition (Aug 2014 - Dec 2014)
  3. Image Quality Assessment using Deep Learning (Jan 2014 - Apr 2014)
  4. Sports video classification using SVMs
  5. Analyzing the performance of various kernel functions for classification of video data

Positions / Responsibilities

  1. Treasurer for Spic Macay IIT Hyd chapter (Nov 2015 - Nov 2016)
  2. CSE Department PhD Representative, Student Gymkhana of IIT-H (May 2015 - May 2016)
  3. Student representative in Senate (Academic body of IIT-H) (Apr 2013 - Jun 2014)
  4. Student representative in Disciplinary Action Committee of IIT-H (Apr 2013 - Jun 2014)
  5. Postgraduate Representative, Student Gymkhana of IIT-H (Apr 2013 - Jun 2014)
  6. Mess Secretary, Student Gymkhana of IIT-H (Apr 2012 - Apr 2013)
free hits
Track My Website