In this blog post I discuss the state of "Un-supervised learning". This blog post is motivated from the talk given by Prof Chiranjib Bhattacharyya at ACM-MSR: Academic Research Summit 2018: A Future with AI. This is only of my favorite talk of all thought provoking, stupendous talks held as part of the summit. The below is the excerpt I drew from the talk and I wanted to share with you all. By and large the area of unsupervised learning is still at its initial stage and might take another 40-50 years to reach its pinnacle. One of the main stumbling block in achieving progress in this particul ..read more

In this post, I am sharing few open problems in the area of Deep Learning that needs attention. Explainability: This is one of the main concerns the community is currently facing. A huge amount of criticism from theoretical machine learning researchers is that neural network models are like Black box. The existing models fail to answer the question of "How a decision is being made".  DARPA has initiated a new program to look into this particular issue. I anticipate that in the next 2-3 years in deep learn this area will see new frontiers. Robust Neural Models: This is one other ..read more

Nesterov accelerated gradient(NAG) is one the most successful and also the mysterious working algorithm. This post tries to explain the history behind the invention of this great method and its contributions to bridge the gap between lower bounds in oracle black box optimization. This post is a summary of Chapter 3 from "Convex Optimization: Algorithms and Complexity" We define a function is $\beta$-smooth if the gradient is $\beta$-Lipschitz. A function is $\beta$-Lipschitz if $|| \nabla f(x) -\nabla f(y) || \leq \beta ||x-y|| $, and if the function is twice differentiable then, $f ..read more

Hi, Summer School on Deep Learning for Computer Vision The intended audience of this work shop are fresh PhD students and senior master students. The Scope of Deep Learning is to cover the following ares Fundamentals of Neural Networks Architectures such as - Convolutional Neural Nets - Recurrent Neural Nets - Auto Encoders New Architectures and Applications Recent Advances and Case Studies Summer school will have: Lectures Talks Tutorials Practicals Second Workshop on Machine Learning IIT Kanpur is organizing its second Workshop on Machine Learning from July 1-3, 2016. This works ..read more

Eigen values and Eigen vectors have a huge applications across various spectrum of engineering. Few include the following PCA EIGEN FACES DIAGONALIZATION PAGE RANK CONDITION NUMBER OF MATRIX RANDOM MATRICES CONVERGENCE OF MARKOV MATRIX CONVERGENCE OF ITERATIVE ALGORITHMS POWERS OF A MATRIX SOLVING SYSTEM OF DIFFERENTIAL EQUATIONS SYMMETRIC POWERS OF SYMMETRIC MATRIX STRUCTURAL GEOLOGY TO FIND DIRECTIONS OF PRINCIPAL STRAIN GRAPH PARTITIONING ALGORITHMS ..read more

I always believe in class room lectures than self reading. On the same lines, I recommend the following few good lectures to get a good understanding of Linear Algebra. Linear Algebra is an essential subject for any Machine Learning researcher to master. Here are few Massive open online courses which help you to understand Linear Algebra. Linear Algebra Instructor: Gilbert Strang (If you complete this course, you need not even go beyond this. Gilbert Strang had written a book on Linear Algebra which is followed world wide, If you listen to these lectures and if you can solve the exercise pro ..read more

The following are some good Learning/vision conferences in India that will happen in this year. IEEE Workshop on Computational Intelligence: Theories, Applications and future directions from Dec 14-17 at IIT Kanpur. NCVPRIPG-2015 (National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics) from Dec 16-19 at IIT Patna, India. CVIP-2016 (International conference on computer vision and image processing) from Feb 26-28, 2016 at IIT Roorkee, India iKDD: ACM India special interest group on knowledge discovery and data mining from March 13-16,2016 at Pune ..read more

Microsoft is going to organize a Summer school on Machine Learning in IISc Bangalore. Follow the updates here This is a small write-up about MSR Summer school on ML in IISc Bangalore, based on my experiences. Every speaker who presented there are stalwarts in their particular area. People like Suvrit Sra, Chih-Jen, SanjoyDas who are regarded a lot in the ML community presented their work. This summer school is very special as Microsoft organised this for the first time in India. Researchers working in ML across India in all IIT's, IISc, BITS, ISI came together and attended the summer school ..read more

https://github.com/ChristosChristofidis/awesome-deep-learning Really Awesome Deep learning stuff ..read more

The Deep Learning Summer School 2015 will be held in Montreal in August 2015.  Organizers: Yoshua Bengio, Roland Memisevic, Yann LeCun The list of speakers are very good. https://sites.google.com/site/deeplearningsummerschool ..read more