About Me


Image of Me

I graduated from Indian Institute of Technology, Kanpur in June 2016 and would be joining Professor Philip Torr’s Lab for an internship at Oxford from January. I previously worked at Toyota Technological Institute at Chicago with Professor Mohit Bansal and am currently working as an intern at a startup named We Create Problems.

I have been extremely passionate about Artificial Intelligence especially Deep Learning applications on Computer Vision and Natural Language Processing.

Some of my key projects include :

  • Contextual RNN GAN for Abstract Reasoning Diagram Generation (AAAI-17)
  • Message Passing Multi Agent Generative Adversarial Networks
  • Improving Visual Question Answering with Dense Captions

I have also worked in the field of High Performance Computing and Parallel Programming . I worked in developing a novel parallel algorithm for solving the Boolean Satisfiability Problem and have implemented it on a NVidia GPU.

I had the chance to work in the High Performance Computing Lab at Carnegie Mellon University and we developed a model for estimating the slowdown of an application while running alongside other applications in a multicore processor .It resulted in a Publication in MICRO 2015.

I also have been quite frequent in the field of Competitive Programming , constantly participating in online Coding platforms such as Topcoder and Codeforces .

Apart from the technical side I am an avid quizzer , having been a National Finalist of the Prestigious Tata Crucible Campus Quiz 2014 , a city winner of the Limca Book of Records Quiz 2011 , a runner up of the esteemed TCS IT WIZ .

I have just recently been interested in Photography and have been learning photography ever since.I love to read novels and basically anything which is a treasure of knowledge .I love to play Squash and hit the gym quite frequently .


My major aim is to write about the state of art in Artificial Intelligence in the hope of making it more clear to one and all.



Copyright 2015 Arnab Ghosh. Last update (ET): 2016-12-06 07:48:12 +0000. Atom Feed