Bay Area NLP Reading Group
Details
Doors at 6:00 PM, discussion from 6:30-7:30, networking 7:30-8:00.
Venue: 44 Tehama St, San Francisco, CA 94105
This will be the latest session of an ongoing NLP reading group for industry folks, and we will be doing something a little different: Karunakar Singamreddy (SK) will be presenting about his research and work on Question Answering.
Abstract:
Natural Language Processing is solving human needs of language translation, reading comprehension and question answering (QA). This talk will discuss the current state of knowledge in QA and NLP, training neural networks for QA with news sources like CNN and Daily Mail, and progress made by Microsoft and Facebook on QA on other datasets (like bAbi). It will also cover the results of SK's experiment on teaching a computer to comprehend an Indian epic story.
Optional (but recommended) Reading:
Teaching Machines to Read and Comprehend. (https://arxiv.org/pdf/1506.03340.pdf) Karl Moritz Hermann, Toma ́sˇ Kocˇisky ́, Edward Grefenstette, Lasse Espeholt, Will Kay, Mustafa Suleyman and Phil Blunsom. 2015.
Ask Me Anything: Dynamic Memory Networks for Natural Language Processing. (https://arxiv.org/pdf/1506.07285.pdf) Ankit Kumar, Ozan Irsoy, Peter Ondruska, Mohit Iyyer, James Bradbury, Ishaan Gulrajani, Victor Zhong, Romain Paulus and Richard Socher. 2016.
The Goldilocks Principle: Reading children’s books with explicit memory representations. (https://arxiv.org/pdf/1511.02301.pdf) Felix Hill, Antoine Bordes, Sumit Chopra, and Jason Weston. 2015.
---
Karunakar Singamreddy (SK) is a tech start-up entrepreneur and a technologist and is interested in solving industrial problems with machine learning techniques. In his recently acquired start-up, he developed a solution to compare and rank documents. He is also curious about recommendation engines, machine learning techniques and deep learning models. Check him out on Github (https://github.com/skreddy99) and his writing on LinkedIn (http://www.linkedin.com/in/s-k-reddy-3473763/recent-activity/posts/).