Information Retrieval using Deep Learning
Submitted by shashank gupta (@shash273) on Wednesday, 26 April 2017
Neural networks are current state-of-the-art in almost all Computer Vision, Natural Language Processing and Speech tasks. Convolution Neural Networks, a deep learning model are go-to choice in Computer Vision. Similarly Recurrent Neural Networks (RNNs) are popular choice in NLP. The area of information retrieval is no different. Neural nets are slowly progressing towards becoming state-of-the-art in this field too. In this talk, we will discuss how deep learning can be used to design information retrieval systems. We will start with introduction to some key concepts in Neural IR followed by a practical example of Content based article recommendation system using deep learning.
We will cover the following topics :
- Introduction to information Retrieval
- Word Embeddings
- Word Embeddings for IR
- Introduction to Deep Neural Nets
- Metric Learning
- Deep Learning for IR (Siamese network)
- Practical examples
7.1. Content based article recommendation system
In the practical example, we will go through how to apply the concepts learned to build a content based article recommendation system.
Shashank is a MS by Resaerch student in Information Retrieval and Extraction lab (search.iiit.ac.in) at IIIT-Hyderabad.
He is working towards the application of Deep Learning in Social Media Analytics and Inforamation Retrieval. He recently published two papers in World Wide Web (WWW) conference held in Perth, Australia. WWW is a very prestigious and top tier conference in the area of Web and Social Media Analytics. His work was on detecting hateful and abusive text from Social media.
His work was awarded the best poster award at the conference.
His WWW work was featured in Times of India and other news outlets too.
Currently he is working on application of Machine learning and Deep Learning in healthcare domain.
His recent publications are:
- Deep Learning for Hate Speech Detection in Tweets.
- Simultaneous Inference of User Representations and Trust.