Welcome to the third module of this course. In this module, you’ll learn Recurrent Neural Networks or RSSs. RSSs are specially designed to work with sequential data, i.e. data where there is a natural notion of a ‘sequence’ such as text (sequences o words, sentences etc.), videos (sequences of images), speech etc. RNNs have been able to produce state-of-the -art results in fields such as natural language processing, computer vision, and time series analysis.
One particular domain RNNs have revolutionised is natural language processing. RNNs have given, and continue to give, state-of-the-art results in areas such as machine translation, sentiment analysis, question answering systems, speech recognition, text summarization, text generation, conversational agents, handwriting analysis and numerous other areas. In computer vision, RNNs are being used in tandem with CNNs in applications such as image and video processing.
Many RNN-based applications have already penetrated consumer products. Take, for example, the auto-reply feature which you see in many chat applications, as shown below:
Similarly, when you talk to a support team of a food delivery app, or any other support team for that matter, you get an auto-reply in the initial stages of your interaction where the support team asks about details such as order date, problem description and other basic things. Many of these conversational systems, informally called ‘chatbots‘, are trained using RNNs.
RNNs are also being used in applications other than NLP. Recently, OpenAI, a non-profit artificial intelligence research company came really close to defeating the world champions of Dota 2, a popular and complex battle arena game. The game was played between a team of five bots (from OpenAI) and a team of five players (world champions). The bots were trained using reinforcement learning and recurrent neural networks.
One particular domain RNNs have revolutionised is natural language processing. RNNs have given, and continue to give, state-of-the-art results in areas such as machine translation, sentiment analysis, question answering systems, speech recognition, text summarization, text generation, conversational agents, handwriting analysis and numerous other areas. In computer vision, RNNs are being used in tandem with CNNs in applications such as image and video processing.
Many RNN-based applications have already penetrated consumer products. Take, for example, the auto-reply feature which you see in many chat applications, as shown below:
There are various companies who are generating music using RNNs. Jukedeck is one such company.
Similarly, when you talk to a support team of a food delivery app, or any other support team for that matter, you get an auto-reply in the initial stages of your interaction where the support team asks about details such as order date, problem description and other basic things. Many of these conversational systems, informally called ‘chatbots‘, are trained using RNNs.
RNNs are also being used in applications other than NLP. Recently, OpenAI, a non-profit artificial intelligence research company came really close to defeating the world champions of Dota 2, a popular and complex battle arena game. The game was played between a team of five bots (from OpenAI) and a team of five players (world champions). The bots were trained using reinforcement learning and recurrent neural networks.
There are various companies who are generating music using RNNs. Jukedeck is one such company.
There are many other problems which are yet to be solved, and RNNs look like a promising candidate to solve them. You could be the one to make an impact in those areas. With that in mind, let’s start the module.
In this session:
- What are sequences?
- The architecture of an RNN
- Types of RNNs
- Drawbacks of RNN and motivation for its other variants
Prerequisites
There are no prerequisites for this session other than knowledge of the previous modules on Neural Networks and the courses on Statistics and ML.
Guidelines for In-Module Questions
The in-video and in-content questions for this module are not graded. Note that graded questions are given on a separate page labelled ‘Graded Questions’ at the end of this session. The graded questions in this session will adhere to the following guidelines:
People you will hear from in this session:
Subject Matter Expert:
Professor, IIIT-Bangalore
The International Institute of Information Technology, Bangalore, commonly known as IIIT Bangalore, is a premier national graduate school in India. Founded in 1999, it offers Integrated M.Tech., M.Tech., M.S. (Research) and PhD programs in the field of Information T
People you will hear from in this session:
Subject Matter Expert:
Professor, IIIT-Bangalore
The International Institute of Information Technology, Bangalore, commonly known as IIIT Bangalore, is a premier national graduate school in India. Founded in 1999, it offers Integrated M.Tech., M.Tech., M.S. (Research) and PhD programs in the field of Information Technology.
Report an error