Firstly, this is going to be a very long blog.
What is this about?
This is about building a nano-Watson on your laptop that can do deepQA.
Why this experiment?
This is a result of inspiration from a set of blogs by John Miedema. His blogs on building a junior Watson in your basement really kicked me started. (you won't find his blogs anywhere now!!) The book Taming Text was another friend I found very close to heart.
What is deepQA?
DeepQA is a term (never heard before the IBM's Watson hype) that means "answering a question through search and cognitive capabilities". Well, at least that how I define it.
Why the mention of Watson?
This project was inspired by IBM Watson's DeepQA capability. If you haven't yet heard about Watson, please google, there are tons of articles out there.
The system I built is extremely small compared to Watson's space and speed so I thought I could call it nano-Watson.
How long did this project take?
This experiment was done in less than 30 working days and considering my total ignorance in the field of text search and language processing, I say, this was good.
Resources Used:
Hardware:
Software:
All components used are open source.
(These I didn't integrate due to lack of time)
What is this about?
This is about building a nano-Watson on your laptop that can do deepQA.
Why this experiment?
This is a result of inspiration from a set of blogs by John Miedema. His blogs on building a junior Watson in your basement really kicked me started. (you won't find his blogs anywhere now!!) The book Taming Text was another friend I found very close to heart.
What is deepQA?
DeepQA is a term (never heard before the IBM's Watson hype) that means "answering a question through search and cognitive capabilities". Well, at least that how I define it.
Why the mention of Watson?
This project was inspired by IBM Watson's DeepQA capability. If you haven't yet heard about Watson, please google, there are tons of articles out there.
The system I built is extremely small compared to Watson's space and speed so I thought I could call it nano-Watson.
How long did this project take?
This experiment was done in less than 30 working days and considering my total ignorance in the field of text search and language processing, I say, this was good.
Resources Used:
Hardware:
- Any machine with a decent OS on it.
- Not a requirement but a 4 GB RAM will get it rolling really smooth.
Software:
All components used are open source.
- Java Oracle 1.7
- Apache SOLR 4.10
- Apache OpenNLP 1.5.3
- WordNet 3.0
(These I didn't integrate due to lack of time)
- OpenCyC
- Apache Jena
The Story on how I built my nano Watson
Chapter 0: Introduction (You are reading this)
No comments:
Post a Comment