Natural Language Processing (NLP)


NLP is the application of AI to understand human language as its spoken/written. Historically humans have interacted with computers using "languages" that computers can understand. These languages are highly structured, precise and devoid of ambiguity. While computers using such languages have brought about tremendous social and economical benefits, as we are seeing all around us, use of computers have always been limited when it came to applications that required understanding of written and spoken human language as its normally done between two humans.



For instance, while databases have evolved in different branches and their storing and retrieval and speeds, method of organization, processing time and several other aspects have improved manifolds, it still requires the use of syntactical "languages" to manipulate the databases (add, retrieve, update, etc.). Would it not be nice if a human can write/speak a statement that can directly query a database and expect the query to be translated correctly? This is one such situation where NLP comes is.



You could ponder over the following:

  • How would it be if a computer can understand your mood from your tone and sentence construction?
  • How would it be if you could automate your entire customer service/support mechanism (which are anyways being done) and do so with emotions (sympathy, empathy and concesrn)?
  • How would it be if you could have documents sent to you automatically processed and just the “intent” conveyed to you? Imagine the reduction in the number of legal and contractual documents that you have to read as part of your work life?
  • How would it be if while you were in the midst of your household chores, talk to an “agent” that can help you plan a holiday like a expert and have your tickets booked?

  • The applications are endless...



    Whirldata understands that NLP is a fledgling concept but with a very valuable application potential. Whirldata ensures a sustained in-house Research and Development effort for building NLP-based products/solutions in the financial/retail/energy/healthcare/legal domain. Some products that are in the pipeline - Deep/Shallow Question Answering, Search, ChatBots to enable various operational tasks, client assistance based on inferred client needs, Virtual Assistant to sales teams and advisors/consultants.



    Every member of Whirldata’s NLP development/testing team is very curious, possess an outstanding sense of humour, possess the inherent ability to share knowledge on a sustained basis and of course technically sound. After all they need to indulge in hours and hours of conversations with the inanimate!



    Knowledge/Expertise areas:

  • Natural Language Toolkit (NLTK): The comprehensive toolset for the entire range of NLP needs.
  • Pattern – Set of web mining tools for NLP
  • TextBlob – NLP API built on Pattern & NLTK .
  • Scikit-learn: Machine learning libraries Python
  • spaCy – NLP with Python & Cython.
  • Stanford Core NLP
  • Gensim – Vector space and topic modeling toolkit
  • LP / IR like OpenNLP, SOLR, CoreNLP
  • Big Data platforms like Hadoop/Mahout, Spark/MLlib
  • Amazon Web Services