• codyguy

    Would love to hear your thoughts. Here is a direct trial link.
    http://classy.thatneedle.com

    posted in Product read more
  • codyguy

    http://www.thatneedle.com/natural-language-classifier.html

    This is a natural language classifier trained for classifying emails into refunds, booking or other. Please try out the demo. Your feedback is very welcome.
    Based on our solid natural language processing kernel. We are in a position to customize text classifier for handling more categories/sub-categories or for different verticals (at a cost of course).

    posted in Product read more
  • codyguy

    @rudimk The system learning indeed and going deep. What I am really pleased about is the accuracy of the deductions. Please challenge the system notcruders!
    Let me know if you see anything funny.

    posted in Tech read more
  • codyguy

    Magical ! NLP API just demonstrated real time deduction. Basic but accurate.
    example query
    looking for a cheap frost free
    output was:
    {"item": "refrigerator", "other_attr": ["frost free"], "hint": ["deal"]}

    This will keep getting better. Please try it out and report using the red button!!
    http://thatneedle.com/filterd

    posted in Tech read more
  • codyguy

    @paraschopra Not tried it yet. Too many things on my plate right now with ThatNeedle API.

    posted in Tech read more
  • codyguy

    This link has some comparison benchmarks for Theano, Torch and others but tensorflow should be added in time. I'll try to update when tensorflow (and hopefully Veles) gets added to the list of benchmarked tools.
    https://github.com/soumith/convnet-benchmarks

    posted in Tech read more
  • codyguy

    Google announced TensorFlow : http://tensorflow.org/
    Samsung announced Veles : https://velesnet.ml/
    Interesting times ahead for Machine learning and AI.

    posted in Tech read more
  • codyguy

    Thank you. And it's getting better every day!!

    posted in Tech read more
  • codyguy

    Thank you for featuring us in your newsletter.

    posted in Tech read more
  • codyguy

    What do think about the pricing, should we bump it up? What would you be willing to pay for something like this?

    posted in Tech read more
  • codyguy

    @Rohit-Agarwal : Yes I would be very happy to share some insights. Probably good for a blog post. I'll keep that on my todo list !

    posted in Tech read more
  • codyguy

    @rudimk : Thank you for your encouraging words. We are not covering books & movies for now. Have a grand plan around book search which will deviate from how it's usually done in the industry today. I should probably mention scope on the demo page itself. (Like try for electronics etc)

    Yes, that's the value. Any retailer can just apply this thin layer on their queries and improve their search experience. Instead of being a difficult niche area for developers, NLP could just be a simple blackbox component. That's how it should always have been.
    Interesting part is, I started out looking for such a component over 5 years ago but didn't see anything that was fast and accurate.

    posted in Tech read more
  • codyguy

    @rudimk Here's the direct trial link for Retail API:
    http://www.thatneedle.com/filterd

    Please report errors by clicking on the red button. I haven't provided a space for comments just yet. But would appreciate details/comments. You can also reach me by
    email: mail@thatneedle.com

    posted in Tech read more
  • codyguy

    @paraschopra Here's the direct trial link for Retail API:
    http://www.thatneedle.com/filterd

    Please report errors by clicking on the red button. I haven't provided a space for comments just yet. But would appreciate details/comments. You can also reach me by
    email: mail@thatneedle.com

    posted in Tech read more
  • codyguy

    @paraschopra : Nothing right now for without-signiup trial of the API. I am pondering if I should create one and make it public. Please give me a few days.
    What do you think are the pros-cons of a signup-free trial?

    posted in Tech read more
Internal error.

Oops! Looks like something went wrong!