If you have a Google Wave account, then try my Robot extension

I wrote a Robot extension that reads text on a wave (and any child blips added to it), and adds its own child blips with some NLP analysis (entity extraction and auto-tagging). No big deal, but fun to write.

Give it a try: knowledge-books@appspot.com

Comments

  1. mb9:17 PM

    Mark, I added your bot to my wave, and it's going berserk. It's created 50+ blips on my wave, and they're all empty:

    knowledge-books@appspot.com:
    blip submitted events: root blip: title: text:
    results: NLP Analysis:
    Summary:
    ...
    Auto tags: []
    Names: []
    Places: []
    Companies: []
    Products: []

    ReplyDelete
  2. mb9:27 PM

    Mark, I tried the NLP demo on http://knowledge-books.appspot.com

    For some reason it thinks every text block refers to computers and AI.

    Here's a paragraph from the Kama Sutra -- for some reason it's getting tagged with "computers", "ai_nlp" and "c++"

    -----

    Those who believe that destiny is the prime mover of all things say: We should not exert ourselves to acquire wealth, for sometimes it is not acquired although we strive to get it, while at other times it comes to us of itself without any exertion on our part. Everything is therefore in the power of destiny, who is the lord of gain and loss, of success and defeat, of pleasure and pain. Thus we see that Bali 3 was raised to the throne of Indra by destiny, and was also put down by the same power, and it is destiny only that call reinstate him.

    Auto tags: [computers:1.0, religion:0.3575419, computers_ai_textmining:0.3575419, religion_christianity:0.31284916, religion_islam:0.25139666, computers_ai_nlp:0.21787709, computers_programming_c++:0.2122905

    ReplyDelete
  3. Mark Watson, author and consultant6:19 AM

    Hello MB, thanks for giving it a try.

    I had to use a very small in-memory linguistic database in order to meet memory limits (and response limits) of the Java AppEngine. Both the robot and the demo form use this tiny database. I really need to convert from an in-memory store (actually, a few hundred megabytes is all I need - that is why I use in-memory data structures) to use persistent storage and JDO. So, the results are bad for now.

    re: getting 50 blips - I have not seen that, but I am looking at the problem right now. My intention was for one robot blip to be created for each original wave, and then for each non-robot added blip (this is where the problem is).

    Anyway, I am very excited by the possibilities of Wave, but I am still going through a learning curve.

    Best regards,
    Mark

    ReplyDelete
  4. Mark Watson, author and consultant7:45 AM

    Hello MB,

    Please give it another try. I modified:

    1. still using a tiny linguistic database, but I found and fixed a bug in the code that created the tiny database.

    2. I now no longer have the robot reply to empty waves (this was the issue of you getting 50 replies to your empty wave)

    3. I now no longer have the robot reply to blips added to an original wave.

    -Mark

    ReplyDelete
  5. mb8:01 AM

    Tried it again, but now there are no replies from your bot.

    ReplyDelete
  6. Mark Watson, author and consultant8:10 AM

    Hello MB,

    That is correct: I just looked at my AppEngine logs: it looks like you created a blank wave so the robot did not respond to it. If you create a new wave, add some text to it, click done, then when asked to invite participants add my robot, it should work fine.

    -Mark

    ReplyDelete
  7. mb9:33 PM

    Sorry, Mark, it's not working.

    I replied to your bot on an existing wave, and there was no response.

    Then created a new wave, added your bot, got the "I'm alive" blip, replied to that blip, and got nothing back.

    ReplyDelete
  8. mb9:41 PM

    Update -- I added more blips to the wave shared with your bot, and got no responses.

    Then when I edited the root wave, I got 20+ blips from your bot.

    ReplyDelete
  9. Mark Watson, author and consultant6:11 AM

    Hello Mahlon,

    I just created a Wave and invited you to participate so we can discuss this a little more interactively. From my logs, it looks like you access my robot about once a second; are you trying my robot from a robot of your own? Anyway, please accept my Wave invite.

    I disabled creating replies to blips, BTW. My intention, at least for now, is for my robot to only reply when it is added to a newly created wave.

    Thanks,
    Mark

    Thanks,
    Mark

    ReplyDelete
  10. Mark Watson, author and consultant6:18 AM

    Hello again Mahlon,

    I discovered the problem: my bug, or rather my misunderstanding.

    I had enabled the capability:

    capability name="document_changed" content="true"


    So if my robot was added to a blib, then text added, then my robot responded for just about every keypress while editing the blip/wave.

    I am now changing it to ony respond later the document is created and the user clicks "Done".

    Thanks,
    Mark

    ReplyDelete
  11. Mark Watson, author and consultant6:21 AM

    Mahlon, I just tested with the "document_changed" capability removed - now the robot is behaving the way I want it to :-)

    When I tested, I created a Wave, clicked "Done", then added my robot -- that is why I did not see this problem.

    Thanks again for your help,
    Mark

    ReplyDelete

Post a Comment

Popular posts from this blog

Custom built SBCL and using spaCy and TensorFlow in Common Lisp

Here are some of my of my recent notes that might save you some time, or teach you a new trick. I have had good results using the py4cl library if I wrap API calls to TensorFlow or spaCy in a short Python library that calls Python libraries and returns results in simple types like strings and dictionaries. I just committed a complete example (Python library and Common Lisp client code) to the public repo for my book  Loving Common Lisp, or the Savvy Programmer's Secret Weapon that will be added to the next edition of my book. Here is a link to the subdirectory with this new example in my repo: https://github.com/mark-watson/loving-common-lisp/tree/master/src/spacy I frequently make standalone executable programs using SBCL and I just noticed a great tip from  Zach Beane for compressing the size of standalone executables. Start with rebuilding SBCL from source to add the compression option; get the source code and: ./make.sh --with-sb-thread --with-sb-core-compression sh in
Read more

I have tried to take advantage of extra time during the COVID-19 pandemic

My wife Carol and I have been practicing social distancing and wearing masks for shopping for over 5 months now. Welcome to the new normal and a crazy world in which entertaining and seeing friends is done by meeting in people's yards and everyone bringing their own "meal in a bag." I enjoy writing so I have been updating my recent books, starting with Loving Common Lisp, or the Savvy Programmer's Secret Weapon and  A Lisp Programmer Living in Python-Land: The Hy Programming Language . These are free to read online and licensed with Creative Commons Share and Share Alike, No Commercial Reuse, so you can also find copies on the web (hopefully up to date copies!). Last month I started a much larger project: I have not updated my book Practical Artificial Intelligence Programming With Java since the fourth edition was published in 2013. I have discarded a lot of older material like exert systems, and have three new chapters on the semantic web and also a new chapter on
Read more

GANs and other deep learning models for cooking recipes

I retired this spring after working on artificial intelligence projects since the 1980s. Freedom from having to work on large projects for other people and companies is liberating and frees up time for thinking about new ideas. Currently I am most interested in deep learning models for generating and evaluating recipes - for now I am using a GAN model (which I am calling RecipeGAN). When I managed a deep learning team at Capital One, I used GANs to synthesize data. During a Saturday morning quiet-time hacking sprint the first month at my new job, I had the idea to take an example program SimpleGAN that generated MINST digits and instead generate numeric spreadsheet data (using the Wisconsin Cancer Data Set that I had previously used in my books as example machine learning data). I was really surprised how well this worked: I could generate fake Wisconsin cancer data, train a classification model on the fake data, and get classification prediction accuracy on real data samples that wa
Read more