chatbots

Dead Drop no. 3

Posted on by Jon K.

More Bots

http://pushpullfork.com/2017/09/botnet-cometh/

I’m following the public facing war between disinformation and the public. Yes, this is not a conspiracy theory, but an actual war, on social networks, aided and abetted by profiteering conspiracy theorists who want to make millions (in some cases) off paranoia. It’s fascinating because it really means the end of being online as a non-combat participant. Everyone is involved in this.

https://moz.com/blog/chat-bot

https://tutorials.botsfloor.com/opensource-ai-chat-bot-framework-with-natural-language-understanding-and-conversational-abilities-7c6b71e2c461

One of the important aspects of successful bots are the natural chat sophistication. Mapping conversations is part of that. Also, this open sourced framework for conversation construction is going to be extremely helpful.

https://booking.design/7-things-you-need-to-know-about-designing-a-chat-bot-4f07886d3fd0

Amazing how Medium is taking over from where WordPress used to dominate. More tips about chatbot design. Really interesting stuff here, some of it straight forward and obvious, but I really dig reading about how other people made decisions.

https://github.com/google/bottery

Google’s chat engine/bot framework.

No comments yet

Dead Drop no.2

Posted on by Jon K.

A second hodgepodge of things I’ve found on the Internet.

How to Build a Twitter Bot

http://www.zachwhalen.net/posts/how-to-make-a-twitter-bot-with-google-spreadsheets-version-04/

https://www.labnol.org/internet/write-twitter-bot/27902/

https://www.fastcompany.com/3031500/how-twitter-bots-fool-you-into-thinking-they-are-real-people

The third article is the most interesting because it’s from 2014 (remember Klout scores?) and really shows how little (and how much) has changed in the Twitter bot universe. Essentially the heart of the strategy is the same, but mixed with a way to learn (through AI, or natural language processing which is clearly being done at some more sophisticated campaigns) and adapt depending on where we are, I suspect that we are on the cusp of what the future looks like when everyone you meet online could not actually be a person. My Philip K. Dick collection became a lot less sci-fi and more prescient.

More Bots (Cisco Spark/Microsoft Teams):

https://msdn.microsoft.com/en-us/microsoft-teams/botsconversation

https://developer.ciscospark.com/bots.html

Both Cisco Spark and Microsoft Teams are available where I work. Both have vague collaboration mandates, both can be used for communication. At this point, Spark might be more advanced with third party enhancements and bots. Teams might end up being the platform of choice if students are involved. I’m hoping to build a chatbot to handle common sorts of requests that our LMS support team get (ie. the stuff we have a predefined reply in our ticketing system, which isn’t attended to after 4:30 PM).

Natural Language Processing:

https://research.google.com/pubs/NaturalLanguageProcessing.html

https://medium.com/@jrodthoughts/ambiguity-in-natural-language-processing-part-ii-disambiguation-fc165e76a679

https://chatbotsmagazine.com/the-problem-with-chatbots-how-to-make-them-more-human-d7a24c22f51e

Natural language is key for the chatbot above, and of course the first attempt will be ruthlessly primitive I’m sure.

Chatbots in General:

https://www.theguardian.com/technology/2016/sep/18/chatbots-talk-town-interact-humans-technology-silicon-valley

No comments yet