Don Patrick - May 30, 2014:

Are you a roboticist, theorist, psychologist, programmer, or are you in fact a human?

Hello All, I am Alan McDonley and I am an addict - or perhaps “Senior AADDict”.

I get high on seeing a computer do what I told it do, and have an “unrequited” childhood musing on a future with intelligent computers (e.g. “Gay, Bounce!” - late 20s childhood).

In Don’s question ontology, although I have given life to three limited-autonomy mobile robots, and several chatbot knowledge bases, and have a distinct tendency to anthropomorphize computer entities, I cannot claim to be a roboticist;  I am certainly not a theorist or psychologist.  (Maybe I’m a “runner”;  I create a lot of run-on sentences.)

While I have written programs in nearly every computer language in vogue from 1967 to today: from FORTRAN, Dartmouth Basic and APL, to C/C++, Java and HTML5/CSS/Dojo/JavaScript, as well as specialized languages such as WSDL, VoiceXML/SpeechRecoGrammars and commercial rule-based development environments, lately the software industry has been hesitant to let me be a programmer.

In {Don,Dave,Hans,8plaTom,Vincent}‘s new here? societal function ontology, overlaid with a belief net, {dilettante, engineer, implementer} score highest.

In the chatbot.org profile type ontology, I selected “Mixture” as both Chat Technolog[ist] and Enthusiast are possible choices (at the risk of self aggrandizement.)

The engineer me likes to use science to solve technical problems people around me are having.  I never studied business sciences, and claim that technology is but 10% or less of a successful product so that I can blame all product failures on the business folk messing up.

Early in my career, I managed to work on some expert/knowledge-based reasoning systems, sensor fusion, ANN, and path planning projects (e.g. implemented A* and traveling salesman in Common Lisp on a DARPA Unmanned Air Vehicle contract), and investigated using semantic web analysis of text messages for recognizing plans and operational status of mobile missile divisions in another project.

In my spare time, (after penance paid to Gods, wives, and spawn), I investigated mobile robotics concepts, platforms, and sensors, running a RugWarrior Robot Yahoo! Group to support others using that platform.  http://alanmcdonley.koolhost.com/photos/pogo3.jpg

What am I here (chatbots.org) for? 
  - “Check-in” on the current capabilities of DIY linguistic computer entities and tools.
  - Investigate what alternatives to AIML/ProgramAB/Pandorabots exist
    for creating a conversational front end to a machine learning investigation,
    I would like to build.
  - I am working my way through the great papers linked from chatbots.org

I’ve recently started playing with an overly stripped down AIML rule set to allow dialog on the topic of ML, but as you all probably know, educating a severely stripped down AIML bot to discuss a narrow domain is tedious, and will likely reach a dead-end when I want to interface to external knowledge and reasoning.

(BTW, If anyone knows of an open source “Watson capability” knowledge representation, reasoning, dialog, and unsupervised learning platform that will run on my three computers, please help me join your dream.)

Alan

Hi and welcome, Alan!

That’s an impressive… well, I won’t call it a “resume”, but…

I’m Dave, and I’m the lead developer for Program O, which is a PHP/MySQL-based AIML interpreter/chatbot engine. I’m also one of the admins here, and aside from my work on Program O, one of the “least qualified” individuals here, in the area of AI. In a former life, I was a professional auto mechanic, but now I split time between taking care of my 95 year old Grandmother (who still works out in her garden) and a role as web developer with a gaming consultancy firm. Everything I know about AI and chatbots (including the underlying skills in web development and programming) is all through self education. The only “formal” education I have is an AS in electronics from back in the mid-eighties. That said, however, I’ve got a great deal of passion for the area of Artificial Intelligence, and a love for this community we have here, so it works out well, I think.

Anyway, feel free to poke around here, see what there is to see, and share your thoughts on whatever captures your interest. We’re not as picky about reanimating old posts here as other places seem to be, just as long as the addition is relevant in some way. And of course, if you have any information that you wish to share about any projects you’re working on, please feel free to do so.

Now as to “open source Watson Capability”... Trust me when I say that most of us are already on that particular “quest”. But if any of us DO find something like that, we’ll all know about it, almost immediately.

Hi Alan,

Welcome to chatbots.org. Based on your comments so far we may find ourselves with a great deal to talk about. I’ve done an enormous amount of research into the state of the art in natural language understanding (the real deterministic state of the art, not all the stochastic rubbish that is currently so fashionable) and have developed a large library of code that puts it into practice.

I took the liberty of finding your profile on Google+ and circling you there. There is also a chatbots.org community there but our G+ activities are limited to those rare occasions when real-time communication (instant messaging and/or video conferencing) are more fruitful than posting messages on this forum.

https://plus.google.com/u/0/communities/117236447102577073584

In answer to one of your questions posed above, you may find the following site useful. It is an open source knowledge base expressed as first order logic which I am currently working with.

http://www.ontologyportal.org

Other useful things to search for along these lines are ConceptNet and OpenCyc.

Andrew Smith - May 31, 2014:

Hi Alan,
... you may find the following site useful. It is an open source knowledge base expressed as first order logic ...
http://www.ontologyportal.org

Other useful things to search for along these lines are ConceptNet and OpenCyc.

Phenomenal!  I had thought about tagging knowledge with a source and trust values, but had not thought about the knowledge licensing tags I see in ConceptNet.  I think we will definitely need IPv6 if OpenCyc creates a URL for every word nrelation word1..wordN concept found in Wiki and the greater Inter(Knowledge)net.  Pretty cool stuff.

Thanks - Humans have managed to compact what knowledge is useful into a mobile, off-line database.  Theoretically, a computer entity could save space (at the expense of response time…) by storing the URLs of the most useful external landing spots into the online knowledge bases, from the edges of the entity’s locally (self learned) knowledge base.  This is too exciting - my wife is going to think I’m looking at porn.

Alan

In that case you should probably take a look at YAGO2 as well.

http://www.mpi-inf.mpg.de/departments/databases-and-information-systems/research/yago-naga/yago/

Alan,

I think (this is my personal opinion) if you take a chatbot, and you give it advanced mathematics lessons, and teach it to use a lexical database such as WordNET… You may at least approach being comparatively as smart as IBM Watson, considering it is a homegrown super intelligent chatbot.

http://elizabot.com/entry/ - May 31, 2014:

Judge: What is ontology?
Elizabot: An ordering or classification of knowledge, such as the Dewey decimal system or the design of a knowledge base.  Metaphysics, Ontology would be: The metaphysical study of the nature of being and existence.

Judge: How much is 2348 times 7937?
Elizabot: My calculation is 2348*7937 therefore my answer is 18636076

We are, afterall, comparing an IBM supercomputer to a website running on a webserver. Some supercomputers or mainframes may be water cooled with radiators and water pumps like car engines.  Compare that to the small fan cooling a personal computer.