What if Johnny 5 was reading WWW, not only books?

A chinese lion statue

Johnny5 was a military robot from Short Circuit. One day it gets struck by lightning and becomes... conscious? self-aware? a being with emotions? (hard to say, right?) My idea is not to put robots outdoors during stormy weather but rather to give it basic needs (e.g. to serve us) and to program it to collect (by using NLP) as much background knowledge as possible from the biggest source that exist - the Web. My main topics are Common Sense and Emotions, the Yin and Yang of our intelligence, and everything what happens between them - for example Imagination, Humor or Ethics. If consciousness is needed - I will give it average WWW personality and you should be tricked :) But seriously - it's time to do something fresh. I decided not to dispute if logic or knowledge base approaches are better. I decided to bet on the minimalistic cognitive architecture with a simple instinct - like a bacterium - living its life in the virtual world of text from which it can learn about us and our world. AI people looked into our brains for too long now. We need to concentrate on the world our brains live in. Without all other brains around - it wouldn't be the same brain anyway.

↓ Below you will find some news written in rather bloggish style.
↑ Above you have links to some other info. You can also subscribe to my feed.

Machine Ethics

Posted on May 5, 2009

After more than three years, my ideas on combining common sense retrieval and emotions processing to achieve a moral reasoning algorithm started to get some code. I want to present its results during JSAI 2009 if only the University let me go - I am back from TSC 2009 on June 15 and plan to fly to Shikoku the very next day, however my office says I have to stay home for 7 days after coming back from an infected country... Because none of my students is interested (yet) in Machine Ethics, I decided to invite a young ethicist from outside Japan. He is more in children's moral rules acquisition (Kohlberg's Development) and that's what my approach is missing. Stay tuned for more!

Roger Schank - Network of Wise Men

Posted on February 20, 2009
Roger Schank

One of the main reasons I went to IUI 2009 was to listen to this man. Somebody whose theories gave me plenty of ideas, gave an invited talk for Common Sense community during Story Understanding Workshop. He's right - we need network of specialists knowledge we don't possess. Yes, we surely don't need any knowledge from blogs about somebody switching off the lights and going to bed. We, humans, don't, but machines - they surely do. Plans, Goals, Scripts - these are A. I. basics but I believe many of them could have a chance for second youth in data explosion era. If you are not familiar with Prof. Schank's works and you don't have time for reading his books - Calvin Ashmore gives you some critical look into the most famous one.

Rare topics

Posted on December 20, 2008

During COLING 2008 and EMNLP 2008 our lab members presented their work on artificial humor and a chatbot using associations from the Web. Some researchers ask me - is this a research on jokes or joke about research? Some time ago a reviewer asked one of my student - why do you call it a dialog agent? It's just a chatbot!

Is it really so hard to imagine that joking or chatting needs more than a dose of intelligence? Maybe such topics are not profitable right now, but it is easy to imagine that many people would rather buy talkative funny housekeeping robot than a cold-hearted silent one :)

Get involved

Rafal Rzepka

I often hear that my average intelligence will result in boring robot which is a Jack of all trades but master of none. Well, that's maybe true, but aren't most of us so? Maybe we are far from the super-intelligence because we want to skip low-level daily life intelligence?
If you are a student interested in interdisciplinary approach to A. I. and you wish to come to Japan for couple of years - check the scholarship possibilities and contact me. If you have some new ideas and love to challenge topics that others are afraid of - you can count on my help.

Araki Lab Links