Friday, February 23, 2018

Dear Blabby: A.I. and Me

If the evaluator cannot reliably tell the machine from... The Turing Test

 ...thanks to the rise of algorithms that can quickly learn tasks on their own, research in conversational computing is advancing.... NY Times
In a natural language computer class I took at the CUNY grad center in the summer of 1988, the project I chose was to create a program to emulate a love advice columnist. I called it "Dear Blabby," using the famous Eliza therapist program, written in the mid-60s, as a model.
The most famous script, DOCTOR, simulated a Rogerian psychotherapist and used rules, dictated in the script, to respond with non-directional questions to user inputs. As such, ELIZA was one of the first chatterbots, but was also regarded as one of the first programs capable of passing the Turing Test. [There is some question about this claim].

ELIZA's creator, Joseph Weizenbaum regarded the program as a method to show the superficiality of communication between man and machine, but was surprised by the number of individuals who attributed human-like feelings to the computer program, including Weizenbaum’s secretary. the human, the machine is said to have passed the test.... While ELIZA was capable of engaging in discourse, ELIZA could not converse with true understanding.[5] However, many early users were convinced of ELIZA’s intelligence and understanding, despite Weizenbaum’s insistence to the contrary.
If you want to know how Eliza did that I'll have to charge you-- though now I could code my way out of a paper bag -- I might have to check with Jeff Kaufman.

Anyway, in that summer of 1988 I did get some kind of conversation going in Dear Blabby, which was trying to converse like your bubba would, but didn't get as far as I hoped, spending long hours in the computer lab using the LISP language, which was a heavy load for the desktops we were using, which made for very slow going. [That summer my sabbatical year followed by a year off without pay was ending, so I may have been under some duress and needed a therapist program myself.]

Every time I see a story about artificial intelligence I want to write a piece on my own experiences with AI 30 years ago when I was going for my MA in computer science at Brooklyn College and CUNY. Many of the courses I chose were AI, which was looking like gold in the mid-late 80s before crashing in the late 90s. There were many sub-fields in AI in the 80s - natural language, artificial vision, machine learning, expert systems and more.

My final courses were pattern recognition (the basis of artificial vision) and one of the earliest classes in neural networks, which is the hottest thing going (I kept my textbooks but lost them in Super Sandy 5 years ago. If only I didn't go back to teaching in 1987 and stuck with it I may have been a contender.

Today's business section of the NY Times has an interesting piece,  To Give A.I. the Gift of Gab, Silicon Valley Needs to Offend You
which talks about chatbots - the hot rage on all your help lines.
These systems do not simply repeat what is said to them or respond with canned answers. They teach themselves to carry on a conversation by carefully analyzing reams of real human dialogue. At Microsoft, for instance, a new system learns to chat by analyzing thousands of online discussions pulled from services like Twitter and Reddit. When you send this bot a message, it chooses a response after generating dozens of possibilities and ranking each according to how well it mirrors those human conversations.
Alan Turing defined the border being breached between humans and computers in
The Turing Test - wikipedia
The Turing test, developed by Alan Turing in 1950, is a test of a machine's ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human. Turing proposed that a human evaluator would judge natural language conversations between a human and a machine designed to generate human-like responses.
Let me know when you can't tell the computer from a human (Where are you Hal?). It may not be long in coming. And then there are the robots coming which will replace us all as we head for Mars with Elon -- but that one another time.

And how about Arnold and his pals in the first Terminator movie where the aim is to wipe out the human race. Thank you very much. I think there's a good chance we won't need robots to do that.

No comments: