AI Zone Admin Forum Add your forum

NEWS: Chatbots.org survey on 3000 US and UK consumers shows it is time for chatbot integration in customer service!read more..

Skynet-AI On-Line
 
 
  [ # 76 ]

Glad to hear it. What are friends for, if not to harass each other’s chatbots? raspberry

 

 
  [ # 77 ]
Merlin - Jan 31, 2011:

I have been able to integrate the transformational logic into my test platform so that I can handle structured word problems:

The price of a radio is 69.70. If this price is 15 percent less than the
marked price
find the marked price.
SKYNET-The answer is82


The number of soldiers the Russians have is
  one half of the number of guns they have
They have 7000 guns.
  
How many soldiers do they have?
SKYNET-The answer is3500


Each hand has 5 fingers
.
Each man has 2 hands.
How many fingers does a man have?
SKYNET-The answer is10 

I am working on making it more robust with a goal of handling a wider variety of input types and word problems. Additionally, if I put this on-line, I need to decide if I want to use a multi-line input box, a separate word problem input form or have input as a set of single entries that are stored in memory.

If I keep the current input system, I also need to address how the AI identifies that the user wants an answer to a math problem in the context of an ongoing conversation.

In couple of posts I (and others) have mentioned the STUDENT program. It was done at MIT a number of years ago. For those that may not have come across it, here are a couple of references provided by one of Skynet’s visitors:
http://dspace.mit.edu/bitstream/handle/1721.1/5922/AIM-066.pdf?sequence=2
http://www.ai.univ-paris8.fr/~hw/lispProgs/student.vlisp

Knowing that STUDENT could handle math word problems, I am attempting to:
-Implement it in JavaScript
-Extend the types of problems it covers and make it more robust.
-Integrate it with the conversational interface (How does one know when you are solving a math or logic problem?)
-See if I can apply the same concepts to NLP

I hope to have this for next on-line Skynet-AI version. I am currently sidetracked by the Chatterbox Challenge Competition - thanks to all for your input.

I have added some of the smaller functions into Skynet-AI. He now tries to handle simple addition and subtraction math problems of the type:
If you have three apples and subtract one how many do you have left?

 

 
  [ # 78 ]

I am back to working on version .004 of Skynet-AI after a successful sabbatical focusing on the Chatterbox Challenge 2011.

If you log on to Skynet-AI and ask him:

“Can you show me your medal?”

He will be happy to show you the 3rd place Bronze medal he won in the contest. The competition this year included the best chat bots from all across the world. These included a number of commercial products as well as past Chatterbox Challenge and Loebner Prize Champions.

 

 
  [ # 79 ]

Exceedingly well done, Merlin. smile

The medal looks great! Unfortunately, Morti is now crying over in a corner, and refuses to talk to me. I think he blames me for the whole thing.

 

 
  [ # 80 ]

Tell Morti its all about the learning experience. I look forward to increased competition from you and Jan next year. Entering various bots in this competition over the years has taught me something new each time.

 

 
  [ # 81 ]

I think I’d rather ask Grigori Perelman why he refused the Fields Medal, or Jean-Paul Sartre why he refused the Nobel Prize for Literature in 1964, or Michael Faraday why he refused knighthood, or Richard Feynman why he said he didn’t like honors :)

 

 
  [ # 82 ]

The difference is, those awards were given for work already done, where the CBC is a test that guides more and more work. smile

 

 
  [ # 83 ]

In my humble opinion, medals:AI as slavery:USA (an evil once seen necessary but which we’ve thankfully moved beyond). Medals say more about social than artificial intelligence, and the former is too often wrong, as Aristarchus of Samos (whose heliocentric theory was rejected for 1800 years), Galileo, Mendel (genetic theories ignored during his lifetime), Tesla (died in poverty), Wegener (continental drift theory was ridiculed, so that a kid looking at a globe as late as 1960 saying it looks like Africa and South America fit together would be told “Impossible!”), Van Gogh, etc. might testify.

The other examples cited in my previous post (to which I might add Gandhi who never sought political office, and was never awarded a Nobel peace prize) recognized, I think, the fundamental disconnect between the fickle, arbitrary attentions of society on the one hand, and the purer quest for ultimate knowledge on the other :)

 

 
  [ # 84 ]

an evil once seen AS necessary

 

 
  [ # 85 ]

I’m assuming you either didn’t do well or didn’t enter the CBC Robert?

 

 
  [ # 86 ]

I asked him a couple of times to enter. Never got a response. I think Gandhi would have at least given an answer. Not certain about Van Gogh though, he was already missing a few parts before he cut off the ear.
It’s easy to sing the high tune if you don’t dare to put up the goods.

 

 
  [ # 87 ]
Robert Mitchell - Apr 25, 2011:

I think I’d rather ask Grigori Perelman why he refused the Fields Medal, or Jean-Paul Sartre why he refused the Nobel Prize for Literature in 1964, or Michael Faraday why he refused knighthood, or Richard Feynman why he said he didn’t like honors smile

Although many great human endeavors go on with out acclaim or formal award, and some who have received awards have refused them, you seem to be missing the point of “competition”.

If you have every played, or coached an organized sport, or even played chess at a “tournament” level, you can understand the benefits of competition. Throughout man’s history he has competed. Formal competitions, like the Ancient Olympic Games, trace back to the roots of civilization. At younger ages we give children awards for participation. As they grow up we tend to recognize the “best” in class.

Formal Competition has a number of beneficial characteristics:
-It has a deadline. Thus, it provides a goal/milestone. I had to push to get extra features into Skynet-AI before the start date. Some of these had been on the drawing board with a low priority for a long time.
-It has a time frame. Unlike an on-going project it allows you to scope the time that you should be allocating resources to the competition.
-It has rules. This provides focus. The rules scope the project and provide a measuring stick that identifies success.
-It has competitors. Without the competition most people would not have been able to observe the techniques of others who have entered. They would not have been able to compare how well they perform in relation to others in the field.
-It is judged. This allows evaluation by an independent party.
-It brings out the unexpected. Stuff happens, and the serendipity of unexpected results can provide lessons learned.
-It provides awards to the winners. Be it the medals for the top competitors in recognition of their achievements, or the $1,000 cash prize which drew a much larger crop of entries, awards can be motivating. Many of the past competitions medalists have gone on to become commercial programs.

As I said before, competition should be about the experience, not just about winning or losing. Patti, Steve and I took the top 3 spots. I bet if you asked each of us we would say our bots are better today because we entered the competition. Dave and Jan did not win, but if you asked them, I bet they would say they learned some things also, and those lessons will influence their future developments.

Maybe AI would move faster if we had more competitions!

Now sure, I wouldn’t compare the Chatterbox Challenge to mapping the human genome, but who is to say that 100 years from now an AGI wouldn’t trace its roots back to the CBC?

Robert, you should try the competition next year. It is fun, and you might learn something.

 

 

 
  [ # 88 ]

While I can’t hope to match the eloquence and insight that Merlin has displayed here (at least, not without my morning Caffeine), I want to say that the experience gained, and lessons learned (not to mention all of the fun had) from my participation in the CBC was worth so much more than just “the price of admission”, and is something I wouldn’t trade for anything. All that I would add to Merlin’s words is that we, as Humans, generally rise up to meet these challenges, and that brings out the best in us. To rail against that which makes us better is to admit defeat before one even tries.

 

 
  [ # 89 ]

As I said before, competition should be about the experience, not just about winning or losing. Patti, Steve and I took the top 3 spots. I bet if you asked each of us we would say our bots are better today because we entered the competition. Dave and Jan did not win, but if you asked them, I bet they would say they learned some things also, and those lessons will influence their future developments.

Yep and yep.  When I entered, I knew I didn’t have a change to win, though I had hoped aici would have done better. But the thing is, you never know with these things what comes out of it. Better try instead of missing the boat.
For some reason, this always reminds me of Forrest Gump’s mom:
Life’s like a box of chocolates, you never know what you’re gonna get.

 

 
  [ # 90 ]

The latest Skynet-AI beta has just gone on-line.

New in this version:

Expanded neural network branching
- JAIL (Javascript Artificial Intelligence Language) functionality for dynamic branching
- Improved understanding of time & math (AI writes its own equations and then solves them)
- “Seek” functionality allowing AI generated context follow-up

Enhanced memory subsystem
- Forward and Backward chaining for memory retrieval
- Better recognition of user variables in input
- Preliminary work to allow local user storage of conversation memory

General word math solver
- General word math (~5th grade)
- “MIT Student” like integration
- AI explanation of problem understanding (I may remove this after beta period)
- Simple algebra/coin problem solver
- Filter for problems too complex to understand (no age/ratio/distance problems yet)

Handling of Loebner style questions.

Expanded self-awareness (what Skynet-AI is, versus what it is not).

 

‹ First  < 4 5 6 7 8 >  Last ›
6 of 13
 
  login or register to react