I think the term “alicebot clone” may be a demotion for The A.L.I.C.E. AI Foundation.
It seems the term does more to demote the adoption of the A.L.I.C.E. and AIML.

Yet one could argue the term follows in the tradition of ELIZA written at MIT by Joseph Weizenbaum.

Why do chatbot masters and A.I. designers disrepute their own branch of knowledge?

8PLA • NET - Mar 27, 2011:

[...]Why do chatbot masters and A.I. designers disrepute their own branch of knowledge?

Tell me more about “disrepute”.

I think, in this particular context, his intention more closely means “denigrate”. Of course, I’m not entirely certain, but that’s how it seems.

I believe AIML is a great tool for developing a chatbot and never understood why some people (usually those who don’t use AIML) class it as a second rate method.

As for the clone part, the ALICE database is an excellent way to get started and I see no point in amending those replies that work such as, “How are you - I am fine thanks”. I don’t see how anyone who works extensively to modify and enhance the ALICE database for their own chatbot can be accused of cloning?

These bots are as much a clone of ALICE as the newest motor car is a clone of the Model T Ford. They both have 4 wheels and an engine. The ALICE database is just a place to start, that’s all.

Also, while the term “ALICE clone” may be a disparaging term, I don’t feel that it’s disparaging toward ALICE, herself (itself? maybe). Nor should it be. I see the ALICE chatbot as a milestone in AI, and to be compared to ALICE isn’t saying something bad about ALICE (or, at least, it shouldn’t be). While it’s true that the field of AI has advanced past ALICE and other AIML chatbots, that doesn’t discount their value today as a good beginning platform for new botmasters. My only complaint about the whole thing is that these “new” botmasters don’t always understand which of the tens of thousands of response categories should have the highest priority for alteration. I’ve been working on Morti for nearly a year now, and STILL have some categories that I have missed (as was evident from the CBC questions). Truthfully, unless a botmaster makes no effort to make their bot unique, a little slack is probably called for, since it’s a HUGE project to make enough changes to the AIML files to make an AIML chatbot seem unique. At least, that’s how I see it.

Yes Dave,

That is correct, and an excellent choice of synonyms

———

Arthur,

To tell you more about : “disrepute”. 
Another word would be: “discredit”.

———

Used in a sentence:

After building ELIZA, Joseph Weizenbaum disreputed A.I. by warning it could take away jobs from humans.

Steve,

You’re right.

It’s like saying AIML is an XML clone.

The Pandorabots implementation is one of the easiest ways for a newcomer to start with a chatbot. Steve’s Mitsuku has also proved that AIML can be used to create an excellent bot. People get confused though between a number of things:

- The Language
AIML is based on XML, I believe the latest spec is:
Artificial Intelligence Markup Language (AIML)
Version 1.0.1 http://www.alicebot.org/TR/2001/WD-aiml/
A.L.I.C.E. AI Foundation Working Draft
25 October 2001 (rev 006)

Although this is listed as a working draft I believe this is final version of the spec. If someone know different please correct me.

- Language Interpreters
http://www.alicebot.org/downloads/programs.html
The Pandorabots interpreter is implemented in common LISP.
AIML interpreters have been implemented in a variety of programming languages.
AIML interpreters look for an exact match on an input pattern to identify an appropriate response. The interpreter is able to handle a large set of input through the use of recursion and reductionism.
http://www.alicebot.org/documentation/srai.html

- The ALICE data set
http://www.alicebot.org/aiml/aaa/
The AAA data set provides a large foundation of input/response pairs (know as categories in AIML)
AIML data sets are normally used with an AIML interpreter. But, this is not always the case. Because the data set is XML it can be converted relatively easily to other forms and used in other interpreters.
In 2009 Casey Kirsle ported the Alice AIML set to RiveScript.
http://www.rivescript.com/?p=blog&id=8

Depending on the goals of your bot there may be good reasons not to use an AIML interpreter. I explained some of mine in a discussion with Dr. Wallace on the “JAIL” thread.

Dave explained some of his on the “Translation Project: AIML to ChatScript” thread.

Dave Morton - Mar 24, 2011:

Hi, Oliver, and welcome to chatbots.org.

The biggest reason I’m doing the translation from AIML to ChatScript is because ChatScript has the potential to be much more flexible and efficient than AIML is. With AIML, any wildcard in your pattern has to match at least one word in order for the pattern to fire. So the pattern “I love you *” won’t match “I love you”. This means that with AIML, you need more categories to accomplish the same desired responses.

Also, ChatScript has a special type of wildcard, called a concept, that adds a whole new dimension to matching. Let’s say you want to create some responses in AIML that covers certain hobbies. You need, then, to either create a category for each hobby you want to deal with, or create a category that would accept inputs that may not describe a hobby:

<category>
<pattern>I like *</pattern>
<template><star /> sounds like a good hobby.</template>
</category> 

The same response in ChatScript would look like the following:

concept: ~hobbies (reading “stamp collecting” woodcraft “chasing girls”)
s: (I like _~hobbies) _0 sounds like a good hobby.

As you can see, not only is the rule more flexible and efficient, it’s also easier to read. There are other reasons for my migrating to ChatScript, but this is the biggest reason. Bruce Wilcox has a post in these forums that describes in greater detail the other advantages of ChatScript. I recommend it highly.

Dr. Wallace and I discussed some of the pros and cons on the JAIL thread.
http://www.chatbots.org/ai_zone/viewthread/317/

8PLA • NET, could you share with us some of the reasons you are running the RiveScript version of the Alice data set instead of using one of the more traditional interpreters?

AIML (the language and the interpreters) along with the AAA data set are great resources for developers interested in chatbots and may be the most popular way to currently put a chatbot on the net.

I do agree with the wildcard being a little annoying. As well as * and _, I would love another symbol (maybe ~) to indicate some or no words. eg

I am happy ~

would match “I am happy” or “I am happy to see you”.

Steve, how many categories are you up to?
I think Dave said he was at around 40,000 now.

I have around 6,000 in Skynet-AI. One of the other benefits I have is the “OR” tag to allow multiple patterns (or pieces in a single pattern) to match to a response.

Merlin - Mar 28, 2011:

8PLA • NET, could you share with us some of the reasons you are running the RiveScript version of the Alice data set instead of using one of the more traditional interpreters?

Yes Merlin,

Thank you for your question. In my personal opinion, my main reason (which is listed in my profile) is that I enjoy making chatbot history as the first to do something.  It was awesome when Dr. Wallace confirmed my official facebook app http://apps.facebook.com/eight-pla-dot-net was the first AIML chatbot on facebook.  So when Dr. Wallace, during one of his excellent, interesting speeches, makes a reference to how AIML bots are now joining social networks, he is referring to my efforts to promote AIML.

So I saw a similar opportunity with RiveScript to be the first at something in chatbot history. It was very exciting when the creator of RiveScript, Casey Kirsle, confirmed that 79b.org is the first RiveScript chatbot to enter the Chatterbox Challenge.  And I hope to be the first RiveScript chatbot to enter the Loebner Prize Competition in the UK.

Merlin - Mar 28, 2011:

Steve, how many categories are you up to?
I think Dave said he was at around 40,000 now.

I have ten update files, each containing between 7,000 and 9,000 categories as well as other AIML files (http://www.square-bear.co.uk/aiml) I would say I was bordering on 90,000+ custom categories written over a period of 8 years or so. This is in addition to the original ALICE files of about 20,000 categories.

Wow, lots of hard work. No wonder Mitsuku performs so well.

I realized I miscounted mine (I included the responses in the count), I am only at about 3,000, but it is hard to do a direct comparison because the “OR” tag allows me to map multiple inputs to a single output.

Do you use any special tool to manage/update the files?

Merlin - Mar 29, 2011:

I realized I miscounted mine (I included the responses in the count), I am only at about 3,000, but it is hard to do a direct comparison because the “OR” tag allows me to map multiple inputs to a single output.

That’s not so different from the categories that <SRAI> to other categories, Merlin.

I also found out yesterday that I still have a small percentage of categories that are “duplicates” of others, so when I get back, I’m going to have to write a script that searches out these duplicates and displays them for review, so that I can delete the extra ones. I expect that this may cause the number of categories to shrink a bit from the current total of 46,372, and may make the total drop below 46 thousand.

Dave Morton - Mar 29, 2011:

Merlin - Mar 29, 2011:

I realized I miscounted mine (I included the responses in the count), I am only at about 3,000, but it is hard to do a direct comparison because the “OR” tag allows me to map multiple inputs to a single output.

That’s not so different from the categories that <SRAI> to other categories, Merlin.

True, but all those different SRAI categories that map to a single response are in just one category on my system.

Dave Morton - Mar 29, 2011:

I also found out yesterday that I still have a small percentage of categories that are “duplicates” of others, so when I get back, I’m going to have to write a script that searches out these duplicates and displays them for review, so that I can delete the extra ones. I expect that this may cause the number of categories to shrink a bit from the current total of 46,372, and may make the total drop below 46 thousand.

I may have a tool that can help you with that. Let me know when you get back and we can talk about it.

You may want to bear in mind that, while I have nearly a dozen AIML “files”, they’re all stored in a MySQL database in a single table. I have a script that can export that table into the various files, but I usually just work on the AIML categories directly within the DB.