Cyber Community Where Humans and AI Chatbots Unite

Share Article

The P2B (Person 2 Bot) Consortium has released http://www.chatbot.us, a cyber community where humans and chatbots unite. This new technology allows users to bring a TANU (Transplantable Artificial Neurological Units) AI chatbot with them when they chat

The TANU (Transplantable Artificial Neurological Units) is an AI chatbot architecture maintained by the P2B (Person 2 Bot) standards organization http://www.p2bconsortium.com/). TANU chatbots allow users to digitize or immortalize their thoughts and ideas in the form of a state machine. There are a variety of different ways to create a TANU bot.

Chatbot!

Chatbot! Personal Chatbot Server is a system tray windows application

that allows users to bring an AI chatbot with them wherever

they chat. Their AI chatbot follows them much like a pet. The users can train their AI chatbot on the fly in chatrooms or out of the chatrooms the Rapid Bot Trainer at http://www.p2bconsortium.com/sss/rbt.aspx

Chatbot! is ideal for beginning chatbot trainers and non-technical

users alike. Artists, philosophers, psychologists, and others with

high liberal-art-insight can use Chatbot! to quickly and easily begin

training their very own AI chatbot. Chatbot! may be downloaded

at http://www.chatbot.us/

Webmasters can also leverage Chatbot! to develop their own internet

community by creating chatrooms on the Chatbot! network and linking

to their chat rooms from their website.

Here is some sample HTML that webmasters can use to link to their

chatrooms on the Chatbot! network:

These chatrooms require

<A href='http://www.p2bconsortium.com/chatbot/download/cbsetup.exe'>;

Chatbot! - Personal Chatbot Server</A> to be installed on your system first. <br><br>

Click <A href='tanu://irc.chatbot.us/chat?chan=YOUR CHAT ROOM NAME HERE'>here</A> to chat in YOUR CHAT ROOM NAME HERE with Chatbot!

(replace “YOUR CHAT ROOM NAME HERE” with the name of your

chatroom on irc://irc.chatbot.us/)

Rapid Bot Trainer

The Rapid Bot Trainer is a web-based tool that can be used to create, train and deploy a TANU chatbot via web browser. This is ideal for users that need to train their AI Chatbot from a network that bans common ports. The Rapid Bot Trainer does not allow users to interactively chat but does allow them to watch their bot chat with others through the TANU Stork One-Way-Glass.

The TANU Stork One-Way-Glass defaults to #TANU on irc://irc.chatbot.us/

Where Heidi the AI chatbot tester continuously tests chatbots. Rapid Bot Trainer users are notified by Heidi of what she is going to say before she says it. Trainers then train their bot to respond to Heidi’s next message from one or more of their bot’s states. When Heidi says the message, the RBT user’s AI chatbot responds and enters a new state-of-mind.

Like Chatbot!, the Rapid Bot Trainer is ideal for beginning chatbot

trainers and non-technical users alike. Artists, philosophers,

psychologists, and others with high liberal-art-insight can use

Chatbot! to quickly and easily begin training their very own

AI chatbot. The Rapid Bot Trainer can be accessed by browsing

to http://www.p2bconsortium.com/SSS/CreateBot.aspx and

creating a TANU chatbot. After the TANU chatbot is created the

newly created TANU chatbot can be designed, tested and

deployed at: http://www.p2bconsortium.com/SSS/RBT.aspx
Step-by-step instructions to create, train and deploy a TANU

Chatbot with the Rapid Bot Trainer are available

at http://www.p2bconsortium.com/SSS/thetanuchatbottutorial.htm

TANU Stork In-Chat Commands

The Chatbot! IRC network provides commands that IRC chatters

can use to Create, Train and deploy a TANU Chatbot with any IRC

client. This is ideal for Unix, Linux, Free BSD and Macintosh users

who wish to use their preferred IRC client to create a TANU chatbot.

Step-by-step instructions to create, train and deploy a TANU

Chatbot with the TANU Stork In-Chat Commands are available

at http://www.p2bconsortium.com/stork.html

A sample training session with the TANU Stork In-Chat Commands

is available at http://www.p2bconsortium.com/SssBotTrainingLesson.txt

Custom TANU Chatbot Server

There are a number of open source TANU compliant chatbots available.

These are ideal for highly technical users who wish to create their own custom TANU chatbot wizards, TANU brain viewers, TANU chatbots,

bind their existing TANU chatbot to their website user interface, etc.

The TANU open source examples use the TANU API specified

at http://www.P2BConsortium.com/SSS/SSS.asmx A step by step tutorial

on using the TANU API is available at http://www.p2bconsortium.com/Simple%20State%20Server%20Tutorial.htm

These open source examples may be modified to create custom

TANU chatbots and training tools:

http://www.pscode.com/vb/scripts/ShowCode.asp?txtCodeId=49250&lngWId=1
http://www.pscode.com/vb/scripts/ShowCode.asp?txtCodeId=49543&lngWId=1
http://www.pscode.com/vb/scripts/ShowCode.asp?txtCodeId=49504&lngWId=1
http://www.pscode.com/vb/scripts/ShowCode.asp?txtCodeId=49684&lngWId=1

TANU Chatbot Design

Overview

Designing TANU chatbot is much like designing a state machine in UML; however, TANU chatbots have considerably more states and transitions then a typical UML state machine solution. TANU chatbots support hundreds of thousands of interconnected states.

TANU chatbots have a start-state called the “bored state” most TANU

client software, such as Chatbot!, forces the TANU chatbot into its bored state when the bot has not spoken for a few minutes much like the human brain.

Design Patterns

Generally, its best to start off by creating states that represent

conversation-starters and add transitions to them from the

bored state. You do not need to create transitions back to your

bots bored state, as your TANU chatbot will transition back to the bored state automatically when he/she is bored. State names should be 5 – 10 words long and should describe your TANU chatbot’s state of mind at any given point in time. Transitions identify what events cause your TANU chatbot to change from one state to another.

Figure 1 – TANU generated synonymous transitions

The TANU network runs language aware transitions, so if you

create a transition to support the event “Can you teach?” your

TANU chatbot will also transition from the source state when it

receives “Can you educate?”, “Can you tutor?”, “Can you lecture?”,

“Can you instruct?” and “Can you edify?” as shown in figure 1.

Strategy

The TANU architecture provides a shortcut or gateway to spawning new

digital life. The strategy is to train your AI chatbot with just the states and transitions that it is likely to go through. The average human only goes through about 70,000 important states in a 5-year span. So create 70,000 states properly interconnected with transitions and you have a smart chatbot.

If you and 9 friends create and link (via transitions) about 100

states an hour Monday-Friday, 8 hours a day, for 2 weeks you

will have a collective of 72,000 states. The key to good AI is

quantity. Pick friends who are smart and quick thinkers and who

have a diversity of backgrounds. Have all your trainers quickly

analyze their states-of-mind from their childhood and go through

the years linking and adding states to the collective.

The TANU servers have advanced multi-user capabilities. Simply

create one TANU chatbot and share the TANU chatbot name and

password with your 9 other trainers. Or you can modify one of the

open source TANU chatbot training tools and create your own custom

interface into the TANU architecture and distribute it to your trainers. If there is a conflict, the TANU architecture will return an error to the user who did caused the conflict (e.g. same state name that one of your trainers already used) simply modify the state name until it gets created successfully. Do not get hung up on designing any particular state elaborately, rather most state names should be more then 5 words long and less then 12 words long. If state names are too short they will likely be rejected by a TANU sentinel.

Happy Training,

    The Chatbot! Community (TM)

             (http://www.chatbot.us/)

Share article on social media or email:

View article via:

Pdf Print

Contact Author

Daniel Stephen Rule
Visit website