teaching tools

40 views
Skip to first unread message

BP

unread,
May 5, 2009, 5:15:59 PM5/5/09
to Python Artificial Intelligence SIG
I will be teaching an artificial intelligence course in the fall.
While we use PROLOG, I'd also like to show what a language like PYTHON
can do with AI. Does anyone know of a PYTHON implimentation of SHRDLU
or an expert system shell?

Petro Verkhogliad

unread,
May 5, 2009, 5:20:51 PM5/5/09
to ai...@googlegroups.com
http://pyke.sourceforge.net/

That looks good. I can't say how good it is, since I haven't used it.

Petro

BP

unread,
May 5, 2009, 5:55:08 PM5/5/09
to Python Artificial Intelligence SIG
It does look interesting. Thanks. Maybe I'll post some comments on it.

On May 5, 5:20 pm, Petro Verkhogliad <vpe...@gmail.com> wrote:
> http://pyke.sourceforge.net/
>
> That looks good. I can't say how good it is, since I haven't used it.
>
> Petro
>

Ralph Miller

unread,
May 6, 2009, 6:22:10 AM5/6/09
to ai...@googlegroups.com


On the dfwpython.org site in the svn repo is PyIE. It is released under
the MIT license and is quite full featured. I am trying to make time for
a formal notification of availability.

https://www.dfwpython.org/repo/Projects/PyIE

From the Users Guide:

PyIE - Python Inference Engine


PyIE is a hypothesis based, agenda driven, object oriented inference
engine written in Python. Inferencing modes include back chaining,
opportunistic forward chaining and explicit forward chaining. The
first-class object base supports metaclasses, classes, objects and
multiple, dynamic inheritance. All objects are first class objects and
all attributes are slot values, i.e. data members attached to some object.


PyIE uses a TMS (truth maintenance system) for first level (condition
change) non-monotonic reasoning. User defined metaclasses are being
explored to support second level (belief change) non-monotonic reasoning.


The agenda contains only hypothesis that are placed onto seven priority
levels according to the reason for insertion.


Agenda priority levels:

0. Bounded response and truth maintenance system
1. Back chaining
2. Value changes
3. Forward chaining (Opportunistic)
4. Forward chaining Explicit (RHS)
5. Prompt Request (into)
6. Request slot value

Features include:

∑ Command line interface
∑ GUI runtime interface
∑ Remote invocation using PyRO ( including GUI runtime interface )
(currently disabled)
∑ Embeddable
∑ Multiple Inheritance in object base
∑ Dynamic Inheritance in object base
∑ Linking to / from classes / objects for object behavior acquisition
∑ System methods ( in Python )
∑ Class methods ( in Python )

If you do not have svn, access the .zips and .tar.gz files in the
download folder.

I will help if you need any assistance, it is an on-going, personal,
project I have been working on in my spare time.

Hope it helps.

Have fun,
Ralph

Reply all
Reply to author
Forward
0 new messages