Dear Behrouz,
Thank you for your kind words about GeNIe!
I just wanted to correct a couple of things. While BayesFusion has
decided to continue offering GeNIe free for academic research and
teaching use, it is a commercial product nowadays. The link to
commercial downloads is:
https://download.bayesfusion.com/files.html?category=Business
I suggest, however, the general address:
https://www.bayesfusion.com/
and link to Downloads. Commercial version runs in a trial mode for 30 days.
GeNIe runs native on Windows but also on Mac and Linux under Wine.
SMILE is written in C++, is fully portable and runs on a large variety
of platforms. SMILE wrappers are available for Java, .NET and .COM.
Through these, SMILE can be used from R, Python, Excel, Matlab, Ruby, etc.
GeNIe's implementation of influence diagrams includes multiple decision
nodes, multiple utility nodes, multi-attribute utility functions
(additive-linear and general). There are two versions of the inference
algorithm for influence diagrams: (1) an incredibly fast one that picks
the best policy, suitable for autonomous applications like robots, and
(2) a very fast one :-) policy evaluation that calculates the expected
utilities for every combination of decision options and every
combination of states of nodes that should be observed before making a
decision. The latter algorithm is suitable for interactive policy analysis.
Finally, GeNIe is hybrid these days (since last summer), which means
that you can freely mix discrete and continuous variables in the same
model. We will be supporting dynamic influence diagrams and their
relatives soon.
I hope this helps.
Marek
----------------------------------------------------------------------
Marek J. Druzdzel http://www.pitt.edu/~druzdzel
On 1/28/2018 12:38 PM, Robert Goldman wrote:
Thanks for the advice. I should have mentioned that I'm at a business
(although it's a research business, a funded research lab), so I'm not
eligible for free academic licenses. I realize that this narrows the
field a bit.
Note that I'd be willing to use something that required some
programming. For now I've found a python library, but it's quite fussy
(easy to get the input wrong, limited in the queries it will answer), so
I would like to do better.
An Octave or R library would be great, for example, although something
with a GUI would be better.
Bonus points if it's not limited to discrete RVs.
Best,
r
On 27 Jan 2018, at 16:50, behrouz wrote:
Hi
Last year I surveyed a number of options and I ended up using GeNIe
and SMILE. They are freely available for the academic user.
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.bayesfusion.com%2Facademic-users&data=01%7C01%7Cmarek%40sis.pitt.edu%7C520e171131424c9d850b08d566a5fdfe%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=5B9vG%2Fi5Nzt4zYrvFmB88fsarnODFL%2FSPVZpGp%2BCCvY%3D&reserved=0
The user interface (which was really nice) is only available for
windows. I think this GUI is the fastest means for 'playing' with
examples.
The SMILE library however supports Linux.
Best,
Behrouz Babaki
On 01/27/2018 12:43 AM, Robert Goldman wrote:
Can anyone point me at some free code that will let me play around
with some simple influence diagrams?
There seem to be a bunch of really great, capable, commercial
systems, but for now, I just need something that will let me do some
meatball examples.
If I can do some convincing examples, I might eventually want a
capable library, but right now, I just need to do some minimal examples.
thanks,
r
_______________________________________________
uai mailing list
uai@ENGR.ORST.EDU
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsecure.engr.oregonstate.edu%2Fmailman%2Flistinfo%2Fuai&data=01%7C01%7Cmarek%40sis.pitt.edu%7C520e171131424c9d850b08d566a5fdfe%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=DjfgzFoHR67be49K1lB6vWbNyGXDqh8cdP%2BTT4Kjjhs%3D&reserved=0
_______________________________________________
uai mailing list
uai@ENGR.ORST.EDU
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsecure.engr.oregonstate.edu%2Fmailman%2Flistinfo%2Fuai&data=01%7C01%7Cmarek%40sis.pitt.edu%7C520e171131424c9d850b08d566a5fdfe%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=DjfgzFoHR67be49K1lB6vWbNyGXDqh8cdP%2BTT4Kjjhs%3D&reserved=0
_______________________________________________
uai mailing list
uai@ENGR.ORST.EDU
https://secure.engr.oregonstate.edu/mailman/listinfo/uai
