BayesFusion Support Forum

The patched version solved the problem in the minimal example above, but it still causes a "segmentation fault" on my main project when calling UpdateBeliefs(). With the older version, it still works on my main project.
I'm sorry that I don't have time to make a new minimal example in the near ...

Thank you :)

I found an older version of the library in an older project I had.
With this version, it worked without any errors.

Version:
SMILE ACADEMIC version 1.6.0 / 2020-09-28
The rest is identical as in the original post

Hi, I'm experiencing a weird bug in the smile library for c++ in the following case:
I have a network consisting of at least one temporal node
"simple.xdsl"
<?xml version="1.0" encoding="UTF-8"?>
<!-- This network was created in GeNIe Academic, which can be used for academic teaching and research ...

Great!

The problem seems to be solved!

Nice, thanks :)

When trying to copy an equation node from one file to another, I get the following error:
paste_error.png

When I then save the file, and try to re-open it the file is corrupted:
corrupted_file.png

The error is related to this new entry in the file:
code_error.png

There is no entry for the ...

Chance node are 0-indexed:
chance_node_states.png

While equation nodes are 1-indexed:
equation_node_states.png

So if you make a chance node and then turn it into an equation node you get this weird numbering (note that State1 is followed by State3):
chance_into_equation.png


Would be nice ...

The feature I was looking for was to discretize the network once, and then use the discretized network in an online application interfacing the SMILE library. But it worked to simply change all nodes to chance nodes :)

Aha! Is it correct that in the new version you can simply change nodes from equation to chance and it will be correctly discretized?

Hi, I can't seem to find the button for making a discretized version of the equation based/hybrid model in Genie 4.0?
I'm thinking about the button that opens a new file where all nodes are discretized.

Best regards
Sverre

Hi Marek, thanks for the reply,

Regarding the 0/1 probabilities, they have worked well for me. The network evaluation will fail if it is impossible to explain the observed evidence, but this is as expected. I see the danger with divisions by 0, but have not experienced any problems yet. Anything i ...

Hi!
How is the computational complexity of having multiple nodes with small CPTs and target set on only the top node vs one single node with a large CPT?
I often stand between making a single node with few inputs, or merging them together to one node with many inputs. The first option is easier to ...

Thanks for the reply Marek :)
The first article you referred to was actually the one I read to understand virtual evidence, its quite good som maybe it could be linked to in the wiki? :)

I understand that it is a hazel to allow unnormalized virtual evidence and that this is not a priority. But, at ...