BayesFusion Support Forum

There's a new feature in GeNIe that lets you marginalize a node into its child. It removes the node and instead adds the node's parents to the child's parents. Also, Smile has a method int DSL_network::MarginalizeNode(int theParent, int theChild) that lets you marginalize a node into its child. Howe...

There's a new feature in GeNIe that lets you marginalize a node into its child. First I want to say that's a great new feature! :D Right-click on a node and select Marginalize. The node gets removed. Its CPD gets merged into the CPD of its child, and the node's parents become parents of the child. I...

Hello everyone, this is a theory question about finding noisy influences and automatically divorcing parents, and how to do this in Smile. Nodes can be automatically divorced if they have a noisy definition: build a binary tree from the parents to the child, put the noisy influences into the nodes n...

Oh, you're correct. I tested with a slightly different model; I'll PM that to you.

Too much honor! I can't wait to see it

I'm using C++ directly, and I've seen in the task manager that the program does use the whole ram when I'm just generating enough test cases... so it's definitely not an out-of-memory issue. Btw I noticed that the assertion also fails when my relevance flags are set like this: SetRelevanceLevelFlags...

Hi orzech,
I've isolated the failing case and wrote a function that does nothing else than applying the case and crash. Plus, I'm sitting on 8GB ram here

Update: after conducting experiments, I found that the bug occurs:

• in ALL approximate algorithms, plus in LauritzenOld,
• when noisy relevance is enabled, and the model contains noisy nodes,
• only when evidence is set.

(Though I don't know if the latter is valid in general or only in my cases.)

Update: The crash is a segfault because of a NULL pointer dereference. The bug lies in 15yo legacy code and is therefore not probable to be fixed soon.
Workaround: don't use this combination

Update: My evidence WAS conflicting. The conflict check at SetEvidence time is not almighty, and if you really need to verify beyond any doubt that a certain set of evidence is non-conflicting, you have to perform a full inference without targets and then look if all values are set valid. The newest...

Hello, I stumbled upon my next crash :) I hope it's not a bug again since I haven't seen any fixes for the previous two I submitted :oops: So I used plain Lauritzen and I have some evidences and targets set. I disable relevance reasoning with DSL_network::DeactivateRelevance() and then call UpdateBe...

I'll see if the error occurs on an obfuscated model as well, and PM you about it.
Thanks for taking care of this

Hello, I recently discovered the possibility to activate/deactivate relevance reasoning in SMILE and promptly stumbled over some crashes. It happened with other algorithms as well (eg EpisSampling) but one case that I have isolated is using Backsampling with noisy relevance reasoning enabled (Enable...

Hello Adam, Yes I understand the concept of ordering the strengthes of parents' states in the noisyMAX definition. I implemented a simple heuristic (in fact, in my network it's always the first state which is the leak state) but I don't know how to use it: If I create the noisyMAX from the CPT first...

It happens regularly while generating random evidence in the network. I can extract one of the situations as a "Case" and store it in a new .xdsl file; I open it in GeNIe, use the Case, and perform inference. Same result: some nodes don't get valid values and the little question mark at bo...