BayesFusion Support Forum

Hi, I just wanted to report a bug in GeNIe/SMILE: 'marginalize' option fails for the case of a noisy-max node. Original network had 3 parents to the noisy-max node, I selected option 'marginalize' on one of the parents. In the result the noisy-max parameters (leak) got totally corrupted, as can be s...

Tomek,

You are genius (it's not like it's the first time I think so...).

It worked flawlessly. Thanks!

Adam

Sorry this time it should work: genie_problem.png I prepared the second screenshot, that should let you appreciate the scale of the problem. One of quick workarounds would be to allow for zoom greater that 200%, because even 200% is a bit limiting. Keep in mind that the picture is actually re-scaled...

Dear GeNIe developer, I've just installed GeNIe on a new laptop with high resolution screen (3840x2160) and GeNIe toolbars and canvas space) are extremely small (making them virtually unusable). Basically, whole layout is rather a mess. I wonder if it would be possible to fix in the near future, as ...

Hi, I was wondering if it is possible to access cases created using Case Manager (that hidden feature that you can access using a secret key combination) using JSMILE? If not, is there any other mechanism that allows to save evidence set in the model and later retrieve it from JSMILE? I'm not doing ...

Andy, There's no mistake in SMILE -- the result should be 54. In your calculations you make incorrect assumption that S1 and S2 are independent, while clearly in your model they aren't. Instead of P(S1=state_x1)*P(S2-state_x2)*U(S1=state_x1,S2=state_x2) you should use P(S1=state_x1)*P(S2-state_x2|S1...

Jochen, It's not theoretically necessary, but just a feature. For algorithms, anyway SMILE converts the original network into a different data structures. But please be aware that SMILE does not exploit noisy-MAX for inference -- there are no special algorithms implemented that take advantage of noi...

Jochen, The parent strengths are necessary because the same CPT can be a parent to two different noisy-MAX nodes and those nodes can have different states of the common parent CPT node as distinguished for them. Technically, for the noisy-MAX, only the distinguished state matters, the order of other...

Jochen The short answer is: it's impossible in SMILE to define the 'distinguished' (false) state of the noisy-MAX as the outcome with number 0. A possible workaround would be to introduce an auxiliary node which would be a child to the noisy-MAX gate and had the reverse states. It would have a deter...

- What reasoning takes place about noisy-MAX definitions? (EnableNoisyMAXRelevance, where it says "Adam's noisy-MAX inference algorithm/improvements 11/04") Jonnie, EnableNoisyMAXRelevance() enables an extra step in relevance. It is related to the noisy-OR/noisy-MAX definitions. When a no...

Johannes, I think I may be missing something from your description. If your original distribution if noisy-OR, you can directly enter it by creating a noisy-MAX node in SMILE and simply provide noisy-OR/MAX parameters (not a full CPT!) directly using SetDefinition() from the DSL_noisyMAX class. How ...

Johannes, SMILE does not provide means to identify the optimal order of parents' states. You should do that on your own. I'm assuming that you are aware of the functionality for assigning strengths of the influences (order of parents' states) that is implemented in the DSL_noisyMAX class (GetParentO...

Hi,

I was wondering if it makes sense and it would be possible to remove the constraint in the Confidence parameter in the 'Learn Parameters with EM' dialog box that limits values to integers. Theoretically they can be any non-negative values, right?

Cheers,
Adam

By any chance, is it:
.. that would be a bit unorthodox approach...

Hi, I wonder how to set the options for multiple fault algorithm in SMILE for DIAG_Network. There is no documantation for that or nothing obvious in the DIAG_Network class. Where to plug these options: ??? // Approaches for multiple fault diagnosis #define DSL_DIAG_MARGINAL 1 #define DSL_DIAG_INDEPE...