Incorrect sensitivities when a sigmay expression depends on a different observable #2561
Assignees
Comments
dweindl
added a commit
to dweindl/AMICI
that referenced
this issue
Oct 23, 2024
…ndent sigmas Due to PEtab not allowing and observables in noiseFormula and AMICI not allowing state variables in sigma expressions, we are trying to replace any state variables by matching observables during PEtab import. However, if there were multiple observables with the same observableFormula, or one observableFormula was contained in another one, the substition of observableFormula for observableId could - depending on the ordering - result in noiseFormulas depending on other observables, which could lead to incorrect sensivities due to AMICI-dev#2561. This change ensures that we are only using the observableId from the same row of the observable table as the noiseFormula.
dweindl
added a commit
to dweindl/AMICI
that referenced
this issue
Oct 23, 2024
…ndent sigmas Due to PEtab not allowing and observables in noiseFormula and AMICI not allowing state variables in sigma expressions, we are trying to replace any state variables by matching observables during PEtab import. However, if there were multiple observables with the same observableFormula, or one observableFormula was contained in another one, the substition of observableFormula for observableId could - depending on the ordering - result in noiseFormulas depending on other observables, which could lead to incorrect sensivities due to AMICI-dev#2561. This change ensures that we are only using the observableId from the same row of the observable table as the noiseFormula. --------- Co-authored-by: Dilan Pathirana <[email protected]>
dweindl
added a commit
to dweindl/AMICI
that referenced
this issue
Oct 23, 2024
dweindl
added a commit
to dweindl/AMICI
that referenced
this issue
Oct 23, 2024
dweindl
added a commit
to dweindl/AMICI
that referenced
this issue
Oct 23, 2024
dweindl
added a commit
that referenced
this issue
Oct 24, 2024
…ndent sigmas (#2562) Due to PEtab not allowing and observables in noiseFormula and AMICI not allowing state variables in sigma expressions, we are trying to replace any state variables by matching observables during PEtab import. However, if there were multiple observables with the same observableFormula, or one observableFormula was contained in another one, the substition of observableFormula for observableId could - depending on the ordering - result in noiseFormulas depending on other observables, which could lead to incorrect sensitivities due to #2561. This change ensures that we are only using the observableId from the same row of the observable table as the noiseFormula. --------- Co-authored-by: Dilan Pathirana <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
If
sigmay[i]depends ony[j], with i !=j, then sensitivities will be incorrect.We need to raise an error during import or support that.
The text was updated successfully, but these errors were encountered: