Skip to content

add optional probs parameter to log_likelihood functions. Using this … #28

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 3 commits into from
Closed

add optional probs parameter to log_likelihood functions. Using this … #28

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
0fce752
add optional probs parameter to log_likelihood functions. Using this …
colibri-coruscans Jul 9, 2017
3012e56
Merge remote-tracking branch 'origin/develop' into develop
colibri-coruscans Aug 2, 2017
aeb3eb0
Merge remote-tracking branch 'origin/develop' into develop
colibri-coruscans Aug 3, 2017
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 12 additions & 5 deletions packages/pygsti/tools/likelihoodfns.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -173,7 +173,7 @@ def fill_count_vecs(mxToFill, spam_label_rows, dataset, gatestring_list):
def logl_terms(gateset, dataset, gatestring_list=None,
minProbClip=1e-6, probClipInterval=(-1e6,1e6), radius=1e-4,
evalTree=None, countVecMx=None, totalCntVec=None, poissonPicture=True,
check=False, gateLabelAliases=None):
check=False, gateLabelAliases=None, probs=None):
"""
The vector of log-likelihood contributions for each gate string & SPAM label

Expand All @@ -200,7 +200,6 @@ def logl_terms(gateset, dataset, gatestring_list=None,
spamLabels = gateset.get_spam_labels() #this list fixes the ordering of the spam labels
spam_lbl_rows = { sl:i for (i,sl) in enumerate(spamLabels) }

probs = _np.empty( (len(spamLabels),len(gatestring_list)), 'd' )
if countVecMx is None:
countVecMx = _np.empty( (len(spamLabels),len(gatestring_list)), 'd' )
fill_count_vecs(countVecMx, spam_lbl_rows, dataset, ds_gatestring_list)
Expand All @@ -219,7 +218,9 @@ def logl_terms(gateset, dataset, gatestring_list=None,
if evalTree is None:
evalTree = gateset.bulk_evaltree(gatestring_list)

gateset.bulk_fill_probs(probs, spam_lbl_rows, evalTree, probClipInterval, check)
if probs is None:
probs = _np.empty( (len(spamLabels),len(gatestring_list)), 'd' )
gateset.bulk_fill_probs(probs, spam_lbl_rows, evalTree, probClipInterval, check)
pos_probs = _np.where(probs < min_p, min_p, probs)

if poissonPicture:
Expand Down Expand Up @@ -253,7 +254,7 @@ def logl_terms(gateset, dataset, gatestring_list=None,
def logl(gateset, dataset, gatestring_list=None,
minProbClip=1e-6, probClipInterval=(-1e6,1e6), radius=1e-4,
evalTree=None, countVecMx=None, totalCntVec=None, poissonPicture=True,
check=False, gateLabelAliases=None):
check=False, gateLabelAliases=None, probs=None):
"""
The log-likelihood function.

Expand Down Expand Up @@ -314,6 +315,12 @@ def logl(gateset, dataset, gatestring_list=None,
the dataset. Defaults to the empty dictionary (no aliases defined)
e.g. gateLabelAliases['Gx^3'] = ('Gx','Gx','Gx')

probs : numpy array, optional
Two-dimensional numpy array whose rows correspond to the gate's spam
labels (i.e. gateset.get_spam_labels()). Each row contains the
expected probability for that spam label for each gate string in gatestring_list.
Use gateset.bulk_fill_probs(...) to generate this quantity once for multiple
evaluations of the log-likelihood function which use the same gateset.

Returns
-------
Expand All @@ -323,7 +330,7 @@ def logl(gateset, dataset, gatestring_list=None,
v = logl_terms(gateset, dataset, gatestring_list,
minProbClip, probClipInterval, radius,
evalTree, countVecMx, totalCntVec, poissonPicture,
check, gateLabelAliases)
check, gateLabelAliases, probs)
return _np.sum(v) # sum over *all* dimensions


Expand Down