the current inspire classifier model is trained by assuming all of quant-ph is core. This is not the case, so it should be retrained by taking core/non-core based on INSPIRE content:
Note that this is proably equivalent to stoping the special-casing of quantum, but we should still make sure we have enough samples in the training set, as it's a very small fraction of the database currently.
the current inspire classifier model is trained by assuming all of quant-ph is core. This is not the case, so it should be retrained by taking core/non-core based on INSPIRE content:
Note that this is proably equivalent to stoping the special-casing of quantum, but we should still make sure we have enough samples in the training set, as it's a very small fraction of the database currently.