Is it not possible to continuously train a model with negative scenarios and create a pattern like what happens in cyber-security protection? A group of people continue to identify negative data sets that are explicitly tagged as failure/negative data sets while we try to understand why the model is getting influenced by the negative scenarios.
Is it not possible to continuously train a model with negative scenarios and create a pattern like what happens in cyber-security protection? A group of people continue to identify negative data sets that are explicitly tagged as failure/negative data sets while we try to understand why the model is getting influenced by the negative scenarios.