Talking to Models: Stanford U & Microsoft Method Enables Developers to Correct Model Bugs via Natural Language Patches

In the new paper Fixing Model Bugs with Natural Language Patches, researchers from Stanford University and Microsoft Research propose a method that uses declarative statements as feedback for correcting errors in neural models, significantly increasing accuracy without high compute costs.

ai artificial intelligence bugs deep-neural-networks developers language language model machine learning machine learning & data science microsoft ml natural natural language natural language processing nature language tech research stanford technology