Arthur, the machine learning model monitoring company, released a suite of new tools and features for monitoring natural language processing models. Natural language processing is one of the most widely adopted machine learning technologies in the enterprise. But organizations often struggle to find the right tools to monitor these models.

The Arthur platform now offers advanced performance monitoring for NLP models, including tracking data drift, bias detection, and prediction-level model explainability. Monitoring NLP models for data drift involves comparing the statistical similarity of new input documents to the documents used to train the model. The Arthur platform automatically alerts you when your input documents or output text starts drifting beyond pre-configured thresholds.

Arthur now also offers bias detection capabilities for NLP models, allowing data science teams to uncover differences in accuracy and other performance measures across different subgroups to identify and fix unfair model bias. The platform also offers performance-bias analysis for tabular models. The Arthur team has also released a new set of explainability tools for NLP models, providing token-level insights for language models. Organizations can now understand which specific words within a document contributed most to a given prediction, even for black-box models.