Quantifying Local Model Validity using Active Learning
This work addresses the need for cost-effective local validity assessment in regulated machine learning applications, though it is incremental as it builds on active learning techniques.
The paper tackles the problem of efficiently estimating local model validity under regulatory constraints by learning the model error using active learning, resulting in a method that achieves sufficient discriminative properties with relatively small data and increased sensitivity to local validity bounds.
Real-world applications of machine learning models are often subject to legal or policy-based regulations. Some of these regulations require ensuring the validity of the model, i.e., the approximation error being smaller than a threshold. A global metric is generally too insensitive to determine the validity of a specific prediction, whereas evaluating local validity is costly since it requires gathering additional data.We propose learning the model error to acquire a local validity estimate while reducing the amount of required data through active learning. Using model validation benchmarks, we provide empirical evidence that the proposed method can lead to an error model with sufficient discriminative properties using a relatively small amount of data. Furthermore, an increased sensitivity to local changes of the validity bounds compared to alternative approaches is demonstrated.