On the Detection of Anomalous or Out-Of-Distribution Data in Vision Models Using Statistical Techniques
This addresses vulnerabilities in machine learning systems for vision applications, but it appears incremental as it explores an underexplored technique without strong performance claims.
The paper tackles the problem of detecting anomalous or out-of-distribution data in vision models, which can cause incorrect predictions, by assessing Benford's law as a method to quantify differences between real and corrupted inputs, with the result being a proposed filter for such data.
Out-of-distribution data and anomalous inputs are vulnerabilities of machine learning systems today, often causing systems to make incorrect predictions. The diverse range of data on which these models are used makes detecting atypical inputs a difficult and important task. We assess a tool, Benford's law, as a method used to quantify the difference between real and corrupted inputs. We believe that in many settings, it could function as a filter for anomalous data points and for signalling out-of-distribution data. We hope to open a discussion on these applications and further areas where this technique is underexplored.