Igeood: An Information Geometry Approach to Out-of-Distribution Detection
This addresses the problem of unreliable OOD detection for safer ML systems, representing a novel method for a known bottleneck.
The paper tackles out-of-distribution detection in machine learning by introducing Igeood, a method that uses information geometry and Fisher-Rao distance to combine logits and features, resulting in outperforming state-of-the-art methods across various architectures and datasets.
Reliable out-of-distribution (OOD) detection is fundamental to implementing safer modern machine learning (ML) systems. In this paper, we introduce Igeood, an effective method for detecting OOD samples. Igeood applies to any pre-trained neural network, works under various degrees of access to the ML model, does not require OOD samples or assumptions on the OOD data but can also benefit (if available) from OOD samples. By building on the geodesic (Fisher-Rao) distance between the underlying data distributions, our discriminator can combine confidence scores from the logits outputs and the learned features of a deep neural network. Empirically, we show that Igeood outperforms competing state-of-the-art methods on a variety of network architectures and datasets.