Out-of-Distribution Detection through Soft Clustering with Non-Negative Kernel Regression
This addresses the need for efficient OOD detection in general-purpose language models, offering significant computational and storage benefits, though it is incremental as it builds on existing OOD detection methods.
The paper tackles the problem of detecting out-of-distribution instances in language models by proposing a soft clustering approach with non-negative kernel regression, resulting in up to 11x faster inference, 87% lower storage, and up to 4 AUROC point improvements on benchmarks.
As language models become more general purpose, increased attention needs to be paid to detecting out-of-distribution (OOD) instances, i.e., those not belonging to any of the distributions seen during training. Existing methods for detecting OOD data are computationally complex and storage-intensive. We propose a novel soft clustering approach for OOD detection based on non-negative kernel regression. Our approach greatly reduces computational and space complexities (up to 11x improvement in inference time and 87% reduction in storage requirements) and outperforms existing approaches by up to 4 AUROC points on four different benchmarks. We also introduce an entropy-constrained version of our algorithm, which leads to further reductions in storage requirements (up to 97% lower than comparable approaches) while retaining competitive performance. Our soft clustering approach for OOD detection highlights its potential for detecting tail-end phenomena in extreme-scale data settings.