LEEP: A New Measure to Evaluate Transferability of Learned Representations
This work addresses the need for better evaluation metrics in transfer learning for machine learning practitioners, though it is incremental as it builds on existing measures.
The paper tackles the problem of evaluating the transferability of learned representations by introducing LEEP, a simple and easy-to-compute measure that predicts performance and convergence speed in transfer learning, achieving up to 30% improvement in correlation with actual transfer accuracy compared to competing methods on ImageNet to CIFAR100.
We introduce a new measure to evaluate the transferability of representations learned by classifiers. Our measure, the Log Expected Empirical Prediction (LEEP), is simple and easy to compute: when given a classifier trained on a source data set, it only requires running the target data set through this classifier once. We analyze the properties of LEEP theoretically and demonstrate its effectiveness empirically. Our analysis shows that LEEP can predict the performance and convergence speed of both transfer and meta-transfer learning methods, even for small or imbalanced data. Moreover, LEEP outperforms recently proposed transferability measures such as negative conditional entropy and H scores. Notably, when transferring from ImageNet to CIFAR100, LEEP can achieve up to 30% improvement compared to the best competing method in terms of the correlations with actual transfer accuracy.