Fast and Accurate Transferability Measurement for Heterogeneous Multivariate Data
This addresses the challenge of efficient transfer learning for heterogeneous multivariate data, offering a practical solution for applications requiring fast dataset selection, though it is incremental as it builds on existing methods like adversarial networks and encoder-decoder architectures.
The paper tackles the problem of quickly identifying the most useful source dataset for a target task when datasets have different features and distributions, proposing Transmeter, which achieves up to 10.3 times faster performance than competitors while providing accurate transferability measurements.
Given a set of heterogeneous source datasets with their classifiers, how can we quickly find the most useful source dataset for a specific target task? We address the problem of measuring transferability between source and target datasets, where the source and the target have different feature spaces and distributions. We propose Transmeter, a fast and accurate method to estimate the transferability of two heterogeneous multivariate datasets. We address three challenges in measuring transferability between two heterogeneous multivariate datasets: reducing time, minimizing domain gap, and extracting meaningful homogeneous representations. To overcome the above issues, we utilize a pre-trained source model, an adversarial network, and an encoder-decoder architecture. Extensive experiments on heterogeneous multivariate datasets show that Transmeter gives the most accurate transferability measurement with up to 10.3 times faster performance than its competitor. We also show that selecting the best source data with Transmeter followed by a full transfer leads to the best transfer accuracy and the fastest running time.