EFFICIENT MULTI-TASK MODEL SELECTION

Summary from the TLDR AI newsletter

Predicting the performance of pre-trained neural networks on multiple tasks without fine-tuning them is a challenge. The traditional method is computationally costly, while modern methods are limited to single tasks. This paper introduces EMMS, which transforms varied task labels into a unified format, allowing for the prediction of a model's effectiveness across tasks efficiently.