1 paper
Wen-Ding Li, Keya Hu, Carter Larsen +11
When learning an input-output mapping from very few examples, is it better to first infer a latent function that explains the examples, or is it better to directly predict new test…