Recent works along these lines learn task-dependent regularizers.

With the built relation matrix and semantic-level representations, pixel representations can be augmented for pixel prediction.

In this work, we present a dual-stage low-bit post-training quantization (PTQ) method for image super-resolution, namely 2DQuant, which achieves efficient and accurate SR under low-bit quantization.

Imitation learning uses data for training policies to solve complex tasks.