Resource-constrained perception systems such as edge computing and
visio...
Object detectors often suffer from the domain gap between training (sour...
Existing neural architecture search (NAS) methods often return an
archit...
An important step in the task of neural network design, such as
hyper-pa...
DETR is a recently proposed Transformer-based method which views object
...
We propose a method to incrementally learn an embedding space over the d...