Pre-training has been widely adopted in deep learning to improve model
p...
We propose Neural Priming, a technique for adapting large pretrained mod...
Large multimodal datasets have been instrumental in recent breakthroughs...
Compositional representations of the world are a promising step towards
...
Learned representations are a central component in modern ML systems, se...
In visual retrieval systems, updating the embedding model requires
recom...
Learning binary representations of instances and classes is a classical
...
The capacity of neural networks like the widely adopted transformer is k...
We present the Supermasks in Superposition (SupSup) model, capable of
se...
Sparsity in Deep Neural Networks (DNNs) is studied extensively with the ...
Training a neural network is synonymous with learning the values of the
...
Unsupervised image-to-image translation techniques are able to map local...