The standard definition generation task requires to automatically produc...
The first ACM/IEEE TinyML Design Contest (TDC) held at the 41st Internat...
Pruning neural networks before training has received increasing interest...
Recently, pre-trained transformer-based models have achieved great succe...
Existing commonsense knowledge bases often organize tuples in an isolate...
The core issue of cyberspace detecting and mapping is to accurately iden...
Cyber attacks pose crucial threats to computer system security, and put
...
The Barzilai-Borwein (BB) method has demonstrated great empirical succes...
A reasonable prediction of infectious diseases transmission process unde...
One of the major concerns for neural network training is that the
non-co...
Does over-parameterization eliminate sub-optimal local minima for neural...
We customize an end-to-end image compression framework for retina OCT im...
Automatic leaf segmentation, as well as identification and classificatio...
Change detection is a fundamental task in computer vision. Despite
signi...
Object detection models shipped with camera-equipped mobile devices cann...
Lifelong learning, the problem of continual learning where tasks arrive ...
Deep neural networks (DNNs) often suffer from "catastrophic forgetting"
...
State-of-the-art deep model compression methods exploit the low-rank
app...
In this paper, we study the loss surface of the over-parameterized fully...
Recurrent neural networks (RNNs) achieve cutting-edge performance on a
v...
Machine lipreading is a special type of automatic speech recognition (AS...