(i) Linear complexity: we introduce a novel patch attention with linear complexity.
The energy consumption of deep learning models is increasing at a breathtaking rate, which raises concerns due to potential negative effects on carbon neutrality in the context of global warming and climate change.
In our work, we target the optimization of hardware and software configurations on an industry-standard edge accelerator.
To resolve this problem, we propose Isometric Propagation Network (IPN), which learns to strengthen the relation between classes within each space and align the class dependency in the two spaces.
End-to-end training is made possible by differentiable registration and 3D triangulation modules.
As a result, AutoML can be reformulated as an automated process of symbolic manipulation.
In this paper, we study the importance of co-designing neural architectures and hardware accelerators.
Few-shot learning aims to train a classifier given only a few samples per class that are highly insufficient to describe the whole data distribution.
In this paper, we propose NATS-Bench, a unified benchmark on searching for both topology and size, for (almost) any up-to-date NAS algorithm.
Efficient hyperparameter or architecture search methods have shown remarkable results, but each of them is only applicable to searching for either hyperparameters (HPs) or architectures.
In this paper, we propose a Self-Evaluated Template Network (SETN) to improve the quality of the architecture candidates for evaluation so that it is more likely to cover competitive candidates.
Ranked #13 on Neural Architecture Search on NAS-Bench-201, CIFAR-100
To avoid traversing all the possibilities of the sub-graphs, we develop a differentiable sampler over the DAG.
A typical approach is to (1) train a detector on the labeled images; (2) generate new training samples using this detector's prediction as pseudo labels of unlabeled images; (3) retrain the detector on the labeled samples and partial pseudo labeled samples.
Ranked #1 on Facial Landmark Detection on 300W (Full) (using extra training data)
The maximum probability for the size in each distribution serves as the width and depth of the pruned network, whose parameters are learned by knowledge transfer, e. g., knowledge distillation, from the original networks.
Ranked #1 on Network Pruning on CIFAR-10
We propose to automatically search for a CNN architecture that is specifically suitable for the reID task.
Ranked #8 on Person Re-Identification on CUHK03 detected
With asymptotic pruning, the information of the training set would be gradually concentrated in the remaining filters, so the subsequent training and pruning process would be stable.
In this paper, we present supervision-by-registration, an unsupervised approach to improve the precision of facial landmark detectors on both images and video.
Ranked #1 on Facial Landmark Detection on 300-VW (C)
We focus on the one-shot learning for video-based person re-Identification (re-ID).
In this work, we propose a style-aggregated approach to deal with the large intrinsic variance of image styles for facial landmark detection.
Ranked #1 on Facial Landmark Detection on AFLW-Front (Mean NME metric)
For most state-of-the-art architectures, Rectified Linear Unit (ReLU) becomes a standard component accompanied with each layer.
Ranked #11 on Image Classification on SVHN
During co-training process, labels of unlabeled instances in the training pool are very likely to be false especially in the initial training rounds, while the standard co-training algorithm utilizes a “draw without replacement” manner and does not remove these false labeled instances from training.
Experiments on PASCAL VOC'07, MS COCO'14, and ILSVRC'13 indicate that by using as few as three or four samples selected for each category, our method produces very competitive results when compared to the state-of-the-art weakly-supervised approaches using a large number of image-level labels.
Ranked #1 on Weakly Supervised Object Detection on COCO
In this paper, we present a novel and general network structure towards accelerating the inference process of convolutional neural networks, which is more complicated in network structure yet with less inference complexity.