Partial Order Pruning: For Best Speed/Accuracy Trade-Off in Neural Architecture Search

Li, Xin; Zhou, Yiming; Zheng, Pan; Feng, Jiashi

doi:10.1109/cvpr.2019.00936

Cited by 128 publications

(73 citation statements)

References 26 publications

(83 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…architecture depends on the difficulty and size of the dataset at hand. While these findings may encourage an automated neural architecture search, such an approach is hindered by the limited computational resources [19], [20], [21], [22], [23]. Alternatively, we propose an ensemble architecture, which combines U-Nets of varying depths into one unified structure.…”

Section: Table Imentioning

confidence: 99%

UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation

Zhou

Siddiquee

Tajbakhsh

et al. 2020

IEEE Trans. Med. Imaging

2,002

876

View full text Add to dashboard Cite

The state-of-the-art models for medical image segmentation are variants of U-Net and fully convolutional networks (FCN). Despite their success, these models have two limitations:(1) their optimal depth is apriori unknown, requiring extensive architecture search or inefficient ensemble of models of varying depths; and (2) their skip connections impose an unnecessarily restrictive fusion scheme, forcing aggregation only at the samescale feature maps of the encoder and decoder sub-networks. To overcome these two limitations, we propose UNet++, a new neural architecture for semantic and instance segmentation, by (1) alleviating the unknown network depth with an efficient ensemble of U-Nets of varying depths, which partially share an encoder and co-learn simultaneously using deep supervision; (2) redesigning skip connections to aggregate features of varying semantic scales at the decoder sub-networks, leading to a highly flexible feature fusion scheme; and (3) devising a pruning scheme to accelerate the inference speed of UNet++. We have evaluated UNet++ using six different medical image segmentation datasets, covering multiple imaging modalities such as computed tomography (CT), magnetic resonance imaging (MRI), and electron microscopy (EM), and demonstrating that (1) UNet++ consistently outperforms the baseline models for the task of semantic segmentation across different datasets and backbone architectures; (2) UNet++ enhances segmentation quality of varying-size objects-an improvement over the fixed-depth U-Net; (3) Mask RCNN++ (Mask R-CNN with UNet++ design) outperforms the original Mask R-CNN for the task of instance segmentation; and (4) pruned UNet++ models achieve significant speedup while showing only modest performance degradation. Our implementation and pre-trained models are available at https://github.com/MrGiovanni/UNetPlusPlus.

show abstract

Section: Table Imentioning

confidence: 99%

UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation

Zhou

Siddiquee

Tajbakhsh

et al. 2020

IEEE Trans. Med. Imaging

2,002

876

View full text Add to dashboard Cite

show abstract

“…The architecture search step is based on: 1) the evolutionary algorithm to mutate the best architecture on the Pareto front; 2) Partial Order Pruning method (Li et al 2019a) to prune the architecture search space with the prior knowledge that deeper models and wider models are better. Our algorithm can be parallelized on multiple computation nodes (each has 8 V100 GPUs) and lift the Pareto front simultaneously.…”

Section: Multi-objective Search Algorithmmentioning

confidence: 99%

“…We also found that MobileNet V2 is dominated by other models although it has mush less FLOPs in Figure 3-2. This is because it has higher memory access cost thus is slower in practice (Li et al 2019a). Therefore, using the direct metric, i.e.…”

Section: Experiments Architecture Search Detailsmentioning

confidence: 99%

“…With the improved training strategy, our search can be conducted directly on the detection datasets without ImageNet pre-training. For an efficient search, we combine evolutionary algorithms (Real et al;Real et al 2017; with Partial Order Pruning technique (Li et al 2019a) for a fast search and parallelize the whole searching algorithm in a distributed training system to further speed up the whole process.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

SM-NAS: Structural-to-Modular Neural Architecture Search for Object Detection

Yao

Zhang

et al. 2020

AAAI

View full text Add to dashboard Cite

The state-of-the-art object detection method is complicated with various modules such as backbone, RPN, feature fusion neck and RCNN head, where each module may have different designs and structures. How to leverage the computational cost and accuracy trade-off for the structural combination as well as the modular selection of multiple modules? Neural architecture search (NAS) has shown great potential in finding an optimal solution. Existing NAS works for object detection only focus on searching better design of a single module such as backbone or feature fusion neck, while neglecting the balance of the whole system. In this paper, we present a two-stage coarse-to-fine searching strategy named Structural-to-Modular NAS (SM-NAS) for searching a GPU-friendly design of both an efficient combination of modules and better modular-level architecture for object detection. Specifically, Structural-level searching stage first aims to find an efficient combination of different modules; Modular-level searching stage then evolves each specific module and pushes the Pareto front forward to a faster task-specific network. We consider a multi-objective search where the search space covers many popular designs of detection methods. We directly search a detection backbone without pre-trained models or any proxy task by exploring a fast training from scratch strategy. The resulting architectures dominate state-of-the-art object detection systems in both inference time and accuracy and demonstrate the effectiveness on multiple detection datasets, e.g. halving the inference time with additional 1% mAP improvement compared to FPN and reaching 46% mAP with the similar inference time of MaskRCNN.

show abstract

“…Moreover, the input resolution of our model is the original high resolution of 1024 × 2048, but our FSFNet is the fastest model, which has a small number of parameters and high accuracy. Comparing our FSFNet to the DF1-Seg-d8 [50] and Fasterseg [35] models, mIoU is 2.2% and 2.3% lower, but the inference speed is faster by 48% and 23%, respectively. Thus, we can confirm that our FSFNet has a good trade-off between semantic segmentation accuracy and computational resources in existing state-of-the-art semantic segmentation.…”

Section: 72mentioning

confidence: 99%

Accelerator-Aware Fast Spatial Feature Network for Real-Time Semantic Segmentation

2020

View full text Add to dashboard Cite

Partial Order Pruning: For Best Speed/Accuracy Trade-Off in Neural Architecture Search

Cited by 128 publications

References 26 publications

UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation

UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation

SM-NAS: Structural-to-Modular Neural Architecture Search for Object Detection

Accelerator-Aware Fast Spatial Feature Network for Real-Time Semantic Segmentation

Contact Info

Product

Resources

About