ECCV 2018 accepted papers on Architectures and efficiency
46 of the 776 papers accepted at ECCV 2018 (5.9%) are on Architectures and efficiency, compared with 6.0% at ECCV 2016. They fall into 5 topics.
See ECCV 2018 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
CNN design · 20 papers
- Scale Aggregation Network for Accurate and Efficient Crowd Counting (773 citations)
- SkipNet: Learning Dynamic Routing in Convolutional Networks (688 citations)
- Bi-Real Net: Enhancing the Performance of 1-bit CNNs with Improved Representational Capability and Advanced Training Algorithm (589 citations)
- Shift-Net: Image Inpainting via Deep Feature Rearrangement (539 citations)
- Learning SO(3) Equivariant Representations with Spherical CNNs (395 citations)
- Convolutional Networks with Adaptive Inference Graphs (373 citations)
- Quaternion Convolutional Neural Networks (214 citations)
- Learning to Zoom: a Saliency-Based Sampling Layer for Neural Networks (159 citations)
- CubeNet: Equivariance to 3D Rotation and Translation (133 citations)
- TBN: Convolutional Neural Network with Ternary Inputs and Binary Weights (105 citations)
- Grassmann Pooling as Compact Homogeneous Bilinear Pooling for Fine-Grained Visual Classification (100 citations)
- Clustering Convolutional Kernels to Compress Deep Neural Networks (85 citations)
and 8 more in the interactive map.
RNNs and normalization · 10 papers
- Group Normalization (4,036 citations)
- NetAdapt: Platform-Aware Neural Network Adaptation for Mobile Applications (607 citations)
- Dynamic Task Prioritization for Multitask Learning (395 citations)
- A Modulation Module for Multi-task Learning with Applications in Image Retrieval (135 citations)
- Adding Attentiveness to the Neurons in Recurrent Neural Networks (105 citations)
- Revisiting the Inverted Indices for Billion-Scale Approximate Nearest Neighbors (104 citations)
- Constrained Optimization Based Low-Rank Approximation of Deep Neural Networks (60 citations)
- Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching (55 citations)
- Training Binary Weight Networks via Semi-Binary Decomposition (18 citations)
- Lifting Layers: Analysis and Applications (8 citations)
Quantization · 7 papers
- LQ-Nets: Learned Quantization for Highly Accurate and Compact Deep Neural Networks (740 citations)
- Value-aware Quantization for Training and Inference of Neural Networks (200 citations)
- Quantized Densely Connected U-Nets for Efficient Landmark Localization (141 citations)
- Constraint-Aware Deep Neural Network Compression (61 citations)
- Statistically-motivated Second-order Pooling (50 citations)
- LSQ++: Lower running time and higher recall in multi-codebook quantization (37 citations)
- Learning Compression from Limited Unlabeled Data (29 citations)
Pruning and efficient tuning · 5 papers
- AMC: AutoML for Model Compression and Acceleration on Mobile Devices (1,477 citations)
- Data-Driven Sparse Structure Selection for Deep Neural Networks (598 citations)
- A Systematic DNN Weight Pruning Framework using Alternating Direction Method of Multipliers (494 citations)
- Neural Network Encapsulation (53 citations)
- Interpolating Convolutional Neural Networks Using Batch Normalization (6 citations)
Neural architecture search · 4 papers
- ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design (7,645 citations)
- Progressive Neural Architecture Search (1,988 citations)
- DPP-Net: Device-aware Progressive Search for Pareto-optimal Neural Architectures (200 citations)
- MaskConnect: Connectivity Learning by Gradient Descent (54 citations)
