AI Paper Atlas

Research map / Architectures and efficiency / Quantization

ReActNet: Towards Precise Binary Neural Network with Generalized Activation Functions

Zechun Liu, Zhiqiang Shen, Marios Savvides, Kwang-Ting Cheng

ECCV 2020 · Architectures and efficiency › Quantization · approach: Binary neural networks · 408 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add concatenated duplicated downsampling blocks (ablation)baseline network · ImageNet valtop-1 accuracy: 59.6 -> 62.5Improved
add distributional loss (ablation)proposed baseline network · ImageNet valtop-1 accuracy: 61.1 -> 62.5Improved
add distributional loss (ablation)baseline without concatenation · ImageNet valtop-1 accuracy: 58.2 -> 59.6Improved
add RSign (ablation)proposed baseline network · ImageNet valtop-1 accuracy: 62.5 -> 66.1Improved
add RPReLU (ablation)proposed baseline network · ImageNet valtop-1 accuracy: 62.5 -> 67.4Improved
add both RSign and RPReLU (ablation)proposed baseline network · ImageNet valtop-1 accuracy: 62.5 -> 69.4Improved
use RSign and RPReLU instead of Sign and PReLU (ablation)proposed baseline network · ImageNet valtop-1 accuracy: 65.5 -> 69.4Improved
add PReLU (ablation)proposed baseline network · ImageNet valtop-1 accuracy: 62.5 -> 65.5Improved
apply ReAct operations (ablation)Bi-RealNet-18 · ImageNet valtop-1 accuracy: 56.4 -> 65.5Improved
replace 1-bit downsampling convs with group real-valued convsReActNet-A · ImageNet valtop-1 accuracy: 69.4 -> 70.1Improved
replace 1-bit downsampling convs with full real-valued convsReActNet-A · ImageNet valtop-1 accuracy: 69.4 -> 71.4Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy