AI Paper Atlas

Research map / Architectures and efficiency / Quantization

SemanticNN: Compressive and Error-Resilient Semantic Offloading for Extremely Weak Devices

Jiaming Huang, Yi Gao, Fuchang Pan, Renjie Li, Wei Dong

AAAI 2026 · Architectures and efficiency › Quantization · approach: Binary neural networks · 0 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
use SemanticNN instead of offloading raw featuressplit ResNet-50 · ImageNet-200 with transmission errorstop-1 accuracy (%): 32.84 -> 62.93Improved
use SemanticNN instead of offloading raw imagessplit ResNet-50 · CIFAR-100 with transmission errorstop-1 accuracy (%): 45.43 -> 53.96Improved
use SemanticNN instead of offloading raw featuresYOLOv3 · VHR-10 with transmission errorsmAP: 0.68 -> 0.81Improved
use SemanticNN instead of offloading raw featuresYOLOv3 · VOC2007 with transmission errorsmAP: 0.81 -> 0.80No clear effect
use SemanticNN instead of DeepCOD* and AgileNN* compressed offloadingsplit ResNet-50 / MobileNetV2 · image classification with transmission errorstop-1 accuracy: +7.11% to 30.88%Improved
use CNN-based SemanticNN instead of Transformer-based DeepSCsplit ResNet-50 · ImageNet-200transmitted data (KB): 98 -> 24.5Improved
use CNN-based SemanticNN instead of Transformer-based DeepSCsplit ResNet-50 · ImageNet-200parameters (KB): 9165 -> 1875Improved
add two-stage Feature-augmentation Learning (denoising autoencoder pretraining) before task training (ablation)SemanticNN · image classificationaccuracyImproved
add attention module in decoder for dynamic BER (ablation)SemanticNN · varying BERsaccuracyImproved
set distribution loss weight alpha to 2 (ablation)SemanticNN · quantization center balanceentropy of center frequencyImproved
set XAI loss weight gamma to 1 (ablation)SemanticNN · XAI losslossImproved
use symmetric encoder-decoder instead of asymmetric (E1D3, E1D5) (ablation)ADJSCC · ResNet-50 offloadingaccuracyImproved

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