Transformers Can Reuse Hidden Token Representations Without Retraining
A new inference-time framework reuses highly similar hidden token states in transformers, preserving sequence length without retraining while largely maintaining ...
Tag: model efficiency
Federated decision transformer sharpens simulated industrial scheduling
A federated Decision Transformer reached 92.5% scheduling accuracy in a simulated industrial environment while reducing parameter transmission by 63% relative ...
Lightweight AI Network Brings Accurate Skin-Lesion Segmentation to Edge Devices
Researchers have developed LMANet, a lightweight multiscale attention network that achieves high-accuracy skin-lesion segmentation with only 5.6 million parameters by ...
Tiny AI Model Reads Tumor Shapes to Map Brain Cancer With a Fraction of the Parameters
A new lightweight AI framework called RAML embeds tumor subregion identity and geometric descriptors directly into its architecture, trading a ...
Attention-Boosted Neural Networks Face Off in the Hunt for Deepfakes
A controlled comparison shows attention-enhanced DenseNet121 reaching 97.4 percent accuracy for deepfake detection while a lighter MobileNetV2 variant offers a ...
New Self-Attention Method Strips Redundant Filters From CNNs to Boost Plant Disease Detection
Researchers have developed CNN-SA-RFR, a cosine similarity-based self-attention framework that suppresses redundant filters during CNN training, boosting plant disease classification ...
-
Mothers who receive childcare support from maternal grandparents show more parental warmth, finds NTU Singapore study
-
University of Seville Breaks 120-Year-Old Mystery, Revises a Key Einstein Concept
-
Bee body mass, pathogens and local climate influence heat tolerance
-
Researchers record first-ever images and data of a shark experiencing a boat strike
-
Groundbreaking Clinical Trial Reveals Lubiprostone Enhances Kidney Function
Embark on a thrilling journey of discovery with Scienmag.com—your ultimate source for cutting-edge breakthroughs. Immerse yourself in a world where curiosity knows no limits and tomorrow’s possibilities become today’s reality!