NSTL国家科技图书文献中心

1. SSL-Cleanse: Trojan Detection and Mitigation in Self-Supervised Learning NSTL国家科技图书文献中心

Mengxin Zheng | Jiaqi Xue ... - 《Computer Vision - ECCV 2024,Part LXXXVII》 - European Conference on Computer Vision - 2025, - 405～421 - 共17页

摘要：Self-supervised learning (SSL) is a prevalent | -trained SSL* image encoder and subsequently training a* | . The growing adoption of SSL* has led to an increase* | in security research on SSL* encoders and associated* | Trojan attacks. Trojan attacks embedded in SSL* encoders*

关键词： Self-supervised learning | Backdoor attack | Detection

2. Event-VoxelHop: A Polarity-Wise Successive Subspace Learning (SSL) Algorithm for Event-Based Object Classification NSTL国家科技图书文献中心

Choi, Jaewon | Kuo, C. -C. Jay ... - 《IEEE sensors journal》 - 2025,25(2) - 3160～3172 - 共13页

摘要： (SSL). Unlike traditional vision systems that rely on | resolution and low-energy use. We apply an SSL* algorithm, a* | polarity and building upon the SSL* framework by* | In this study, we introduce Event-VoxelHop, an | innovative algorithm designed for object classification

关键词： Cameras | Event detection | Classification algorithms | Sensors | Streams | Feature extraction | Transforms | Robot vision systems | Accuracy | Training...

3. SSL-MBC: Self-Supervised Learning With Multibranch Consistency for Few-Shot PolSAR Image Classification NSTL国家科技图书文献中心

Wenmei Li | Hao Xia ... - 《IEEE journal of selected topics in applied earth observations and remote sensing》 - 2025,18 - 4696～4710 - 共15页

摘要：-supervised learning (SSL) is an effective solution for | SSL in PolSAR classification tasks is still a great | abovementioned issues, we propose an SSL* model with multibranch* | consistency (SSL-MBC) for few-shot PolSAR image | contrastive SSL* framework, without negative samples, is*

关键词： Scattering | Feature extraction | Matrix decomposition | Training | Image classification | Data mining | Data models | Self-supervised learning | Covariance matrices | Coherence

4. SSL Based Encoder Pretraining for Segmenting a Heterogeneous Chronic Wound Image Database with Few Annotations NSTL国家科技图书文献中心

Guillaume Picaud | Marc Chaumont ... - 《Diabetic Foot Ulcers Grand Challenge》 - Diabetic Foot Ulcers Grand Challenge | International Conference on Medical Image Computing and Computer Assisted Intervention - 2025, - 71～80 - 共10页

摘要： (SSL) strategies offer a promising solution to | a lighter one. We also explore the effect of SSL | DINO one. Furthermore, DINO SSL* pre-training on* | Segmentation is crucial in medical imaging | , but obtaining a sufficient quantity of annotated

关键词： Self supervised learning | Dinov2 | Segmentation | Chronic wounds

5. Dynamic Data Selection for Efficient SSL via Coarse-to-Fine Refinement NSTL国家科技图书文献中心

Aditay Tripathi | Pradeep Shenoy ... - 《Computer Vision - ECCV 2024,Part LVIII》 - European Conference on Computer Vision - 2025, - 377～393 - 共17页

摘要：Self-supervised learning (SSL) is critical for | compute requirements of SSL* have focused on identifying* | training dynamics of SSL* methods and adjusts the selected* | learning high-quality representations from unlabeled | images at scale. Earlier efforts at reducing the

关键词： Efficient training | Self supervised learning | Dynamic schedule

6. AsyncGBP+: Bridging SSL/TLS and Heterogeneous Computing Power With GPU-Based Providers NSTL国家科技图书文献中心

Yi Bian | Fangyu Zheng ... - 《IEEE Transactions on Computers》 - 2025,74(2) - 356～370 - 共15页

摘要： promising solution for accelerating the worldwide used SSL | GPUs to the serial mode of the SSL/TLS stack are | The rapid evolution of GPUs has emerged as a | /TLS, which faces performance bottlenecks due to its | underlying heavy cryptographic computations. Nevertheless

关键词： Cryptography | Graphics processing units | Protocols | Internet | Hardware acceleration | Digital signatures | Computer architecture | Standards | Fans | Elliptic curves

7. Preparation and characterization of the ground mixture of rebamipide commercial tablets and Hydroxypropyl Cellulose-SSL by ball-milling: Application to the dispersoid of mouthwash suspension NSTL国家科技图书文献中心

Mizobuchi, Senri | Hirose, Kaoru ... - 《European journal of pharmaceutics and biopharmaceutics》 - 2025,206 - 114584～114584 - 共10页

摘要：) tablets and hydroxypropyl cellulose (HPC-SSL) samples | properties of the ground mixture of HPC-SSL* and the powder* | formed with HPC-SSL* also avoided agglomeration on the* | . Furthermore, the ground mixture of HPC-SSL* and the powder* | -SSL) samples can be applied as a powdered suspension

关键词： Hydroxypropyl cellulose | Rebamipide | Commercial tablet | Ball-milling | Mouthwash | Oral mucositis | Suspension

8. Efficient diagnosis of retinal disorders using dual-branch semi-supervised learning (DB-SSL): An enhanced multi-class classification approach NSTL国家科技图书文献中心

Malik, Muhammad Hamm... | Wan, Zishuo ... - 《Computerized Medical Imaging and Graphics》 - 2025,121 - 102494～102494 - 共23页

摘要：-SSL) approach that leverages both labeled and | The early diagnosis of retinal disorders is | essential in preventing permanent or partial blindness | . Identifying these conditions promptly guarantees early | treatment and prevents blindness. However, the challenge

关键词： Retinal disease diagnoses | Deep learning | Semi-supervised learning | Multi-class classification | Color fundus photography

9. Diagnosing TBM Parameters to Understand Their Behavior while Identifying Soil Types by SSL During EPBM Excavation NSTL国家科技图书文献中心

Sharmin Sarna | Marte Gutierrez - 《Information technology in geo-engineering》 - International Conference on Information Technology in Geo-engineering - 2025, - 359～368 - 共10页

摘要：It is beneficial to know the soil type on the | go during excavation to ensure optimal Earth | pressure balance tunnel boring machine(EPBM | )operation.This encouraged many researchers to perform data | -driven analyses to develop prediction models.These

关键词： EPBM | Data-driven analysis | Semi-supervised learning

10. Advances in OpenASR21 Evaluation with Increased Temporal Resolution for Speech Self-supervised Learning Models NSTL国家科技图书文献中心

Vishwa Gupta - 《Speech and Computer,Part I》 - International Conference on Speech and Computer - 2025, - 69～81 - 共13页

摘要： can use any self supervised learning (SSL) model to | computed from the speech SSL* models from 20 ms to 10 ms* | . The temporal resolution of the SSL* models is in* | done without retraining the SSL* models. The resulting* | SSL models in the low resource OpenASR21 languages

摘要：Self-supervised learning (SSL) is a prevalent | -trained SSL* image encoder and subsequently training a* | . The growing adoption of SSL* has led to an increase* | in security research on SSL* encoders and associated* | Trojan attacks. Trojan attacks embedded in SSL* encoders*

关键词： Self-supervised learning | Backdoor attack | Detection

摘要： (SSL). Unlike traditional vision systems that rely on | resolution and low-energy use. We apply an SSL* algorithm, a* | polarity and building upon the SSL* framework by* | In this study, we introduce Event-VoxelHop, an | innovative algorithm designed for object classification

关键词： Cameras | Event detection | Classification algorithms | Sensors | Streams | Feature extraction | Transforms | Robot vision systems | Accuracy | Training...

摘要：-supervised learning (SSL) is an effective solution for | SSL in PolSAR classification tasks is still a great | abovementioned issues, we propose an SSL* model with multibranch* | consistency (SSL-MBC) for few-shot PolSAR image | contrastive SSL* framework, without negative samples, is*

关键词： Scattering | Feature extraction | Matrix decomposition | Training | Image classification | Data mining | Data models | Self-supervised learning | Covariance matrices | Coherence

摘要： (SSL) strategies offer a promising solution to | a lighter one. We also explore the effect of SSL | DINO one. Furthermore, DINO SSL* pre-training on* | Segmentation is crucial in medical imaging | , but obtaining a sufficient quantity of annotated

关键词： Self supervised learning | Dinov2 | Segmentation | Chronic wounds

摘要：Self-supervised learning (SSL) is critical for | compute requirements of SSL* have focused on identifying* | training dynamics of SSL* methods and adjusts the selected* | learning high-quality representations from unlabeled | images at scale. Earlier efforts at reducing the

关键词： Efficient training | Self supervised learning | Dynamic schedule

摘要： promising solution for accelerating the worldwide used SSL | GPUs to the serial mode of the SSL/TLS stack are | The rapid evolution of GPUs has emerged as a | /TLS, which faces performance bottlenecks due to its | underlying heavy cryptographic computations. Nevertheless

关键词： Cryptography | Graphics processing units | Protocols | Internet | Hardware acceleration | Digital signatures | Computer architecture | Standards | Fans | Elliptic curves

关键词： Hydroxypropyl cellulose | Rebamipide | Commercial tablet | Ball-milling | Mouthwash | Oral mucositis | Suspension

摘要：-SSL) approach that leverages both labeled and | The early diagnosis of retinal disorders is | essential in preventing permanent or partial blindness | . Identifying these conditions promptly guarantees early | treatment and prevents blindness. However, the challenge

关键词： Retinal disease diagnoses | Deep learning | Semi-supervised learning | Multi-class classification | Color fundus photography

摘要：It is beneficial to know the soil type on the | go during excavation to ensure optimal Earth | pressure balance tunnel boring machine(EPBM | )operation.This encouraged many researchers to perform data | -driven analyses to develop prediction models.These

关键词： EPBM | Data-driven analysis | Semi-supervised learning

摘要： can use any self supervised learning (SSL) model to | computed from the speech SSL* models from 20 ms to 10 ms* | . The temporal resolution of the SSL* models is in* | done without retraining the SSL* models. The resulting* | SSL models in the low resource OpenASR21 languages

关键词： OpenASR21 | Low-resource | Speech recognition | SSL models | Temporal resolution

4008-161-200 800-990-8900

国家科技图书文献中心

摘要：Self-supervised learning (SSL) is a prevalent | -trained SSL image encoder and subsequently training a | . The growing adoption of SSL has led to an increase | in security research on SSL encoders and associated | Trojan attacks. Trojan attacks embedded in SSL encoders

关键词： Self-supervised learning | Backdoor attack | Detection

摘要： (SSL). Unlike traditional vision systems that rely on | resolution and low-energy use. We apply an SSL algorithm, a | polarity and building upon the SSL framework by | In this study, we introduce Event-VoxelHop, an | innovative algorithm designed for object classification

关键词： Cameras | Event detection | Classification algorithms | Sensors | Streams | Feature extraction | Transforms | Robot vision systems | Accuracy | Training...

摘要：-supervised learning (SSL) is an effective solution for | SSL in PolSAR classification tasks is still a great | abovementioned issues, we propose an SSL model with multibranch | consistency (SSL-MBC) for few-shot PolSAR image | contrastive SSL framework, without negative samples, is

关键词： Scattering | Feature extraction | Matrix decomposition | Training | Image classification | Data mining | Data models | Self-supervised learning | Covariance matrices | Coherence

摘要： (SSL) strategies offer a promising solution to | a lighter one. We also explore the effect of SSL | DINO one. Furthermore, DINO SSL pre-training on | Segmentation is crucial in medical imaging | , but obtaining a sufficient quantity of annotated

关键词： Self supervised learning | Dinov2 | Segmentation | Chronic wounds

摘要：Self-supervised learning (SSL) is critical for | compute requirements of SSL have focused on identifying | training dynamics of SSL methods and adjusts the selected | learning high-quality representations from unlabeled | images at scale. Earlier efforts at reducing the

关键词： Efficient training | Self supervised learning | Dynamic schedule

摘要： promising solution for accelerating the worldwide used SSL | GPUs to the serial mode of the SSL/TLS stack are | The rapid evolution of GPUs has emerged as a | /TLS, which faces performance bottlenecks due to its | underlying heavy cryptographic computations. Nevertheless

关键词： Cryptography | Graphics processing units | Protocols | Internet | Hardware acceleration | Digital signatures | Computer architecture | Standards | Fans | Elliptic curves

摘要：) tablets and hydroxypropyl cellulose (HPC-SSL) samples | properties of the ground mixture of HPC-SSL and the powder | formed with HPC-SSL also avoided agglomeration on the | . Furthermore, the ground mixture of HPC-SSL and the powder | -SSL) samples can be applied as a powdered suspension

关键词： Hydroxypropyl cellulose | Rebamipide | Commercial tablet | Ball-milling | Mouthwash | Oral mucositis | Suspension

摘要：-SSL) approach that leverages both labeled and | The early diagnosis of retinal disorders is | essential in preventing permanent or partial blindness | . Identifying these conditions promptly guarantees early | treatment and prevents blindness. However, the challenge

关键词： Retinal disease diagnoses | Deep learning | Semi-supervised learning | Multi-class classification | Color fundus photography

摘要：It is beneficial to know the soil type on the | go during excavation to ensure optimal Earth | pressure balance tunnel boring machine(EPBM | )operation.This encouraged many researchers to perform data | -driven analyses to develop prediction models.These

关键词： EPBM | Data-driven analysis | Semi-supervised learning

摘要： can use any self supervised learning (SSL) model to | computed from the speech SSL models from 20 ms to 10 ms | . The temporal resolution of the SSL models is in | done without retraining the SSL models. The resulting | SSL models in the low resource OpenASR21 languages

关键词： OpenASR21 | Low-resource | Speech recognition | SSL models | Temporal resolution

4008-161-200 800-990-8900

国家科技图书文献中心

摘要：Self-supervised learning (SSL) is a prevalent | -trained SSL* image encoder and subsequently training a* | . The growing adoption of SSL* has led to an increase* | in security research on SSL* encoders and associated* | Trojan attacks. Trojan attacks embedded in SSL* encoders*

摘要： (SSL). Unlike traditional vision systems that rely on | resolution and low-energy use. We apply an SSL* algorithm, a* | polarity and building upon the SSL* framework by* | In this study, we introduce Event-VoxelHop, an | innovative algorithm designed for object classification

摘要：-supervised learning (SSL) is an effective solution for | SSL in PolSAR classification tasks is still a great | abovementioned issues, we propose an SSL* model with multibranch* | consistency (SSL-MBC) for few-shot PolSAR image | contrastive SSL* framework, without negative samples, is*

摘要： (SSL) strategies offer a promising solution to | a lighter one. We also explore the effect of SSL | DINO one. Furthermore, DINO SSL* pre-training on* | Segmentation is crucial in medical imaging | , but obtaining a sufficient quantity of annotated

摘要：Self-supervised learning (SSL) is critical for | compute requirements of SSL* have focused on identifying* | training dynamics of SSL* methods and adjusts the selected* | learning high-quality representations from unlabeled | images at scale. Earlier efforts at reducing the

摘要： can use any self supervised learning (SSL) model to | computed from the speech SSL* models from 20 ms to 10 ms* | . The temporal resolution of the SSL* models is in* | done without retraining the SSL* models. The resulting* | SSL models in the low resource OpenASR21 languages