CIFAR-100
PulseAugur coverage of CIFAR-100 — every cluster mentioning CIFAR-100 across labs, papers, and developer communities, ranked by signal.
- used by ResNet-18 90%
- instance of Class Incremental Learning 90%
- instance of Tiny-ImageNet 70%
- instance of ResNet-18 70%
- used by Imagenet 1k 70%
- used by ImageNet-100 70%
- instance of residual neural network 70%
- instance of ImageNet-100 70%
- instance of Imagenet 1k 70%
- instance of TinyImageNet 70%
- instance of The Street View House Numbers Dataset 70%
- used by Vision Transformers 70%
12 day(s) with sentiment data
-
New watermarking scheme TwinMark protects AI models from distillation attacks
Researchers have developed TwinMark, a novel watermarking technique designed to protect AI models against distillation attacks. This method uses two complementary linear functionals, one based on feature covariance and …
-
Temperon training method achieves SAM quality with reduced cost
Researchers have introduced Temperon, a novel training method designed to achieve the quality of Sharpness-Aware Minimization (SAM) while significantly reducing computational costs. Temperon utilizes a two-phase approac…
-
LePoKet framework enhances robotic vision with learnable knowledge transfer
Researchers have developed LePoKet, a novel framework for knowledge transfer in robotic vision systems. This method optimizes interaction parameters within a block-wise interface, enabling learnable parameter optimizati…
-
New BARGE method tackles imbalanced learning with noisy labels
Researchers have developed a new method called BARGE (Bounded Adjustment with Reliability-Guided Embeddings) to address challenges in imbalanced learning with noisy labels. This single-stage objective combines a bounded…
-
New continual learning methods tackle catastrophic forgetting in AI models · 4 sources tracked
Researchers are developing new methods to combat catastrophic forgetting in continual learning, a challenge where AI models lose previously acquired knowledge when learning new tasks. One approach, Multiple Embedding Re…
-
New theory refines spectral representation learning, emphasizes task diversity
Researchers have developed a new theoretical framework for spectral representation learning, challenging the assumption of isotropy in self-supervised learning. The study demonstrates that task diversity, rather than sy…
-
ZAPS pipeline enhances Neural Architecture Search by combining proxy signals and topology
Researchers have developed ZAPS, a novel four-stage pipeline designed to improve Neural Architecture Search (NAS) by efficiently combining proxy signals with architectural topology. This method addresses the limitations…
-
New AR-KD method enhances AI model knowledge transfer
Researchers have introduced Adaptive Reciprocal Knowledge Distillation (AR-KD), a novel method designed to enhance knowledge transfer from large teacher models to smaller student models. Unlike traditional one-way disti…
-
New machine unlearning method achieves 82x speedup
Researchers have developed a novel machine unlearning framework that significantly speeds up the process of removing specific data points from trained models. This method identifies correlated data points and uses a clo…
-
New methods accelerate Neural Architecture Search with reduced computational cost · 3 sources tracked
Researchers have developed new methods to improve Neural Architecture Search (NAS), a process that can be computationally expensive. One approach, RiPPLE, uses partial training data from a small set of anchor architectu…
-
New method enhances decentralized federated distillation with multi-modality knowledge collaboration
This paper introduces a novel decentralized federated distillation method designed for clients with heterogeneous models. The approach leverages shared unlabeled public data for collaboration, where each client evaluate…
-
FANS framework optimizes model architectures for heterogeneous federated learning
Researchers have developed FANS (Federated Adaptive Network Search), a new framework designed to optimize model architectures in heterogeneous federated learning environments. This approach utilizes a hypernetwork to le…
-
New Oscillatory Predictive Learning framework shows emergent adversarial robustness
Researchers have developed a new framework called Oscillatory Predictive Learning (OPL) that aims to achieve adversarial robustness in computer vision without relying on traditional methods like adversarial training or …
-
Phase transition frequency predicts ResNet accuracy in training
Researchers have identified a new metric, "phase transition frequency," that can predict the test accuracy of ResNet models during training. This metric, which counts discrete class-separability jumps, showed a strong n…
-
New method enhances graph construction for machine learning tasks
Researchers have developed a novel method for constructing graphs in kernelized graph methods, specifically addressing the challenge of selecting an appropriate Gaussian bandwidth (sigma). The proposed approach uses a p…
-
New Temperature Scaling Attack targets model confidence in federated learning
Researchers have developed a new training-time attack called the Temperature Scaling Attack (TSA) that specifically targets the confidence calibration of models in federated learning systems. This attack degrades a mode…
-
AI research reveals distillation bottleneck, label-aware methods improve performance
Researchers have identified a significant geometric bottleneck in knowledge distillation between Vision Transformers and smaller CNNs. Standard cosine distillation causes the learned representations to collapse to a low…
-
New Threat Conditional Network offers unified adversarial robustness
Researchers have introduced the Threat Conditional Network (TCN), a novel approach to achieving robust performance against adversarial attacks across a wide range of threat levels within a single model. TCN utilizes a r…
-
New MFSPNet method slashes neural architecture search costs
Researchers have developed MFSPNet, a novel model-free surrogate-assisted neural architecture search method designed to reduce the computational cost of designing deep neural networks. This approach integrates a lightwe…
-
New CRAD method enhances decentralized federated learning with reliability-aware distillation
Researchers have developed a new method called Class-wise Reliability-Aware Distillation (CRAD) for decentralized federated learning. This approach allows clients to use different model architectures and does not requir…