Statements (55)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:Cloud_Computing_Service
|
| gptkbp:architecturalStyle |
Deep Learning
|
| gptkbp:completed |
Top-5 error rate of 15.3%
|
| gptkbp:contributedTo |
Popularized deep learning in computer vision
|
| gptkbp:dataUsage |
gptkb:ImageNet
|
| gptkbp:developedBy |
gptkb:Geoffrey_Hinton
|
| gptkbp:has |
3 fully connected layers
5 convolutional layers a large number of training images multiple filters per layer overlapping pooling layers |
| https://www.w3.org/2000/01/rdf-schema#label |
AlexNet
|
| gptkbp:influencedBy |
LeNet
|
| gptkbp:inputOutput |
227x227 pixels
|
| gptkbp:inspiration |
gptkb:ResNet
gptkb:GoogleNet Inception model |
| gptkbp:is |
a popular research topic
used in various applications a benchmark for image classification a foundational model in deep learning a model for classification tasks a model for deep feature learning a model for feature extraction a model for fine-tuning a model for image recognition a model for large-scale image classification a model for object detection a model for visual recognition tasks a model that has been widely cited a model that has influenced many architectures a model that uses backpropagation a model that uses batch normalization a significant advancement in neural networks widely used in transfer learning a model that has set new standards in image classification. a_key_model_in_the_resurgence_of_AI |
| gptkbp:isUsedFor |
Caffe
|
| gptkbp:keyIssues |
60 million
|
| gptkbp:numberOfStudents |
8
|
| gptkbp:operates |
gptkb:C++
Python |
| gptkbp:screenSize |
1000 classes
|
| gptkbp:successor |
VGGNet
|
| gptkbp:training |
5-6 days on two GPUs
|
| gptkbp:uses |
Stochastic Gradient Descent
Data augmentation ReLU activation function Local Response Normalization Image preprocessing Dropout regularization GPU_acceleration Max_pooling |
| gptkbp:winner |
gptkb:ImageNet_Large_Scale_Visual_Recognition_Challenge
|
| gptkbp:yearEstablished |
2012
|