Zero-Shot Learning
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
- 하위 태스크
- 7
- 평가표
- 33
- 보존된 평가 기록
- 84
Zero-shot learning (ZSL) is a model's ability to detect classes never seen during training. The condition is that the classes are not known during supervised learning.
Earlier work in zero-shot learning use attributes in a two-step approach to infer unknown classes. In the computer vision context, more recent advances learn mappings from image feature space to semantic space. Other approaches learn non-linear multimodal embeddings. In the modern NLP context, language models can be evaluated on downstream tasks without fine tuning.
Benchmark datasets for zero-shot learning include aPY, AwA, and CUB, among others.
( Image credit: Prototypical Networks for Few shot Learning in PyTorch )
Further readings:
논문
이 태스크와 하위 분야의 관련 논문입니다. 실제 성능 비교는 평가 결과에서 확인하세요.
Motion Feature Network: Fixed Motion Filter for Action Recognition
Learning Class Prototypes via Structure Alignment for Zero-Shot Recognition
Deep Discriminative Model for Video Classification
Correlation Net: Spatiotemporal multimodal deep learning for action recognition
AutoLoc: Weakly-supervised Temporal Action Localization
Towards Privacy-Preserving Visual Recognition via Adversarial Training: A Pilot Study
Few-Shot Adaptation for Multimedia Semantic Indexing
Selective Zero-Shot Classification with Augmented Attributes
Skeletal Movement to Color Map: A Novel Representation for 3D Action Recognition with Inception Residual Networks
Signal Alignment for Humanoid Skeletons via the Globally Optimal Reparameterization Algorithm
ENG: End-to-end Neural Geometry for Robust Depth and Pose Estimation using CNNs
The Globally Optimal Reparameterization Algorithm: an Alternative to Fast Dynamic Time Warping for Action Recognition in Video Sequences
Adding Attentiveness to the Neurons in Recurrent Neural Networks
CTAP: Complementary Temporal Action Proposal Generation
Video-based Person Re-identification via 3D Convolutional Networks and Non-local Attention
Geometric Generalization Based Zero-Shot Learning Dataset Infinite World: Simple Yet Powerful
Spatio-Temporal Instance Learning: Action Tubes from Class Supervision
A Variational Time Series Feature Extractor for Action Prediction
3D Human Action Recognition with Siamese-LSTM Based Deep Metric Learning
A Survey of Knowledge Representation in Service Robotics
Cooperative Learning of Audio and Video Models from Self-Supervised Synchronization
Action Recognition for Depth Video using Multi-view Dynamic Images
A Novel Geometric Framework on Gram Matrix Trajectories for Human Behavior Understanding
YH Technologies at ActivityNet Challenge 2018
Modeling Spatio-Temporal Human Track Structure for Action Localization
Human Action Recognition and Prediction: A Survey
Human Activity Prediction in Smart Home Environments with LSTM Neural Networks
Exploiting Spatial-Temporal Modelling and Multi-Modal Fusion for Human Action Recognition
A Transferable Pedestrian Motion Prediction Model for Intersections with Different Geometries
Context-Aware Pedestrian Motion Prediction In Urban Intersections
Tracking Emerges by Colorizing Videos
CNN-based Action Recognition and Supervised Domain Adaptation on 3D Body Skeletons via Kernel Feature Maps
Global Semantic Consistency for Zero-Shot Learning
Learning Multimodal Representations for Unseen Activities
Complex Gated Recurrent Neural Networks
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.