MetaGAN: An Adversarial Approach to Few-Shot Learning

By Ruixiang Zhang, Tong Che, Zoubin Ghahramani, Yoshua Bengio, Yangqiu Song. Neural Information Processing Systems, 2018

451 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/75dafecd481f134d4f8b9afb4923b35184d3c971

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Few-Shot Learning With Global Class Representations (2019, unknown): “…randomly selected MLSTM [23] 43.44 ± 0.77 60.60 ± 0.71 MN [36] 43.56 ± 0.84 55.31 ± 0.73 MA [9] 48.70 ± 1.84 63.11 ± 0.92 PN [31] 49.42 ± 0.78 68.20 ± 0.66 DLM [35] 50.28 ± 0.80 63.70 ± 0.70 RN [31] 50.44 ± 0.82 65.32 ± 0.70 MG [38] 52.71 ± 0.64 68.63 ± 0.67 MMN [3] 53.37 The averaged accuracy (”
Semantics-Driven Generative Replay for Few-Shot Class Incremental Learning (2022, background): “Adversarial learning-based generative models like MetaGAN [43] and Few-shot Classifier GAN [1] have also been introduced for this purpose which can generate abundance of synthesized samples from the available few labeled data per class.”
Aligning Visual Prototypes with BERT Embeddings for Few-Shot Learning (2021, background): “Most few-shot learning methods can be divided into metric-based [17, 37, 43, 55] and meta-learning based [7, 25, 34] methods, although some other directions have also been explored, such as hallucination based [11, 51, 58] and parameter-generation based [9, 26] methods.”
PARN: Position-Aware Relation Networks for Few-Shot Learning (2019, methodology): “, learning a meta-optimizer [18] or taking some wise optimization strategies [6, 28, 29], to better and faster update the model for new tasks.”
Few-Shot Class Incremental Learning with Generative Feature Replay (2021, methodology): “Generative models are used for generating images samples of previous tasks and these synthetic images are replayed (Nguyen et al., 2017; Wu et al., 2018; Ostapenko et al., 2019; Zhang et al., 2018).”
Meta-Learning with Shared Amortized Variational Inference (2020, unknown): “, 2017; Ravi & Larochelle, 2017), adversarial approaches (Zhang et al., 2018), and Bayesian models (Gordon et al.”
Variational Few-Shot Learning (2019, background): “To alleviate it, GAN-based techs [46, 19, 47] exploit diverse training data patterns and apply it into test phase to expand the support set capacity.”
An Ensemble of Epoch-Wise Empirical Bayes for Few-Shot Learning (2019, background): “mples; (2) memory network methods [37, 46, 40, 35] aim to learn training “experience” from seen tasks and then aim to generalize to the learning of unseen ones; and (3) gradient descent based methods [11, 12, 2, 43, 29, 17, 63, 54] usually employing a meta-learner that learns to fast adapt a NN base”
Limited Data Rolling Bearing Fault Diagnosis With Few-Shot Learning (2019, background): “Zhang et al. [43] proposed a conceptually simple and general framework called MetaGAN for few-shot learning problems.”
DeepEMD: Few-Shot Image Classification With Differentiable Earth Mover’s Distance and Structured Classifiers (2020, unknown): “[4, 12, 14, 20, 27, 28, 37, 45, 49, 57, 60, 61, 65, 69–71,76, 77].”
Mind the Gap: Domain Gap Control for Single Shot Domain Adaptation for Generative Adversarial Networks (2021, background): “In the GAN domain, various models and training strategies have been proposed for few-shot domain adaptation tasks (Bousmalis et al., 2017; ZHANG et al., 2018; Li et al., 2020; Liu et al., 2019).”
Image-to-Image Translation: Methods and Applications (2021, background): “As noted by meta-learning [108], [109] and few-shot learning [110], [111], humans can effectively use prior experiences and knowledge when learning new tasks, while artificial”
Revisiting Meta-Learning as Supervised Learning (2020, background): “…(Riemer et al., 2019; Kaiser et al., 2017; Al-Shedivat et al., 2018; Clavera et al., 2019), transfer and multitask learning (Ying et al., 2018; Zhang et al., 2018b), active learning (Ravi and Larochelle, 2018; Sharma et al., 2018; Bachman et al., 2017; Ravi and Larochelle, 2018; Sharma et…”
Adversarial Feature Hallucination Networks for Few-Shot Learning (2020, methodology): “We speculate the reason is that this dataset is relatively small, less than 60 images per class DataAug MetaGAN [45] Conv-32F NeurIPS’18 on average; a large number of classes only have about 30 images.”
A Closer Look At Feature Space Data Augmentation For Few-Shot Intent Classification (2019, methodology): “For example, generative models are used for data augmentation in image classification (Mehrotra and Dukkipati, 2017; Antoniou et al., 2018; Zhang et al., 2018), text classification (Gupta, 2019), anomaly detection (Lim et al.”
Free Lunch for Few-shot Learning: Distribution Calibration (2021, background): “…use the idea of Generative Adversarial Networks (GANs) (Goodfellow et al., 2014) or autoencoder (Rumelhart et al., 1986) to generate samples (Zhang et al. (2018); Chen et al. (2019b); Schwartz et al. (2018); Gao et al. (2018)) or features (Xian et al. (2018); Zhang et al. (2019)) to augment…”
Counterfactual Vision-and-Language Navigation via Adversarial Path Sampling (2019, background): “Apart from leading the training target, adversarial training is also applied to data augmentation [5, 36].”