Label Efficient Learning of Transferable Representations acrosss Domains and Tasks

By Zelun Luo, Yuliang Zou, Judy Hoffman, Li Fei-Fei. 2017

290 citations recorded by the source corpus.

We propose a framework that learns a representation transferable across different domains and tasks in a label efficient manner. Our approach battles domain shift with a domain adversarial loss, and generalizes the embedding to novel task using a metric learning-based approach. Our model is simultaneously optimized on labeled source data and unlabeled or sparsely labeled data in the target domain. Our method shows compelling results on novel classes within a new domain even when only a few labeled examples per class are available, outperforming the prevalent fine-tuning approach. In addition, we demonstrate the effectiveness of our framework on the transfer learning task from image object recognition to video action recognition.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

GAN Memory with No Forgetting (2020, methodology): “Distinct from existing methods, our GAN memory leverages transfer learning [6, 16, 46, 89, 57] and (image) style transfer [18, 30, 41].”
Taskonomy: Disentangling Task Transfer Learning (2018, background): “compositional modeling [35, 10, 13, 23, 55, 92, 90], homomorphic cryptography [42], lifelong learning [93, 15, 85, 84], functional maps [71], certain aspects of Bayesian inference and Dirichlet processes [54, 91, 90, 89, 37, 39], few-shot learning [81, 25, 24, 70, 86], transfer learning [75, 84, 29,”
A Survey on Deep Transfer Learning (2018, methodology): “[16] utilize a domain adversarial loss, and generalizes the embedding to novel task using a metric learning-based approach to find more tractable features in deep transfer learning.”
Few-Shot Object Detection via Feature Reweighting (2018, background): “[23] consider the problem of adapting to novel classes in a new domain.”
Class-Incremental Domain Adaptation (2020, methodology): “We compare against prior UDA methods DANN [11], OSBP [45], UAN [54], STA [29], and CI methods E2E [4], LETR [33], iCaRL [41], LwF-MC [7], LwM [7].”
Generalizing from a Few Examples (2019, methodology): “Transfer learning methods are popularly used in FSL [7, 82, 85], where the prior knowledge is transferred from the source task to the few-shot task.”
Which Tasks Should Be Learned Together in Multi-task Learning? (2019, background): “Transfer learning, [34, 16, 40, 15, 25, 28, 21, 35, 31], is similar to multi-task learning in that solutions are learned for multiple tasks.”
Learning Semantic Representations for Unsupervised Domain Adaptation (2018, background): “In recent years, few-shot adversarial learning (Tzeng et al., 2015; Motiian et al., 2017a; Luo et al., 2017) have been explored in domain adaptation.”
Boundary-weighted Domain Adaptive Neural Network for Prostate MR Image Segmentation (2019, background): “through minimizing an adversarial loss with respect to a domain discriminator [26], [48].”
On Leveraging Pretrained GANs for Generation with Limited Data (2020, background): “, to reuse the feature extractor of a pretrained classifier) (Bengio, 2012; Donahue et al., 2014; Luo et al., 2017; Zamir et al., 2018), and transfer learning in natural language processing (e.”
MetaSearch: Incremental Product Search via Deep Meta-Learning (2020, background): “Luo et al. [31] considered the problem of adapting to novel categories in a new domain and introduced an adaptive few-shot learning method that achieved good performance.”
A Review of Single-Source Deep Unsupervised Visual Domain Adaptation (2020, background): “It belongs to the sample-efficient learning class [10],”
A Survey on Few-Shot Class-Incremental Learning (2023, background): “As a result, it has broad applications in zero-shot and few-shot domains, including style transfer, feature space transfer for data augmentation, and label-e ffi cient learning of transferable representations across domains (Azadi et al., 2018; Liu et al., 2018; Luo et al., 2017).”
Transferability vs. Discriminability: Batch Spectral Penalization for Adversarial Domain Adaptation (2019, methodology): “…adversarial learning, where a domain discriminator is trained to distinguish the source from the target while feature representations are learned to confuse it simultaneously (Ganin & Lempitsky, 2015; Tzeng et al., 2015; Ganin et al., 2016; Tzeng et al., 2017; Luo et al., 2017; Long et al., 2018).”
Discriminative Partial Domain Adversarial Network (2020, background): “Some help feature extractor to learn domain-invariant features in a min-max game by adding a discriminator [5, 17,25,26].”
Transferable Representation Learning with Deep Adaptation Networks (2019, background): “In particular, adversarial domain adaptation [29], [36], [37], [38], [39] combines adversarial learning and domain adaptation, which performs strongly on many datasets.”
Heterogeneous Graph Attention Network for Unsupervised Multiple-Target Domain Adaptation (2020, background): “FADA [11] pairs the labeled target sample and labeled source sample, and the discriminator is trained to tell whether the pair comes from the same domain and same class, and the cross-category similarity is proposed for semantic transfer [24].”