Domain Generalization for Object Recognition with Multi-task Autoencoders

By Muhammad Ghifary, W. Kleijn, Mengjie Zhang, David Balduzzi. 2015

719 citations recorded by the source corpus.

The problem of domain generalization is to take knowledge acquired from a number of related domains, where training data is available, and to then successfully apply it to previously unseen domains. We propose a new feature learning algorithm, Multi-Task Autoencoder (MTAE), that provides good generalization performance for cross-domain object recognition. The algorithm extends the standard denoising autoencoder framework by substituting artificially induced corruption with naturally occurring inter-domain variability in the appearance of objects. Instead of reconstructing images from noisy versions, MTAE learns to transform the original image into analogs in multiple related domains. It thereby learns features that are robust to variations across domains. The learnt features are then used as inputs to a classifier. We evaluated the performance of the algorithm on benchmark image recognition datasets, where the task is to learn features from multiple datasets and to then predict the image label from unseen datasets. We found that (denoising) MTAE outperforms alternative autoencoder-based models as well as the current state-of-the-art algorithms for domain generalization.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging, Representation & Self-Supervised Learning

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

In Representation & Self-Supervised Learning

Citations in the corpus

Learning Robust Global Representations by Penalizing Local Predictive Power (2019, methodology): “More recent DG approaches incorporate adversarial (or similar) techniques (Ghifary et al., 2015; Wang et al., 2016; Motiian et al., 2017b; Li et al., 2018a; Carlucci et al., 2018), or build ensembles of per-domain models that are then fused representations together (Bousmalis et al.”
Multiple-Source Adaptation with Domain Classifiers (2020, background): “Ghifary et al. [2015] extended auto-encoders to the multi-task setting and minimized the sum of reconstruction errors across domains.”
Learning Robust Representations by Projecting Superficial Statistics Out (2018, methodology): “We compared the performance of HEX/ADV with several methods tested on this data including CAE (Rifai et al., 2011), MTAE (Ghifary et al., 2015), CCSA (Motiian et al., 2017), DANN (Ganin et al., 2016), Fusion (Mancini et al., 2018), LabelGrad, and CrossGrad (Shankar et al., 2018).”
Domain Generalization via Conditional Invariant Representations (2018, methodology): “Domain generalization has been widely applied in classi-ﬁcation tasks (Xu et al. 2014; Duan et al. 2009; Muan-det, Balduzzi, and Sch ¨ olkopf 2013; Ghifary et al. 2017; Ghifary et al. 2015; Erfani et al. 2016).”
Learning to Generalize: Meta-Learning for Domain Generalization (2017, neutral): “The few existing DG methods typically train on multiple source domains and propose mechanisms to extract some domain agnostic representation or model that describes common aspects of known domains (Khosla et al. 2012; Muandet, Balduzzi, and Schölkopf 2013; Ghifary et al. 2015; Li et al. 2017).”
In Search of Lost Domain Generalization (2020, methodology): “Using all available data In Rotated MNIST, whereas the usual version of the dataset constructs all domains from the same set of 1000 digits, we divide all the MNIST digits evenly among domains.”
Deeper, Broader and Artier Domain Generalization (2017, positive): “MTAE-1HNN: 1HNN with multi-task auto encoder [10].”
Unified Deep Supervised Domain Adaptation and Generalization (2017, methodology): “We followed the setting in [24], and randomly selected a set M of 100 images per category from the MNIST dataset (1000 in total).”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “arge set of unlabeled source data to discover repetitive patterns that could be used for the supervised learning task of interest. On the other hand, the relation between DA and domain generalization [151, 254, 97, 94, 155], multi-task learning [26, 76, 178] or few-shot learning [149, 80] is much st”
Domain Generalization by Solving Jigsaw Puzzles (2019, methodology): “This was formalized with the use of deep learning autoencoders in [20, 27], while [33] proposed to learn an embedding space where images of same classes but different sources are projected nearby.”
Test-Time Training for Out-of-Distribution Generalization (2019, background): “…multiple environment distributions, some of which are available during training (source), while others are used for testing (target) (Li et al., 2018; Shankar et al., 2018; Muandet et al., 2013; Balaji et al., 2018; Ghifary et al., 2015; Motiian et al., 2017; Li et al., 2017a; Gan et al., 2016).”
Deep Visual Domain Adaptation: A Survey (2018, methodology): “Reconstructionbased using the data reconstruction as an auxiliary task to ensure feature invariance encoder-decoder reconstruction [4, 26, 24, 114] adversarial reconstruction [105, 113, 50]”
FSDR: Frequency Space Domain Randomization for Domain Generalization (2021, methodology): “Most existing DG methods can be broadly categorized into multi-source DG [44, 13, 12, 34, 16, 33, 43, 42, 59, 36, 2, 10, 35] and single-source DG [64, 70, 77, 50] both of which strive to learn domainagnostic features by either learning a domain-invariant feature space [44, 13, 12, 16, 34, 59, 36, 10”
Deep Stable Learning for Out-Of-Distribution Generalization (2021, background): “Recently, such distribution (domain) shift problems have been intensively studied in the domain generalization (DG) literature [41, 17, 25, 62, 31, 33].”
SelfReg: Self-supervised Contrastive Regularization for Domain Generalization (2021, methodology): “This testbed currently provides seven multi-domain datasets (i.e. ColoredMNIST [1], RotatedMNIST [16], VLCS [13], PACS [26], Ofﬁce-Home [42], and TerraIncognita [2], DomainNet [34]) and provides benchmarks results of 14 baseline approaches (i.e. ERM [41], IRM [1], GroupDRO [36], Mixup [45], MLDG [27”