Assaying Out-Of-Distribution Generalization in Transfer Learning
- URL: http://arxiv.org/abs/2207.09239v1
- Date: Tue, 19 Jul 2022 12:52:33 GMT
- Title: Assaying Out-Of-Distribution Generalization in Transfer Learning
- Authors: Florian Wenzel, Andrea Dittadi, Peter Vincent Gehler, Carl-Johann
Simon-Gabriel, Max Horn, Dominik Zietlow, David Kernert, Chris Russell,
Thomas Brox, Bernt Schiele, Bernhard Sch\"olkopf, Francesco Locatello
- Abstract summary: We take a unified view of previous work, highlighting message discrepancies that we address empirically.
We fine-tune over 31k networks, from nine different architectures in the many- and few-shot setting.
- Score: 103.57862972967273
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Since out-of-distribution generalization is a generally ill-posed problem,
various proxy targets (e.g., calibration, adversarial robustness, algorithmic
corruptions, invariance across shifts) were studied across different research
programs resulting in different recommendations. While sharing the same
aspirational goal, these approaches have never been tested under the same
experimental conditions on real data. In this paper, we take a unified view of
previous work, highlighting message discrepancies that we address empirically,
and providing recommendations on how to measure the robustness of a model and
how to improve it. To this end, we collect 172 publicly available dataset pairs
for training and out-of-distribution evaluation of accuracy, calibration error,
adversarial attacks, environment invariance, and synthetic corruptions. We
fine-tune over 31k networks, from nine different architectures in the many- and
few-shot setting. Our findings confirm that in- and out-of-distribution
accuracies tend to increase jointly, but show that their relation is largely
dataset-dependent, and in general more nuanced and more complex than posited by
previous, smaller scale studies.
Related papers
- Distributed, communication-efficient, and differentially private estimation of KL divergence [15.294136011320433]
Key task in managing distributed, sensitive data is to measure the extent to which a distribution changes.
We describe novel algorithmic approaches for estimating the KL divergence of data across federated models of computation, under differential privacy.
arXiv Detail & Related papers (2024-11-25T15:20:40Z) - Generalizing to any diverse distribution: uniformity, gentle finetuning and rebalancing [55.791818510796645]
We aim to develop models that generalize well to any diverse test distribution, even if the latter deviates significantly from the training data.
Various approaches like domain adaptation, domain generalization, and robust optimization attempt to address the out-of-distribution challenge.
We adopt a more conservative perspective by accounting for the worst-case error across all sufficiently diverse test distributions within a known domain.
arXiv Detail & Related papers (2024-10-08T12:26:48Z) - Deriving Causal Order from Single-Variable Interventions: Guarantees & Algorithm [14.980926991441345]
We show that datasets containing interventional data can be effectively extracted under realistic assumptions about the data distribution.
We introduce interventional faithfulness, which relies on comparisons between the marginal distributions of each variable across observational and interventional settings.
We also introduce Intersort, an algorithm designed to infer the causal order from datasets containing large numbers of single-variable interventions.
arXiv Detail & Related papers (2024-05-28T16:07:17Z) - Measuring Adversarial Datasets [28.221635644616523]
Researchers have curated various adversarial datasets for capturing model deficiencies that cannot be revealed in standard benchmark datasets.
There is still no methodology to measure the intended and unintended consequences of those adversarial transformations.
We conducted a systematic survey of existing quantifiable metrics that describe text instances in NLP tasks.
arXiv Detail & Related papers (2023-11-06T22:08:16Z) - Robustness May be More Brittle than We Think under Different Degrees of
Distribution Shifts [72.90906474654594]
We show that robustness of models can be quite brittle and inconsistent under different degrees of distribution shifts.
We observe that large-scale pre-trained models, such as CLIP, are sensitive to even minute distribution shifts of novel downstream tasks.
arXiv Detail & Related papers (2023-10-10T13:39:18Z) - Multiply Robust Federated Estimation of Targeted Average Treatment
Effects [0.0]
We propose a novel approach to derive valid causal inferences for a target population using multi-site data.
Our methodology incorporates transfer learning to estimate ensemble weights to combine information from source sites.
arXiv Detail & Related papers (2023-09-22T03:15:08Z) - Squared $\ell_2$ Norm as Consistency Loss for Leveraging Augmented Data
to Learn Robust and Invariant Representations [76.85274970052762]
Regularizing distance between embeddings/representations of original samples and augmented counterparts is a popular technique for improving robustness of neural networks.
In this paper, we explore these various regularization choices, seeking to provide a general understanding of how we should regularize the embeddings.
We show that the generic approach we identified (squared $ell$ regularized augmentation) outperforms several recent methods, which are each specially designed for one task.
arXiv Detail & Related papers (2020-11-25T22:40:09Z) - On the Benefits of Invariance in Neural Networks [56.362579457990094]
We show that training with data augmentation leads to better estimates of risk and thereof gradients, and we provide a PAC-Bayes generalization bound for models trained with data augmentation.
We also show that compared to data augmentation, feature averaging reduces generalization error when used with convex losses, and tightens PAC-Bayes bounds.
arXiv Detail & Related papers (2020-05-01T02:08:58Z) - When Relation Networks meet GANs: Relation GANs with Triplet Loss [110.7572918636599]
Training stability is still a lingering concern of generative adversarial networks (GANs)
In this paper, we explore a relation network architecture for the discriminator and design a triplet loss which performs better generalization and stability.
Experiments on benchmark datasets show that the proposed relation discriminator and new loss can provide significant improvement on variable vision tasks.
arXiv Detail & Related papers (2020-02-24T11:35:28Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.