Unsupervised Disentanglement without Autoencoding: Pitfalls and Future
Directions
- URL: http://arxiv.org/abs/2108.06613v1
- Date: Sat, 14 Aug 2021 21:06:42 GMT
- Title: Unsupervised Disentanglement without Autoencoding: Pitfalls and Future
Directions
- Authors: Andrea Burns, Aaron Sarna, Dilip Krishnan, Aaron Maschinot
- Abstract summary: Disentangled visual representations have largely been studied with generative models such as Variational AutoEncoders (VAEs)
We explore regularization methods with contrastive learning, which could result in disentangled representations powerful enough for large scale datasets and downstream applications.
We evaluate disentanglement with downstream tasks, analyze the benefits and disadvantages of each regularization used, and discuss future directions.
- Score: 21.035001142156464
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Disentangled visual representations have largely been studied with generative
models such as Variational AutoEncoders (VAEs). While prior work has focused on
generative methods for disentangled representation learning, these approaches
do not scale to large datasets due to current limitations of generative models.
Instead, we explore regularization methods with contrastive learning, which
could result in disentangled representations that are powerful enough for large
scale datasets and downstream applications. However, we find that unsupervised
disentanglement is difficult to achieve due to optimization and initialization
sensitivity, with trade-offs in task performance. We evaluate disentanglement
with downstream tasks, analyze the benefits and disadvantages of each
regularization used, and discuss future directions.
Related papers
- Unsupervised Model Diagnosis [49.36194740479798]
This paper proposes Unsupervised Model Diagnosis (UMO) to produce semantic counterfactual explanations without any user guidance.
Our approach identifies and visualizes changes in semantics, and then matches these changes to attributes from wide-ranging text sources.
arXiv Detail & Related papers (2024-10-08T17:59:03Z) - Complementary Learning for Real-World Model Failure Detection [15.779651238128562]
We introduce complementary learning, where we use learned characteristics from different training paradigms to detect model errors.
We demonstrate our approach by learning semantic and predictive motion labels in point clouds in a supervised and self-supervised manner.
We perform a large-scale qualitative analysis and present LidarCODA, the first dataset with labeled anomalies in lidar point clouds.
arXiv Detail & Related papers (2024-07-19T13:36:35Z) - Hub-VAE: Unsupervised Hub-based Regularization of Variational
Autoencoders [11.252245456934348]
We propose an unsupervised, data-driven regularization of the latent space with a mixture of hub-based priors and a hub-based contrastive loss.
Our algorithm achieves superior cluster separability in the embedding space, and accurate data reconstruction and generation.
arXiv Detail & Related papers (2022-11-18T19:12:15Z) - Self-Supervised Learning via Maximum Entropy Coding [57.56570417545023]
We propose Maximum Entropy Coding (MEC) as a principled objective that explicitly optimize on the structure of the representation.
MEC learns a more generalizable representation than previous methods based on specific pretext tasks.
It achieves state-of-the-art performance consistently on various downstream tasks, including not only ImageNet linear probe, but also semi-supervised classification, object detection, instance segmentation, and object tracking.
arXiv Detail & Related papers (2022-10-20T17:58:30Z) - Evaluating the Label Efficiency of Contrastive Self-Supervised Learning
for Multi-Resolution Satellite Imagery [0.0]
Self-supervised learning has been applied in the remote sensing domain to exploit readily-available unlabeled data.
In this paper, we study self-supervised visual representation learning through the lens of label efficiency.
arXiv Detail & Related papers (2022-10-13T06:54:13Z) - A Novel Plug-and-Play Approach for Adversarially Robust Generalization [26.29269757430314]
We propose a robust framework that employs adversarially robust training to safeguard the machine learning models against perturbed testing data.
We achieve this by incorporating the worst-case additive adversarial error within a fixed budget for each sample during model estimation.
arXiv Detail & Related papers (2022-08-19T17:02:55Z) - DEALIO: Data-Efficient Adversarial Learning for Imitation from
Observation [57.358212277226315]
In imitation learning from observation IfO, a learning agent seeks to imitate a demonstrating agent using only observations of the demonstrated behavior without access to the control signals generated by the demonstrator.
Recent methods based on adversarial imitation learning have led to state-of-the-art performance on IfO problems, but they typically suffer from high sample complexity due to a reliance on data-inefficient, model-free reinforcement learning algorithms.
This issue makes them impractical to deploy in real-world settings, where gathering samples can incur high costs in terms of time, energy, and risk.
We propose a more data-efficient IfO algorithm
arXiv Detail & Related papers (2021-03-31T23:46:32Z) - Adversarial Examples for Unsupervised Machine Learning Models [71.81480647638529]
Adrial examples causing evasive predictions are widely used to evaluate and improve the robustness of machine learning models.
We propose a framework of generating adversarial examples for unsupervised models and demonstrate novel applications to data augmentation.
arXiv Detail & Related papers (2021-03-02T17:47:58Z) - MAIN: Multihead-Attention Imputation Networks [4.427447378048202]
We propose a novel mechanism based on multi-head attention which can be applied effortlessly in any model.
Our method inductively models patterns of missingness in the input data in order to increase the performance of the downstream task.
arXiv Detail & Related papers (2021-02-10T13:50:02Z) - A Sober Look at the Unsupervised Learning of Disentangled
Representations and their Evaluation [63.042651834453544]
We show that the unsupervised learning of disentangled representations is impossible without inductive biases on both the models and the data.
We observe that while the different methods successfully enforce properties "encouraged" by the corresponding losses, well-disentangled models seemingly cannot be identified without supervision.
Our results suggest that future work on disentanglement learning should be explicit about the role of inductive biases and (implicit) supervision.
arXiv Detail & Related papers (2020-10-27T10:17:15Z) - Manifolds for Unsupervised Visual Anomaly Detection [79.22051549519989]
Unsupervised learning methods that don't necessarily encounter anomalies in training would be immensely useful.
We develop a novel hyperspherical Variational Auto-Encoder (VAE) via stereographic projections with a gyroplane layer.
We present state-of-the-art results on visual anomaly benchmarks in precision manufacturing and inspection, demonstrating real-world utility in industrial AI scenarios.
arXiv Detail & Related papers (2020-06-19T20:41:58Z)
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.