• Corpus ID: 211205013

On Learning Sets of Symmetric Elements

@article{Maron2020OnLS,
  title={On Learning Sets of Symmetric Elements},
  author={Haggai Maron and Or Litany and Gal Chechik and Ethan Fetaya},
  journal={ArXiv},
  year={2020},
  volume={abs/2002.08599}
}
Learning from unordered sets is a fundamental learning setup, recently attracting increasing attention. Research in this area has focused on the case where elements of the set are represented by feature vectors, and far less emphasis has been given to the common case where set elements themselves adhere to their own symmetries. That case is relevant to numerous applications, from deblurring image bursts to multi-view 3D shape recognition and reconstruction. In this paper, we present a… 

Figures and Tables from this paper

Equivariant Subgraph Aggregation Networks
TLDR
A novel framework to represent each graph as a set of subgraphs derived by some predefined policy, and to process it using a suitable equivariant architecture, and it is proved that this approach increases the expressive power of both MPNNs and more expressive architectures.
Improved generalization bounds of group invariant / equivariant deep networks via quotient feature spaces
TLDR
This study develops a novel generalization error bound for invariant and equivariant deep neural networks using a notion of a quotient feature space, which measures the effect of group actions for the properties.
Inference-Time Unlabeled Personalized Federated Learning
TLDR
This work introduces a new learning setup, Inference-Time Unlabeled PFL (ITU-PFL), where a system trained on a set of clients, needs to be later applied to novel unlabeled clients at inference time, and proposes a novel approach, ITUFL-HN, which uses a hypernetwork to produce a new model for the late-to-the-party client.
Learning Prototype-oriented Set Representations for Meta-Learning
TLDR
Viewing each set as a distribution over a set of global prototypes, this paper provides a novel optimal transport (OT) based way to improve existing summary networks and minimize its OT distance to the set empirical distribution over data points, providing a natural unsupervised way to improved the summary network.
Distribution-Based Invariant Deep Networks for Learning Meta-Features
TLDR
A distribution-based invariant deep learning architecture is presented, and leveraged to extract the meta-features characterizing a dataset, to identify the SVM hyper-parameters best suited to a large benchmark of diversified small size datasets.
On the Adversarial Robustness of 3D Point Cloud Classification
TLDR
The first in-depth study showing how adversarial training (AT) behaves in point cloud classification is presented and it is identified that the required symmetric function (pooling operation) is paramount to the model's robustness under AT.
Position-Agnostic Multi-Microphone Speech Dereverberation
TLDR
This paper presents an NN architecture that can cope with microphone arrays on which no prior knowledge is presumed, and demonstrates its applicability on the speech dereverberation problem.
A Simple and Universal Rotation Equivariant Point-cloud Network
TLDR
A much simpler architecture is suggested, it is proved that it enjoys the same universality guarantees and its performance on Modelnet40 is evaluated.
Approximately Equivariant Networks for Imperfectly Symmetric Dynamics
TLDR
By relaxing equivariance constraints, approximately equivariant networks which are biased towards preserving symmetry but are not strictly constrained to do so can outperform both baseline with no symmetry bias and baselines with overly strict symmetry in both simulated turbulence domains and real-world multi-stream jet flow.
Equivariance Discovery by Learned Parameter-Sharing
TLDR
The partition distance is proposed to use the partition distance to empirically quantify the accuracy of the recovered equivariance, and the approach recovers known equivariances, such as permutations and shifts, on sum of numbers and spatially-invariant data.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 73 REFERENCES
Burst Image Deblurring Using Permutation Invariant Convolutional Neural Networks
TLDR
The novel convolutional architecture has a simultaneous view of all frames in the burst, and by construction treats them in an order-independent manner to effectively detect and leverage subtle cues scattered across different frames, while ensuring that each frame gets a full and equal consideration regardless of its position in the sequence.
Deep Models of Interactions Across Sets
TLDR
It is argued that models should hence be Permutation Equivariant (PE): constrained to make the same predictions across such permutations, and presented a parameter-sharing scheme and proved that it could not be made any more expressive without violating PE.
Large-scale celebfaces attributes (celeba) dataset
  • Retrieved August,
  • 2018
Places: A 10 Million Image Database for Scene Recognition
TLDR
The Places Database is described, a repository of 10 million scene photographs, labeled with scene semantic categories, comprising a large and diverse list of the types of environments encountered in the world, using the state-of-the-art Convolutional Neural Networks as baselines, that significantly outperform the previous approaches.
Deep Sets
TLDR
The main theorem characterizes the permutation invariant objective functions and provides a family of functions to which any permutation covariant objective function must belong, which enables the design of a deep network architecture that can operate on sets and which can be deployed on a variety of scenarios including both unsupervised and supervised learning tasks.
Dynamic FAUST: Registering Human Bodies in Motion
TLDR
This work proposes a new mesh registration method that uses both 3D geometry and texture information to register all scans in a sequence to a common reference topology, and shows how using geometry alone results in significant errors in alignment when the motions are fast and non-rigid.
UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
TLDR
This work introduces UCF101 which is currently the largest dataset of human actions and provides baseline action recognition results on this new dataset using standard bag of words approach with overall performance of 44.5%.
ImageNet: A large-scale hierarchical image database
TLDR
A new database called “ImageNet” is introduced, a large-scale ontology of images built upon the backbone of the WordNet structure, much larger in scale and diversity and much more accurate than the current image datasets.
On Universal Equivariant Set Networks
TLDR
It is proved that PointNet is not equivariant universal; and adding a single linear transmission layer makes PointNet universal, and the architecture called PointNetST is argued to be the simplest permutation equivariants universal model known to date.
Incidence Networks for Geometric Deep Learning
TLDR
It is shown that any incidence tensor decomposes into invariant subsets, which leads to a decomposition of the corresponding equivariant linear maps, for which it is proved an efficient pooling-and-broadcasting implementation.
...
1
2
3
4
5
...