Investigating Neural Architectures by Synthetic Dataset Design

• URL: http://arxiv.org/abs/2204.11045v1
• Date: Sat, 23 Apr 2022 10:50:52 GMT
• Title: Investigating Neural Architectures by Synthetic Dataset Design
• Authors: Adrien Courtois, Jean-Michel Morel, Pablo Arias
• Abstract summary: Recent years have seen the emergence of many new neural network structures (architectures and layers) We sketch a methodology to measure the effect of each structure on a network's ability, by designing ad hoc synthetic datasets. We illustrate our methodology by building three datasets to evaluate each of the three following network properties.
• Score: 14.317837518705302
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent years have seen the emergence of many new neural network structures (architectures and layers). To solve a given task, a network requires a certain set of abilities reflected in its structure. The required abilities depend on each task. There is so far no systematic study of the real capacities of the proposed neural structures. The question of what each structure can and cannot achieve is only partially answered by its performance on common benchmarks. Indeed, natural data contain complex unknown statistical cues. It is therefore impossible to know what cues a given neural structure is taking advantage of in such data. In this work, we sketch a methodology to measure the effect of each structure on a network's ability, by designing ad hoc synthetic datasets. Each dataset is tailored to assess a given ability and is reduced to its simplest form: each input contains exactly the amount of information needed to solve the task. We illustrate our methodology by building three datasets to evaluate each of the three following network properties: a) the ability to link local cues to distant inferences, b) the translation covariance and c) the ability to group pixels with the same characteristics and share information among them. Using a first simplified depth estimation dataset, we pinpoint a serious nonlocal deficit of the U-Net. We then evaluate how to resolve this limitation by embedding its structure with nonlocal layers, which allow computing complex features with long-range dependencies. Using a second dataset, we compare different positional encoding methods and use the results to further improve the U-Net on the depth estimation task. The third introduced dataset serves to demonstrate the need for self-attention-like mechanisms for resolving more realistic depth estimation tasks.

Related papers

• Defining Neural Network Architecture through Polytope Structures of Dataset [53.512432492636236]
  This paper defines upper and lower bounds for neural network widths, which are informed by the polytope structure of the dataset in question. We develop an algorithm to investigate a converse situation where the polytope structure of a dataset can be inferred from its corresponding trained neural networks. It is established that popular datasets such as MNIST, Fashion-MNIST, and CIFAR10 can be efficiently encapsulated using no more than two polytopes with a small number of faces.
  arXiv  Detail & Related papers  (2024-02-04T08:57:42Z)
• On Characterizing the Evolution of Embedding Space of Neural Networks using Algebraic Topology [9.537910170141467]
  We study how the topology of feature embedding space changes as it passes through the layers of a well-trained deep neural network (DNN) through Betti numbers. We demonstrate that as depth increases, a topologically complicated dataset is transformed into a simple one, resulting in Betti numbers attaining their lowest possible value.
  arXiv  Detail & Related papers  (2023-11-08T10:45:12Z)
• Memorization with neural nets: going beyond the worst case [5.662924503089369]
  In practice, deep neural networks are often able to easily interpolate their training data. For real-world data, however, one intuitively expects the presence of a benign structure so that already occurs at a smaller network size than suggested by memorization capacity. We introduce a simple randomized algorithm that, given a fixed finite dataset with two classes, with high probability constructs an interpolating three-layer neural network in time.
  arXiv  Detail & Related papers  (2023-09-30T10:06:05Z)
• Homological Convolutional Neural Networks [4.615338063719135]
  We propose a novel deep learning architecture that exploits the data structural organization through topologically constrained network representations. We test our model on 18 benchmark datasets against 5 classic machine learning and 3 deep learning models.
  arXiv  Detail & Related papers  (2023-08-26T08:48:51Z)
• Provable Guarantees for Nonlinear Feature Learning in Three-Layer Neural Networks [49.808194368781095]
  We show that three-layer neural networks have provably richer feature learning capabilities than two-layer networks. This work makes progress towards understanding the provable benefit of three-layer neural networks over two-layer networks in the feature learning regime.
  arXiv  Detail & Related papers  (2023-05-11T17:19:30Z)
• SVNet: Where SO(3) Equivariance Meets Binarization on Point Cloud Representation [65.4396959244269]
  The paper tackles the challenge by designing a general framework to construct 3D learning architectures. The proposed approach can be applied to general backbones like PointNet and DGCNN. Experiments on ModelNet40, ShapeNet, and the real-world dataset ScanObjectNN, demonstrated that the method achieves a great trade-off between efficiency, rotation, and accuracy.
  arXiv  Detail & Related papers  (2022-09-13T12:12:19Z)
• Dive into Layers: Neural Network Capacity Bounding using Algebraic Geometry [55.57953219617467]
  We show that the learnability of a neural network is directly related to its size. We use Betti numbers to measure the topological geometric complexity of input data and the neural network. We perform the experiments on a real-world dataset MNIST and the results verify our analysis and conclusion.
  arXiv  Detail & Related papers  (2021-09-03T11:45:51Z)
• Neural Network Layer Algebra: A Framework to Measure Capacity and Compression in Deep Learning [0.0]
  We present a new framework to measure the intrinsic properties of (deep) neural networks. While we focus on convolutional networks, our framework can be extrapolated to any network architecture.
  arXiv  Detail & Related papers  (2021-07-02T13:43:53Z)
• A neural anisotropic view of underspecification in deep learning [60.119023683371736]
  We show that the way neural networks handle the underspecification of problems is highly dependent on the data representation. Our results highlight that understanding the architectural inductive bias in deep learning is fundamental to address the fairness, robustness, and generalization of these systems.
  arXiv  Detail & Related papers  (2021-04-29T14:31:09Z)
• Neural networks adapting to datasets: learning network size and topology [77.34726150561087]
  We introduce a flexible setup allowing for a neural network to learn both its size and topology during the course of a gradient-based training. The resulting network has the structure of a graph tailored to the particular learning task and dataset.
  arXiv  Detail & Related papers  (2020-06-22T12:46:44Z)
• Stochastic encoding of graphs in deep learning allows for complex analysis of gender classification in resting-state and task functional brain networks from the UK Biobank [0.13706331473063876]
  We introduce a encoding method in an ensemble of CNNs to classify functional connectomes by gender. We measure the salience of three brain networks involved in task- and resting-states, and their interaction.
  arXiv  Detail & Related papers  (2020-02-25T15:10:51Z)

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.