Through a Steerable Lens: Magnifying Neural Network Interpretability via Phase-Based Extrapolation

• URL: http://arxiv.org/abs/2506.02300v3
• Date: Wed, 11 Jun 2025 06:26:42 GMT
• Title: Through a Steerable Lens: Magnifying Neural Network Interpretability via Phase-Based Extrapolation
• Authors: Farzaneh Mahdisoltani, Saeed Mahdisoltani, Roger B. Grosse, David J. Fleet,
• Abstract summary: We propose a novel framework that visualizes the implicit path between classes by treating the network gradient as a form of infinitesimal motion.<n>Experiments on both synthetic and real-world datasets demonstrate that our phase-focused extrapolation yields perceptually aligned, semantically meaningful transformations.
• Score: 26.45789667046442
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Understanding the internal representations and decision mechanisms of deep neural networks remains a critical open challenge. While existing interpretability methods often identify influential input regions, they may not elucidate how a model distinguishes between classes or what specific changes would transition an input from one category to another. To address these limitations, we propose a novel framework that visualizes the implicit path between classes by treating the network gradient as a form of infinitesimal motion. Drawing inspiration from phase-based motion magnification, we first decompose images using invertible transforms-specifically the Complex Steerable Pyramid-then compute class-conditional gradients in the transformed space. Rather than iteratively integrating the gradient to trace a full path, we amplify the one-step gradient to the input and perform a linear extrapolation to expose how the model moves from source to target class. By operating in the steerable pyramid domain, these amplified gradients produce semantically meaningful, spatially coherent morphs that highlight the classifier's most sensitive directions, giving insight into the geometry of its decision boundaries. Experiments on both synthetic and real-world datasets demonstrate that our phase-focused extrapolation yields perceptually aligned, semantically meaningful transformations, offering a novel, interpretable lens into neural classifiers' internal representations.

Related papers

• Physics-Aware Style Transfer for Adaptive Holographic Reconstruction [1.8749305679160366]
  Inline holographic imaging presents an ill-posed inverse problem of reconstructing objects' complex amplitude from recorded diffraction patterns.<n>We present a physics-aware style transfer approach that interprets the object-to-sensor distance as an implicit style within diffraction patterns.<n>We show that the inverse mapping operation can be learned in an adaptive manner only with datasets composed of intensity measurements.
  arXiv  Detail & Related papers  (2025-07-01T06:56:51Z)
• Mapping the Edge of Chaos: Fractal-Like Boundaries in The Trainability of Decoder-Only Transformer Models [0.0]
  Recent evidence from miniature neural networks suggests that the boundary separating these outcomes displays fractal characteristics.<n>This study extends them to medium-sized, decoder-only transformer architectures by employing a more consistent convergence measure.<n>The results show that the trainability frontier is not a simple threshold; rather, it forms a self-similar yet seemingly random structure at multiple scales.
  arXiv  Detail & Related papers  (2025-01-08T05:24:11Z)
• Flow Factorized Representation Learning [109.51947536586677]
  We introduce a generative model which specifies a distinct set of latent probability paths that define different input transformations. We show that our model achieves higher likelihoods on standard representation learning benchmarks while simultaneously being closer to approximately equivariant models.
  arXiv  Detail & Related papers  (2023-09-22T20:15:37Z)
• Latent Traversals in Generative Models as Potential Flows [113.4232528843775]
  We propose to model latent structures with a learned dynamic potential landscape. Inspired by physics, optimal transport, and neuroscience, these potential landscapes are learned as physically realistic partial differential equations. Our method achieves both more qualitatively and quantitatively disentangled trajectories than state-of-the-art baselines.
  arXiv  Detail & Related papers  (2023-04-25T15:53:45Z)
• Unsupervised Domain Transfer with Conditional Invertible Neural Networks [83.90291882730925]
  We propose a domain transfer approach based on conditional invertible neural networks (cINNs) Our method inherently guarantees cycle consistency through its invertible architecture, and network training can efficiently be conducted with maximum likelihood. Our method enables the generation of realistic spectral data and outperforms the state of the art on two downstream classification tasks.
  arXiv  Detail & Related papers  (2023-03-17T18:00:27Z)
• Latent Transformations via NeuralODEs for GAN-based Image Editing [25.272389610447856]
  We show that nonlinear latent code manipulations realized as flows of a trainable Neural ODE are beneficial for many practical non-face image domains. In particular, we investigate a large number of datasets with known attributes and demonstrate that certain attribute manipulations are challenging to obtain with linear shifts only.
  arXiv  Detail & Related papers  (2021-11-29T18:59:54Z)
• Graph Modularity: Towards Understanding the Cross-Layer Transition of Feature Representations in Deep Neural Networks [7.187240308034312]
  We move a tiny step towards understanding the transition of feature representations in deep neural networks (DNNs) We first characterize this transition by analyzing the class separation in intermediate layers, and next model the process of class separation as community evolution in dynamic graphs. We find that modularity tends to rise as the layer goes deeper, but descends or reaches a plateau at particular layers.
  arXiv  Detail & Related papers  (2021-11-24T13:29:17Z)
• Topographic VAEs learn Equivariant Capsules [84.33745072274942]
  We introduce the Topographic VAE: a novel method for efficiently training deep generative models with topographically organized latent variables. We show that such a model indeed learns to organize its activations according to salient characteristics such as digit class, width, and style on MNIST. We demonstrate approximate equivariance to complex transformations, expanding upon the capabilities of existing group equivariant neural networks.
  arXiv  Detail & Related papers  (2021-09-03T09:25:57Z)
• Augmenting Implicit Neural Shape Representations with Explicit Deformation Fields [95.39603371087921]
  Implicit neural representation is a recent approach to learn shape collections as zero level-sets of neural networks. We advocate deformation-aware regularization for implicit neural representations, aiming at producing plausible deformations as latent code changes.
  arXiv  Detail & Related papers  (2021-08-19T22:07:08Z)
• SNARF: Differentiable Forward Skinning for Animating Non-Rigid Neural Implicit Shapes [117.76767853430243]
  We introduce SNARF, which combines the advantages of linear blend skinning for polygonal meshes with neural implicit surfaces. We propose a forward skinning model that finds all canonical correspondences of any deformed point using iterative root finding. Compared to state-of-the-art neural implicit representations, our approach generalizes better to unseen poses while preserving accuracy.
  arXiv  Detail & Related papers  (2021-04-08T17:54:59Z)
• Image-to-image Mapping with Many Domains by Sparse Attribute Transfer [71.28847881318013]
  Unsupervised image-to-image translation consists of learning a pair of mappings between two domains without known pairwise correspondences between points. Current convention is to approach this task with cycle-consistent GANs. We propose an alternate approach that directly restricts the generator to performing a simple sparse transformation in a latent layer.
  arXiv  Detail & Related papers  (2020-06-23T19:52:23Z)
• Fast Symmetric Diffeomorphic Image Registration with Convolutional Neural Networks [11.4219428942199]
  We present a novel, efficient unsupervised symmetric image registration method. We evaluate our method on 3D image registration with a large scale brain image dataset.
  arXiv  Detail & Related papers  (2020-03-20T22:07:24Z)

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.