Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 5 of 5 for “"Equivariant Machine Learning"”.

  1. Learning O(d)-Equivariant Cartesian Tensor Functions

    … them a natural area of study for the field of equivariant machine learning, in particular equivariance to the orthogonal group, $\Od$. We characterize $\Od$-equivariant tensor functions on tuples of tensors as well as tensor images. We also use $\Od$ symmetries to develop any-dimensional …

    jhu Repository record for Learning O(d)-Equivariant Cartesian Tensor Functions (opens in a new tab)

  2. DATA-DRIVEN AND CONSTRAINED APPROACHES TO LEARNING AND ENFORCING SYMMETRY IN NEURAL NETWORKS

    … and physics that plays a central role in machine learning. This dissertation examines two dominant paradigms for leveraging symmetry in neural networks: data augmentation, which exposes models to transformed inputs during training, and equivariant architectures, which guarantee symmetry by …

    penn Repository record for DATA-DRIVEN AND CONSTRAINED APPROACHES TO LEARNING AND ENFORCING SYMMETRY IN NEURAL NETWORKS (opens in a new tab)

  3. Symmetry in Graph Learning

    … Graph neural networks have emerged as popular machine learning tools to learn and reason with graph data. Despite their promise and some success stories, graph neural networks have not managed to absolutely outperform classical methods such as spectral embeddings and graph kernels. Moreover, it …

    jhu Repository record for Symmetry in Graph Learning (opens in a new tab)

  4. Any-dimensional Invariant Universality

    Traditional supervised learning typically operates within fixed input-output dimensions. In contrast, the emerging research field of any-dimensional learning aims to approximate sequences of functions defined over arbitrary dimensions. By leveraging symmetries such as group action invariance and …

    jhu Repository record for Any-dimensional Invariant Universality (opens in a new tab)

  5. Equivariant Sensor Fusion for Autonomous Driving Object Detection

    … architectures. The first, EquiFusion, applies equivariant structure throughout the full detection pipeline. The second, HybridFusion, combines an equivariant stream with a parallel non-equivariant stream in order to balance geometric structure with feature flexibility. We also evaluate the …

    mcmaster Repository record for Equivariant Sensor Fusion for Autonomous Driving Object Detection (opens in a new tab)