Related papers: Hyperbolic Multimodal Continual Learning

Hyperbolic Multimodal Continual Learning

URL: http://arxiv.org/abs/2608.09572v1
Date: Mon, 10 Aug 2026 13:07:26 GMT
Title: Hyperbolic Multimodal Continual Learning
Abstract summary: This work provides a theoretical foundation for representation preservation in hyperbolic space.<n>We show that forgetting in hyperbolic continual learning involves both semantic relation drift and hierarchy-related distortion.<n>We propose a principled continual learning framework that preserves essential geometric structure while allowing effective adaptation to new tasks.
Score: 90.3699346061366
License: http://creativecommons.org/licenses/by/4.0/
Abstract: Hyperbolic geometry has recently emerged as a powerful representation space for multimodal learning, as it naturally captures hierarchical semantic structure across modalities. Despite this progress, how such representations behave under continual learning poses fundamentally different challenges that remain underexplored. This work provides a geometric perspective on this problem and establishes a theoretical foundation for representation preservation in hyperbolic space, showing that preventing forgetting requires cross-modal invariance under a shared hyperbolic isometry. We further show that forgetting in hyperbolic continual learning involves both semantic relation drift and hierarchy-related distortion, motivating preservation of both cross-modal relational structure and hierarchical geometry. Guided by these insights, a principled continual learning framework is derived that preserves essential geometric structure while allowing effective adaptation to new tasks. Experiments on continual multimodal benchmarks corroborate the effectiveness of the proposed approach.

Related papers

AURORA: Contextual Orthogonalization for Geometric Representation Learning in Healthcare Foundation Models [1.7823291802496628]
We introduce AURORA, Adaptive Uncertainty Representations through Orthogonalized observational alignment.<n>AURORA decomposes semantic subspaces into contextual factors learns consistency objectives within each subspace.<n>It consistently outperforms reconstruction and substantially improves contextual disentanglement, neighborhood purity, and under institutional distribution shift.
arXiv Detail & Related papers (2026-05-18T02:32:50Z)
Anisotropic Modality Align [91.23979617826926]
Training multimodal large language models has long been limited by the scarcity of high-quality paired multimodal data.<n>Recent studies show that the shared representation space of pretrained multimodal contrastive models can serve as a bridge, enabling models to perform multimodal training with unimodal data.<n>The core obstacle lies in the persistent Modality Gap in the shared space.
arXiv Detail & Related papers (2026-05-08T14:53:24Z)
When Does Structure Matter in Continual Learning? Dimensionality Controls When Modularity Shapes Representational Geometry [4.121514039516763]
We study how network architecture, task similarity, and representational dimensionality jointly shape learning in a sequential task paradigm.<n>Our findings suggest that representational dimensionality acts as a key organizing variable governing when structural separation becomes functionally relevant.
arXiv Detail & Related papers (2026-04-30T09:50:01Z)
TopoOR: A Unified Topological Scene Representation for the Operating Room [54.50897207203292]
Surgical Scene Graphs abstract the complexity of surgical operating rooms (OR) into a structure of entities and their relations.<n>We introduce TopoOR, a new paradigm that models multimodal operating rooms as a higher-order structure.<n>We also propose a higher-order attention mechanism that explicitly preserves manifold structure and modality-specific features.
arXiv Detail & Related papers (2026-03-10T10:19:42Z)
Achieving Hyperbolic-Like Expressiveness with Arbitrary Euclidean Regions: A New Approach to Hierarchical Embeddings [9.614222676567385]
We present RegD, a flexible Euclidean framework that supports the use of arbitrary geometric regions as embedding representations.<n>RegD achieves hyperbolic-like expressiveness by incorporating a depth-based dissimilarity between regions, enabling it to emulate key properties of hyperbolic geometry.<n>Our empirical evaluation on diverse real-world datasets shows consistent performance gains over state-of-the-art methods.
arXiv Detail & Related papers (2025-01-29T09:44:03Z)
Neuron: Learning Context-Aware Evolving Representations for Zero-Shot Skeleton Action Recognition [64.56321246196859]
We propose a novel dyNamically Evolving dUal skeleton-semantic syneRgistic framework.<n>We first construct the spatial-temporal evolving micro-prototypes and integrate dynamic context-aware side information.<n>We introduce the spatial compression and temporal memory mechanisms to guide the growth of spatial-temporal micro-prototypes.
arXiv Detail & Related papers (2024-11-18T05:16:11Z)
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)
Understanding and Constructing Latent Modality Structures in Multi-modal Representation Learning [53.68371566336254]
We argue that the key to better performance lies in meaningful latent modality structures instead of perfect modality alignment. Specifically, we design 1) a deep feature separation loss for intra-modality regularization; 2) a Brownian-bridge loss for inter-modality regularization; and 3) a geometric consistency loss for both intra- and inter-modality regularization.
arXiv Detail & Related papers (2023-03-10T14:38:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.