![[2602.22823] Hypernetwork-based approach for grid-independent functional data clustering](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22823] Hypernetwork-based approach for grid-independent functional data clustering
Summary
This article presents a novel hypernetwork-based framework for grid-independent functional data clustering, addressing limitations of existing methods that rely on sampled grids.
Why It Matters
The proposed approach enhances the robustness and flexibility of functional data clustering by decoupling cluster assignments from resolution and sampling density. This innovation is significant for researchers and practitioners in machine learning, as it allows for more accurate clustering of functional data across varying conditions.
Key Takeaways
- Introduces a hypernetwork framework for functional data clustering.
- Decouples clustering from sampling resolution and grid choices.
- Demonstrates competitive performance in high-dimensional settings.
Computer Science > Machine Learning arXiv:2602.22823 (cs) [Submitted on 26 Feb 2026] Title:Hypernetwork-based approach for grid-independent functional data clustering Authors:Anirudh Thatipelli, Ali Siahkoohi View a PDF of the paper titled Hypernetwork-based approach for grid-independent functional data clustering, by Anirudh Thatipelli and Ali Siahkoohi View PDF HTML (experimental) Abstract:Functional data clustering is concerned with grouping functions that share similar structure, yet most existing methods implicitly operate on sampled grids, causing cluster assignments to depend on resolution, sampling density, or preprocessing choices rather than on the underlying functions themselves. To address this limitation, we introduce a framework that maps discretized function observations -- at arbitrary resolution and on arbitrary grids -- into a fixed-dimensional vector space via an auto-encoding architecture. The encoder is a hypernetwork that maps coordinate-value pairs to the weight space of an implicit neural representation (INR), which serves as the decoder. Because INRs represent functions with very few parameters, this design yields compact representations that are decoupled from the sampling grid, while the hypernetwork amortizes weight prediction across the dataset. Clustering is then performed in this weight space using standard algorithms, making the approach agnostic to both the discretization and the choice of clustering method. By means of synthetic and real-w...
