Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Dai, Hao, Tang, Chong, Chauhan, Jagmohan
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2507.17368
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913956119969792
author Dai, Hao
Tang, Chong
Chauhan, Jagmohan
author_facet Dai, Hao
Tang, Chong
Chauhan, Jagmohan
contents Continual learning (CL) with long-tailed data distributions remains a critical challenge for real-world AI systems, where models must sequentially adapt to new classes while retaining knowledge of old ones, despite severe class imbalance. Existing methods struggle to balance stability and plasticity, often collapsing under extreme sample scarcity. To address this, we propose ViRN, a novel CL framework that integrates variational inference (VI) with distributional trilateration for robust long-tailed learning. First, we model class-conditional distributions via a Variational Autoencoder to mitigate bias toward head classes. Second, we reconstruct tail-class distributions via Wasserstein distance-based neighborhood retrieval and geometric fusion, enabling sample-efficient alignment of tail-class representations. Evaluated on six long-tailed classification benchmarks, including speech (e.g., rare acoustic events, accents) and image tasks, ViRN achieves a 10.24% average accuracy gain over state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2507_17368
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ViRN: Variational Inference and Distribution Trilateration for Long-Tailed Continual Representation Learning
Dai, Hao
Tang, Chong
Chauhan, Jagmohan
Machine Learning
Continual learning (CL) with long-tailed data distributions remains a critical challenge for real-world AI systems, where models must sequentially adapt to new classes while retaining knowledge of old ones, despite severe class imbalance. Existing methods struggle to balance stability and plasticity, often collapsing under extreme sample scarcity. To address this, we propose ViRN, a novel CL framework that integrates variational inference (VI) with distributional trilateration for robust long-tailed learning. First, we model class-conditional distributions via a Variational Autoencoder to mitigate bias toward head classes. Second, we reconstruct tail-class distributions via Wasserstein distance-based neighborhood retrieval and geometric fusion, enabling sample-efficient alignment of tail-class representations. Evaluated on six long-tailed classification benchmarks, including speech (e.g., rare acoustic events, accents) and image tasks, ViRN achieves a 10.24% average accuracy gain over state-of-the-art methods.
title ViRN: Variational Inference and Distribution Trilateration for Long-Tailed Continual Representation Learning
topic Machine Learning
url https://arxiv.org/abs/2507.17368