Shallow Deep Learning Can Still Excel in Fine-Grained Few-Shot Learning

2025-07-29

AI Summary

Fine-grained few-shot learning (FGFSL) commonly assumes deeper backbones (e.g., ResNet12) are inherently superior, leading to the underutilization of shallow networks like ConvNet-4. Method: This paper challenges the “depth implies stronger representation” assumption and proposes the Location-Aware Constellation Network (LCN-4), a lightweight ConvNet-4 variant enhanced with explicit positional modeling. Its core innovations include: (1) a location-aware feature clustering module that incorporates spatial structural priors; and (2) a unified grid-based positional encoding compensation mechanism coupled with frequency-domain positional embeddings to recover absolute and relative spatial information lost in standard convolutions. Results: Evaluated on three fine-grained few-shot benchmarks, LCN-4 significantly outperforms existing ConvNet-4 baselines and matches or exceeds state-of-the-art ResNet12-based methods—demonstrating that shallow architectures, when augmented with principled position awareness, achieve competitive fine-grained discriminative power.

Technology Category

• Computer Vision: Representation Learning for Vision
• Machine Learning: Representation Learning
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Deep learning has witnessed the extensive utilization across a wide spectrum of domains, including fine-grained few-shot learning (FGFSL) which heavily depends on deep backbones. Nonetheless, shallower deep backbones such as ConvNet-4, are not commonly preferred because they're prone to extract a larger quantity of non-abstract visual attributes. In this paper, we initially re-evaluate the relationship between network depth and the ability to fully encode few-shot instances, and delve into whether shallow deep architecture could effectuate comparable or superior performance to mainstream deep backbone. Fueled by the inspiration from vanilla ConvNet-4, we introduce a location-aware constellation network (LCN-4), equipped with a cutting-edge location-aware feature clustering module. This module can proficiently encoder and integrate spatial feature fusion, feature clustering, and recessive feature location, thereby significantly minimizing the overall loss. Specifically, we innovatively put forward a general grid position encoding compensation to effectively address the issue of positional information missing during the feature extraction process of specific ordinary convolutions. Additionally, we further propose a general frequency domain location embedding technique to offset for the location loss in clustering features. We have carried out validation procedures on three representative fine-grained few-shot benchmarks. Relevant experiments have established that LCN-4 notably outperforms the ConvNet-4 based State-of-the-Arts and achieves performance that is on par with or superior to most ResNet12-based methods, confirming the correctness of our conjecture.

Problem

• Evaluating shallow networks' performance in fine-grained few-shot learning
• Addressing positional information loss in shallow feature extraction
• Improving feature clustering with location-aware techniques in FGFSL

Innovation

• Location-aware constellation network for FGFSL
• Grid position encoding compensates missing info
• Frequency domain embedding offsets location loss