Learning Adaptive Classifiers Synthesis for Generalized Few-Shot Learning

By Han-Jia Ye, Hexiang Hu, De-chuan Zhan, Fei Sha. 2019

66 citations recorded by the source corpus.

Object recognition in the real-world requires handling long-tailed or even open-ended data. An ideal visual system needs to recognize the populated head visual concepts reliably and meanwhile efficiently learn about emerging new tail categories with a few training instances. Class-balanced many-shot learning and few-shot learning tackle one side of this problem, by either learning strong classifiers for head or learning to learn few-shot classifiers for the tail. In this paper, we investigate the problem of generalized few-shot learning (GFSL)—a model during the deployment is required to learn about tail categories with few shots and simultaneously classify the head classes. We propose the ClAssifier SynThesis LEarning (Castle), a learning framework that learns how to synthesize calibrated few-shot classifiers in addition to the multi-class classifiers of head classes with a shared neural dictionary, shedding light upon the inductive GFSL. Furthermore, we propose an adaptive version of Castle (aCastle) that adapts the head classifiers conditioned on the incoming tail training examples, yielding a framework that allows effective backward knowledge transfer. As a consequence, aCastle can handle GFSL with classes from heterogeneous domains effectively. Castle and aCastle demonstrate superior performances than existing GFSL algorithms and strong baselines on MiniImageNet as well as TieredImageNet datasets. More interestingly, they outperform previous state-of-the-art methods when evaluated with standard few-shot learning criteria.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging, Representation & Self-Supervised Learning, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

In Representation & Self-Supervised Learning

In Meta & Continual Learning

Citations in the corpus

Few-Shot Class-Incremental Learning by Sampling Multi-Phase Tasks (2022, methodology): “Typical GFSL algorithms try to solve the problem by classifier weight generalization [72], [74], subspace regularization [75] and attention-based regularization [76].”
Overcoming Catastrophic Forgetting in Incremental Few-Shot Learning by Finding Flat Minima (2021, methodology): “Current few-shot methods mainly include optimization-based methods [12, 23, 32, 39, 45, 46, 55] and metric-based methods [13, 19, 44, 49, 52, 58, 57, 53].”
Co-Transport for Class-Incremental Learning (2021, background): “CIL for multimedia: It is common to observe the emergence of new classes in real-world applications [26, 32, 44, 52, 53, 55, 58, 65– 67].”
Few-Shot Incremental Learning with Continually Evolved Classifiers (2021, methodology): “Optimization-based methods [8,16,24,28,29,36,37,50] and metric-based methods [9, 12, 35, 42, 46, 47, 51, 52] are two main lines of efforts.”
Learning Placeholders for Open-Set Recognition (2021, background): “The aim to boost open-set recognition can be summarized as a calibration problem [8, 34].”
A Survey on Few-Shot Class-Incremental Learning (2023, methodology): “Additionally, Yoon et al. (2020) suggests a method for fusing base features, while Ye et al. (2021) puts forward the idea of synthesizing few-shot classifiers with a shared neural dictionary.”
Progressive One-shot Human Parsing (2020, unknown): “…in OS3, while both base classes and novel classes should be recognized simultaneously during evaluation in OSHP as shown in Figure 1 (b), which is indeed a variant of generalized few-shot learning (GFSL) problem (Gidaris and Komodakis 2018; Ren et al. 2019; Shi et al. 2019; Ye et al. 2019).”
CG-BERT: Conditional Text Generation with BERT for Generalized Few-shot Intent Detection (2020, background): “Ye et al. (2019) proposes a learning framework, Classiﬁer Synthesis Learning (CAS-TLE), which learns how to synthesize calibrated few-shot classiﬁers in addition to the classiﬁers of seen classes, leveraging a shared neural dictionary between seen classes and novel classes.”
SemGIF: A Semantics Guided Incremental Few-shot Learning Framework with Generative Replay (2021, methodology): “The traditional FSL work in [34] acts as our preliminary baseline B1, while the GFSL works in [32, 44] will serve as the baseline comparison B2 for the performance over the combined evaluation over the base and novel dataset with the data from the base dataset still being accessible.”
Few-Shot Learning With a Strong Teacher (2021, background): “For example, [73], [74] leveraged the graph propagation to improve meta-learning; [75], [76] constructed a generalized FSL objective towards adaptive calibration strategies between base and novel classes; [77] investigated specialized data augmentations; [78] studied how to identify invariant featur”