arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.02841cs.LGcs.AI
收藏

语义感知的生成潜在数据增强用于低资源领域的学习

Semantics-Aware Generative Latent Data Augmentation for Learning in Low-Resource Domains

Siebel School of Computing(计算科学系);Data Science, University of Illinois Urbana-Champaign, Illinois, USA(数据科学,伊利诺伊大学厄巴纳-香槟分校,伊利诺伊州,美国)

Jaesung Bae, Minje Kim

AI总结:

GeLDA通过语义感知的生成潜在数据增强方法,在低资源领域提升语音情绪识别和图像分类任务的性能。

摘要:

尽管在数据丰富的环境中表现强劲,深度学习在实践中常见的数据稀缺设置中往往表现不佳。虽然在大规模数据集上训练的基模型(FMs)通过提取通用特征表现出强大的泛化能力,但在下游微调过程中仍可能因缺乏标注数据而遇到困难。为此,我们提出了GeLDA,一种语义感知的生成潜在数据增强框架,该框架利用条件扩散模型在FM诱导的潜在空间中合成样本。由于该空间是低维的,并且比输入空间更集中地包含任务相关的信息,GeLDA能够实现高效、高质量的数据生成。GeLDA通过辅助特征向量对生成过程进行条件控制,这些向量捕捉了类别或子领域之间的语义关系,从而在低资源领域中实现数据增强。我们在两个大规模识别任务中验证了GeLDA:(a)在零样本语言特定语音情绪识别中,GeLDA将Whisper-large基线的无加权召回率提高了6.13%;(b)在长尾图像分类中,它在ImageNet-LT上实现了74.7%的尾类准确率,设定了新的最先进结果。

英文摘要:

Despite strong performance in data-rich regimes, deep learning often underperforms in the data-scarce settings common in practice. While foundation models (FMs) trained on massive datasets demonstrate strong generalization by extracting general-purpose features, they can still suffer from scarce labeled data during downstream fine-tuning. To address this, we propose GeLDA, a semantics-aware generative latent data augmentation framework that leverages conditional diffusion models to synthesize samples in an FM-induced latent space. Because this space is low-dimensional and concentrates task-relevant information compared to the input space, GeLDA enables efficient, high-quality data generation. GeLDA conditions generation on auxiliary feature vectors that capture semantic relationships among classes or subdomains, facilitating data augmentation in low-resource domains. We validate GeLDA in two large-scale recognition tasks: (a) in zero-shot language-specific speech emotion recognition, GeLDA improves the Whisper-large baseline's unweighted average recall by 6.13%; and (b) in long-tailed image classification, it achieves 74.7% tail-class accuracy on ImageNet-LT, setting a new state-of-the-art result.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑