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 2105.03484cs.CL
收藏

面向人类层面 NLP 的预训练 Transformer 实证评估:样本量与维度的作用

Empirical Evaluation of Pre-trained Transformers for Human-Level NLP: The Role of Sample Size and Dimensionality

Stony Brook University(石溪大学)

Adithya V Ganesan, Matthew Matero, Aravind Reddy Ravula, Huy Vu, H. Andrew Schwartz

AI总结:

本文系统评估预训练 Transformer 在小样本人类层面 NLP 任务中的表现,发现预训练降维可缓解微调困难,RoBERTa 表现最佳,且多数任务仅需 1/12 嵌入维度。

摘要:

在人类层面的 NLP 任务中,例如预测心理健康、人格或人口统计学特征,观测数量通常小于现代基于 Transformer 的语言模型中每一层标准的 768 维以上隐藏状态规模,从而限制了有效利用 Transformer 的能力。本文系统研究了降维方法(主成分分析、因子分解技术或多层自动编码器)的作用,以及嵌入向量维度和样本量作为预测性能函数的关系。我们首先发现,用有限数据微调大型模型存在显著困难,而这可通过预训练降维机制克服。RoBERTa 在人类层面任务中持续取得最佳性能;相比其他降维方法,PCA 在更好地处理撰写较长文本的用户方面具有优势。最后,我们观察到,大多数任务仅使用 1/12 的嵌入维度即可获得与最佳性能相当的结果。

英文摘要:

In human-level NLP tasks, such as predicting mental health, personality, or demographics, the number of observations is often smaller than the standard 768+ hidden state sizes of each layer within modern transformer-based language models, limiting the ability to effectively leverage transformers. Here, we provide a systematic study on the role of dimension reduction methods (principal components analysis, factorization techniques, or multi-layer auto-encoders) as well as the dimensionality of embedding vectors and sample sizes as a function of predictive performance. We first find that fine-tuning large models with a limited amount of data pose a significant difficulty which can be overcome with a pre-trained dimension reduction regime. RoBERTa consistently achieves top performance in human-level tasks, with PCA giving benefit over other reduction methods in better handling users that write longer texts. Finally, we observe that a majority of the tasks achieve results comparable to the best performance with just $\frac{1}{12}$ of the embedding dimensions.

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。

↑