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 2504.01650stat.MLcs.LG
收藏

稀疏高斯神经过程

Sparse Gaussian Neural Processes

Helmholtz AI(海德堡人工智能研究所);Technical University of Munich(慕尼黑技术大学);Munich Center for Machine Learning(慕尼黑机器学习中心)

Tommy Rochussen, Vincent Fortuin

AI总结:

本文提出稀疏高斯神经过程模型族,元学习稀疏高斯过程推断,实现快速预测并首次允许在神经过程中手动设定先验,适用于任务数少或有专家知识的场景。

摘要:

尽管概率元学习在近期取得了显著进展,但实践者通常因深度学习模型相对缺乏可解释性而避免使用它们。相反,许多实践者直接使用具有可解释先验的高斯过程等非元模型,并为每个遇到的任务从头开始进行繁琐的模型训练过程。虽然对于数据点数量有限的任务这是合理的,但精确高斯过程推断的立方计算成本使得当每个任务具有大量观测时变得不可行。为解决这一问题,我们引入了一族元学习稀疏高斯过程推断的模型。这不仅使得在新任务上能够通过稀疏高斯过程进行快速预测,而且由于我们的模型作为神经过程家族成员具有清晰的解释,它还首次允许在神经过程中手动引出先验。在观测任务数量较少或专家领域知识可用的元学习场景中,这提供了关键优势。

英文摘要:

Despite significant recent advances in probabilistic meta-learning, it is common for practitioners to avoid using deep learning models due to a comparative lack of interpretability. Instead, many practitioners simply use non-meta-models such as Gaussian processes with interpretable priors, and conduct the tedious procedure of training their model from scratch for each task they encounter. While this is justifiable for tasks with a limited number of data points, the cubic computational cost of exact Gaussian process inference renders this prohibitive when each task has many observations. To remedy this, we introduce a family of models that meta-learn sparse Gaussian process inference. Not only does this enable rapid prediction on new tasks with sparse Gaussian processes, but since our models have clear interpretations as members of the neural process family, it also allows manual elicitation of priors in a neural process for the first time. In meta-learning regimes for which the number of observed tasks is small or for which expert domain knowledge is available, this offers a crucial advantage.

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。

↑