arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.03908cs.LGstat.ML
收藏

在结构化生物医学领域中实现预测不变性

Enforcing Predictive Invariance across Structured Biomedical Domains

Massachusetts Institute of Technology(麻省理工学院)

Wengong Jin, Regina Barzilay, Tommi Jaakkola

摘要:

许多生物化学应用,如分子性质预测,需要模型泛化超出其训练域(环境)。此外,这些任务中的自然环境是结构化的,由复杂的描述符(如分子支架或蛋白质家族)定义。因此,大多数环境要么在训练期间从未出现过,要么只包含单个训练样本。为了解决这些挑战,我们提出了一个新的遗憾最小化(RGM)算法及其扩展的结构化环境。RGM建立在不变风险最小化(RGM)的基础上,根据预测性后悔重新铸造同时最优性条件,找到一种表示,使预测器能够与后见之明访问保留环境的预言竞争。结构化的扩展自适应地突出了由于复杂的环境,通过专门的域扰动的变化。我们评估了我们的方法在多个应用:分子特性预测,蛋白质同源性和稳定性预测,并表明RGM显着优于以前的最先进的基线。

英文摘要:

Many biochemical applications such as molecular property prediction require models to generalize beyond their training domains (environments). Moreover, natural environments in these tasks are structured, defined by complex descriptors such as molecular scaffolds or protein families. Therefore, most environments are either never seen during training, or contain only a single training example. To address these challenges, we propose a new regret minimization (RGM) algorithm and its extension for structured environments. RGM builds from invariant risk minimization (IRM) by recasting simultaneous optimality condition in terms of predictive regret, finding a representation that enables the predictor to compete against an oracle with hindsight access to held-out environments. The structured extension adaptively highlights variation due to complex environments via specialized domain perturbations. We evaluate our method on multiple applications: molecular property prediction, protein homology and stability prediction and show that RGM significantly outperforms previous state-of-the-art baselines.

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。

↑