ITR: Computational Induction of Scientific Process Models

这种跨学科、跨机构的合作研究努力旨在建立一个框架，统一信息技术中两个独立但核心的主题--对解释重要现象的模型进行计算模拟，以及从观察到的数据规律中通过计算归纳知识。与机器学习和数据挖掘中的大多数工作不同，该方法强调以既定的科学形式生成知识的方法，在可能的情况下纳入领域知识，关注因果和解释性模型，解决从观测时间序列数据进行归纳的方法，并嵌入到科学家可以用于模型开发的模拟环境中。研究围绕着一类新的模型，包括相互作用的定量过程和从时间序列数据中归纳出这样的模型的问题。将解决的计算挑战包括减少过拟合和方差、诱导过程条件、处理具有缺失值的大型、不同种类的数据集，以及扩展到复杂模型。由此产生的算法将被包括在一个可训练的模拟环境中，该环境允许用户手动构建模型或从数据中归纳模型，然后模拟他们的行为。实验评估将包括来自罗斯海的地球科学观测和合成数据。可训练的模拟环境应该允许地球科学家系统地搜索候选模型的空间，在更短的时间内产生更准确的模型。此外，新的计算方法应该有助于系统生物学和工程学等其他领域的模型构建。环境和样本模型都将在课程中使用，并可通过网站访问。欲了解更多信息，请访问http://www.isle.org/process.html.。

This interdisciplinary, interinstitutional collaborative research effort aims to develop a framework that unifies two separate but central themes in information technology -- computational simulation of models to explain important phenomena and computational induction of knowledge from observed regularities in data. Unlike most previous work in machine learning and data mining, the approach emphasizes methods that generate knowledge in established scientific formalisms, incorporate domain knowledge where possible, focus on causal and explanatory models, address induction from observational time-series data, and are embedded in a simulationenvironment which scientists can use for model development.The research revolves around a new class of models that consist ofinteracting quantitative processes and the problem of inducing suchmodels from time-series data. Computational challenges that will beaddressed include reducing overfitting and variance, inducing conditions on processes, handling large, heterogeneous data sets with missingvalues, and scaling to complex models. The resulting algorithms will be included in a trainable simulation environment that lets users construct models manually or induce them from data, then simulate their behavior. Experimental evaluation will involve both Earth Science observationsfrom the Ross Sea and synthetic data.The trainable simulation environment should let Earth scientists searchthe space of candidate models systematically, producing more accurate models in much less time. Moreover, the novel computational methodsshould aid model construction in other fields like systems biology andengineering. Both the environment and sample models will be utilizedin courses and accessible through a Web site. More information isavailable at http://www.isle.org/process.html.