Workshop on the Algorithmic, Mathematical, and Statistical Foundations of Data Science

关于数据科学的算法、数学和统计基础的研讨会将于2016年4月28日至30日在弗吉尼亚州阿灵顿举行，该活动将汇集计算机科学、数学和统计学领域的主要研究人员，以解决与数据科学相关的基础问题。研讨会的目标有三个：(i)确定数据科学这一新兴学科的基本领域，在这些领域中，计算机科学家、数学家和统计学家之间的合作是取得重大进展所必需的；（ii）评估计算机科学家、数学家和统计学家之间的合作如何通过推进和改变博士生和博士后的数据科学研究培训，为劳动力发展做出潜在贡献；（三）提出可以显著促进和推进此类合作的不同基础设施模式。研讨会的主要成果是一份白皮书，将作为专业协会和资助机构的指导方针。大数据现象的迅速兴起，带来了机遇，也带来了挑战。虽然大量数据可以生成模型和设计算法，从而提高推理能力，但这些模型和算法在中等规模的数据集上可能不太成功。研究人员面临的挑战是发展理论原则，使推理和学习能够扩展到大规模数据集，以及在数据生成和数据采样过程中存在异质性的情况下控制错误的算法。这些挑战需要理论计算机科学、数学、统计学、机器学习和数据挖掘以及高性能计算研究人员之间的合作。该奖项支持一个研讨会，该研讨会汇集了来自这些研究社区的领导者，讨论在这一发展领域合作工作的挑战和机遇。

A workshop on the Algorithmic, Mathematical, and Statistical Foundations of Data Science will be held April 28-30, 2016 in Arlington, VA. The event will bring together leading researchers in computer science, mathematics, and statistics to address foundational issues related to data science. The objectives of the workshop are three-fold: (i) identify fundamental areas in the emerging discipline of Data Science where collaboration between computer scientists, mathematicians, and statisticians is necessary to achieve significant progress; (ii) Assess how collaboration between computer scientists, mathematicians, and statisticians could potentially contribute to workforce development by advancing and transforming the Data Science research training of Ph.D. students and post-docs; and (iii) Suggest different infrastructure modalities that could significantly promote and advance such collaborations. The main deliverable of the workshop is a white paper that will serve as a guideline for professional societies and funding agencies. The rapid emergence of the Big Data phenomenon presents both opportunities and challenges. While massive data may allow the generation of models and the design of algorithms that have improved inferential power, such models and algorithms may be less successful on modest-sized data sets. The challenge for researchers is to develop theoretical principles that will allow the scaling of inference and learning to massive-scale datasets, and algorithms that control errors even in the presence of heterogeneity in the data generation and data sampling processes. These challenges will require collaborations between researchers representing theoretical computer science, mathematics, statistics, machine learning and data mining, and high performance computing. This award supports a workshop that brings together leaders from these communities of researchers to discuss the challenges and opportunities for collaborative work in this developing field.