Daphne Koller 计算机科学家

shiyi18

Daphne Koller
Computer Scientist | Class of 2004
Devising new strategies and algorithms for making decisions in complex systems under uncertain conditions, and applying statistical reasoning to classical problems in artificial intelligence.

Portrait of Daphne Koller
Photos for download >
Title
Computer Scientist
Affiliation
Stanford University
Location
Stanford, California
Age
36 at time of award
Area of Focus
Computer Science and Electrical Engineering
Published September 28, 2004
ABOUT DAPHNE'S WORK
Daphne Koller is a computer scientist who has developed new computational methods for representing knowledge and reasoning at the intersection of traditional logic, probability, uncertainty, and subjective judgment. Her work bridges a longstanding divide in the field of artificial intelligence between efforts to develop an explicit representation of knowledge (for example, in medical diagnosis) and efforts to categorize data based on statistical properties (such as optical character recognition). Koller significantly expanded the utility of Bayesian networks—computational devices for reasoning based on uncertain information—by showing how these structures can be organized in hierarchical, object-oriented networks. She and her colleagues advanced this one step further by developing "probabilistic relational models", admixtures of logical and statistical representations that can implement standard deductive reasoning without exhibiting the fragility of earlier systems to incorrect, incomplete, or uncertain inputs. Probabilistic relational models are capable of extracting knowledge embedded in large databases into a form that can be used for abstract reasoning. These advances find direct application in areas such as commerce, security, and biomedical research; for example, probabilistic relational models have been used to analyze the yeast genome, identifying regulatory roles for several proteins whose function had been previously uncharacterized. Through her research in graphical modeling, Koller has demonstrated the power of probabilistic methods for tackling the hardest problems in knowledge representation, inference, and learning.

BIOGRAPHY
Daphne Koller received a B.Sc. (1985) and a M.Sc. (1986) from the Hebrew University of Jerusalem and a Ph.D. (1993) from Stanford University. Koller was a postdoctoral fellow (1993-1995) at the University of California, Berkeley. In 1995, she joined the Department of Computer Science at Stanford University as an assistant professor and, in 2001, became an associate professor. Her research has been published in journals such as Games and Economic Behavior, Artificial Intelligence, Science, and Nature Genetics.



Daphne Koller
计算机科学家 | 2004级
为复杂系统在不确定条件下的决策设计新的策略和算法，并将统计推理应用于人工智能的经典问题。

达芙妮-科勒的画像
图片下载 >
标题
计算机科学家
工作单位
斯坦福大学
工作地点
斯坦福, 加州
年龄
获奖时为36岁
重点领域
计算机科学和电气工程
发表于2004年9月28日
关于达芙妮的工作
达芙妮-科勒是一位计算机科学家，她在传统逻辑、概率、不确定性和主观判断的交汇处开发了新的计算方法来表示知识和推理。她的工作弥合了人工智能领域长期存在的鸿沟，即努力开发知识的明确表示（例如在医学诊断中）和努力根据统计属性对数据进行分类（例如光学字符识别）。科勒极大地扩展了贝叶斯网络的效用--基于不确定信息进行推理的计算设备--展示了这些结构如何能够被组织在分层的、面向对象的网络中。她和她的同事们通过开发 "概率关系模型"，即逻辑和统计表征的混合物，可以实现标准的演绎推理，而不会表现出早期系统对不正确、不完整或不确定输入的脆弱性，从而进一步推动了这一点。概率关系模型能够将嵌入大型数据库中的知识提取为一种可用于抽象推理的形式。这些进展在商业、安全和生物医学研究等领域找到了直接的应用；例如，概率关系模型被用来分析酵母基因组，确定几个蛋白质的调节作用，而这些蛋白质的功能以前是没有被描述的。通过她在图形建模方面的研究，科勒展示了概率方法在解决知识表示、推理和学习方面最困难问题的力量。

个人简历
达芙妮-科勒在耶路撒冷希伯来大学获得理科学士（1985年）和理科硕士（1986年），在斯坦福大学获得博士学位（1993年）。科勒曾在加州大学伯克利分校担任博士后研究员（1993-1995）。1995年，她加入斯坦福大学计算机科学系，担任助理教授，2001年成为副教授。她的研究发表在《游戏与经济行为》、《人工智能》、《科学》和《自然遗传学》等杂志上。