ITR: New directions in clustering and learning

集群和学习的新方向面对越来越大量的数据，研究人员、政府机构、企业甚至公众都在寻找工具来帮助他们处理大量信息、识别其中的模式、了解这些模式的含义，并及时对这些信息采取行动。开发这样的工具涉及到算法、统计、人工智能和机器学习的新颖而有趣的混合。该项目在这些领域聚集了一个专家团队（四个来自学术界，两个来自工业界）来解决这些问题的一个有趣和有意义的子集，这些问题具有集群或学习的一般味道。聚类通常是由学习的最终目标驱动的，但也可以被看作是一个学习任务本身，因为它导致了对数据的更紧凑的描述。同样，所有的学习都涉及某种聚类，事实上，这一观点在最近的学习文献中是隐含的。该项目对学习数据模式的整个问题采取了综合的观点，从可能产生数据流的代表性草图的流计算开始，到将数据聚类为有意义的模式的问题（伴随着离群值去除，多目标优化等问题），到学习适合复杂模型的算法（支持向量机，贝叶斯网络，高斯混合等）研究人员认为，所有这些不同的算法努力都有统一的想法。此外，他们的协同方法提出了几个有趣的想法，可能会导致重大进展。示例如下：包括在不同的应用中使用编码理论思想，例如多类学习（一个广泛的学习问题，包括文本和语音分类，词性标记，手势识别等）。在视觉上形成一个轮廓;利用聚类思想进行降维（提供流行的基于SVD的方法的替代方案），and using运用ideas理念fromapproximation近似algorithms算法and clustering聚类to do near近-该项目还包括一个管理和教育计划，涉及通过开发新课程和学习软件来传播这项研究的想法，这些软件将被放置在公众中作为本项目的一部分，开发的算法将在大型数据集上进行测试，包括从Google Inc.一些算法的想法也将在行业（包括谷歌）中实现。

New directions in clustering and learningFaced with ever-larger amounts of data, researchers, government institutions, corporations and even the general public seek tools that help them deal with large bodies of information, identify patterns in it, learn what thesepatterns mean, and act upon that information in a timely fashion. Developing such tools involves a novel and interesting blend of algorithms, statistics, AI, and machine learning. The project assembles a team of experts (fourfrom academia and two from industry) in these areas to attack an interestingand meaningful subset of such problems which have the general flavor ofclustering or learning.The defining philosophy of this proposal is that no clear boundary Separates the twin notions of clustering and learning. Clustering is usually drivenby the end goal of learning, but can also be viewed as a learning taskin itself since it results in a more compact description of the data.By the same token all learning involves clustering of some sort, andin fact this viewpoint is implicit in recent papers in the learning literature. The project takes an integrated view of the entire problem of learningpatterns in data, starting from streaming computations that might producerepresentative sketches of the data as it streams by, to problems of clustering data into meaninful patterns (with attendant problems of outlier removal,multiobjective optimization etc.), to learning algorithms that fitsophisticated models (SVMs, bayesian nets, gaussian mixtures etc.) for inference and reasoning tasks.The investigators believe that all these disparate algorithmic efforts haveunifying ideas. Furthermore, their synergistic approach throws up severalinteresting ideas of its own that could lead to significant advances. Examples: include using coding theoretic ideas in disparate applications such as Multiclass learning (a broad class of learning problems including text and speech categorization, part-of-speech tagging, gesture recognition etc.) and shape recognitionin vision; the use of clustering ideas to do dimension reduction (offeringan alternative to popular SVD based approaches), and using ideas fromapproximation algorithms and clustering to do near-optimal model fittingfor models such as bayesian nets.The project also includes a management and educational plan that involvesdissemination of the ideas of this research through development of new courses and also pieces of learning software that will be placed in the public domain.Algorithms developed in as part of this project will be tested on large datasets, including those obtained from Google Inc. Some algorithmic ideas will also be implemented in industry (including Google).