Multi-view learning with Gaussian Process Latent Variable Models

本研究项目的目的是开发数据高效和可解释的建模方法，从多个视图。重点将放在非参数贝叶斯方法，如基于高斯过程的模型。该项目的目标是开发适用于各种不同场景的机器学习。我们将首先关注使用运动捕捉数据的应用程序。我们这样做的动机是双重的。人类运动是人类容易解释的东西，这使得很容易将“意义”附加到结果上。这意味着我们可以评估我们的第一个目标，创建可解释的模型。其次，我们希望数据高效。动作捕捉数据的收集成本很高，而且需要专门的设备。因此，我们必须从少量的数据中学习，这要求我们以最有效的形式使用它。这也导致了多视图动机，使用运动捕捉数据作为示例，考虑到我们有几个人执行相同的动作，我们将看到这些中的每一个都是相同的底层概念的视图，与我们必须只从一个人那里学习相比，直接扩展了我们的数据集。视图学习适用于各种场景。2几乎所有的东西都可以从多个角度看。比如两个人从不同的角度看到一辆在街上行驶的汽车;或者仅仅是一个人的两只眼睛。汽车的另一个视角是它的声音，甚至是它的气味。一张脸从十个不同的人的角度来看，或者在二十个不同的照明条件下。视角不限于物理对象;它也可以是例如从不同个人或人群进行的视角进行的行走动作。从不同的音乐家表演的角度来看。多视图学习的重点是利用数据中的这些联系来揭示潜在的概念，以联合的方式解释每个视图。这使我们能够了解不同的观点在哪些方面是相似的，在哪些方面是不同的，并在它们之间进行推理。任何学习系统的一个核心方面是能够询问模型。学到了什么？模型的可能性有多大？预测的确定性是什么？除了理解是许多应用的本质之外，如分析或诊断，它对于信任学习的结果、所做的假设、预测以及对进一步发展有用的指导和比较都是至关重要的。在过去的十年里，我们已经看到机器学习成功地应用于大量的新领域。它的许多成功来自于大数据集的可用性。在许多方面，大量的数据减少了学习系统的需求，因为它们可以被允许不那么抽象。然而，对于许多应用程序来说，大型数据集是不可用的（而且很可能是不可用的），这意味着我们需要尽可能有效地使用数据中的可用信息。因此，我们将致力于使用原则性不确定性传播的数据有效模型，以减少对大型数据集的需求。

The aim of this research project is to develop methodology for data-efficient and interpretable modeling frommultiple views. The focus will be on nonparametric Bayesian methods such as models based on Gaussianprocesses. The projects goals are to develop machine learning that will be applicable to a large range ofdifferent scenarios. We will initially focus on applications which use motion capture data. Our motivation forthis is twofold. Human motion is something which is readily interpretable by humans, making it easy to attach"meaning" to results. This means that we can evaluate our first goal, creating interpretable models. Secondly,we wish to be data efficient. Motion capture data is expensive to collect and requires specialist equipment.Therefore we will have to learn from small amounts of data which requires us to use it in its most efficientform. This also leads to the multi-view motivation, using motion capture data as an example, given that we haveseveral people performing the same action we will see each of these as views of the same underlying conceptdirectly expanding our dataset compared to if we had to learn from only one person.Even though motivated by motion capture data multi-view learning is applicable to a large range of scenarios.Virtually everything can be seen from more than one perspective. Such as a car driving on the street seen bytwo people from different angles; or just by the two eyes of a single person. Another perspective of the car isthe sound of it, or even the smell of it. A face from the perspectives of being ten different individuals, or intwenty different lighting conditions. Perspectives are not limited to being of physical objects; it can also be e.g.the action of walking from the perspectives of being carried out by different individuals or groups of people. Ora song from the perspectives of being performed by different musicians. Multi-view learning is focused onexploiting these connections in the data to uncover latent concepts that explains each of the views in a jointmanner. This allows us to understand in what aspects the different views are similar and where they differ; aswell as conducting inference between them. A central aspect of any learning systems is to be able to interrogatea model. What has been learnt? How likely is the model? What is the certainty of predictions? Besidesunderstanding being the essence of many applications, such as analysis or diagnosis, it is vital for trusting theresult of learning, the assumptions made, predictions as well as for guidance and comparison useful for furtherdevelopment. Over the last decade we have seen machine learning successfully applied to a large range of newdomains. Many of its successes comes from availability of large data sets. In many ways, large amounts of datahave reduced the demands of learning systems as they can be allowed to be less abstract. However, for manyapplications large datasets are not (and will most likely not be) available which means we need to use theinformation available in the data as efficiently as possible. Therefore we will work on data efficient models thatuse principled uncertainty propagation in order to reduce the need for large data sets.