PRISM: A Rich Class of Parameterized Submodular Information Measures for Guided Data Subset Selection
PRISM: A Rich Class of Parameterized Submodular Information Measures for Guided Data Subset Selection
复制标题
DOI：
10.1609/aaai.v36i9.21264
复制
发表时间：
2021-02
期刊：
影响因子：
--
通讯作者：
Suraj Kothawade;Vishal Kaushal;Ganesh Ramakrishnan;J. Bilmes;Rishabh K. Iyer
中科院分区：
文献类型：
--
作者：
Suraj Kothawade;Vishal Kaushal;Ganesh Ramakrishnan;J. Bilmes;Rishabh K. Iyer
随着数据集大小的不断增加，子集选择技术对于过多的任务变得越来越重要。通常有必要指导子集选择以实现某些必要的目标，包括聚焦或瞄准某些数据点，同时避免其他数据点。这样的问题的示例包括：i）有针对性的学习，其中目标是找到具有模型正在执行的稀有类或稀有属性的子集，以及ii）引导式摘要，其中数据（例如，图像集合、文本、文档或视频）被概括，以用于具有特定附加用户意图的更快的人类消费。受这些应用的启发，我们提出了PRISM，一类丰富的参数化子模信息测度。通过新颖的功能及其参数化，PRISM提供了各种建模功能，这些功能可以在子集的期望质量（如多样性或表示）与一组数据点的相似性/不相似性之间进行权衡。我们演示了如何PRISM可以应用到上述两个现实世界的问题，这需要指导子集选择。在这样做时，我们表明，PRISM有趣地概括了过去的一些工作，从而加强了其广泛的实用性。通过对不同数据集的广泛实验，我们证明了PRISM在有针对性的学习和引导图像收集摘要方面优于最先进的技术。PRISM是SUBMODLIB（https：//github.com/decile-team/submodlib）和TRUST（https：//github.com/decile-team/trust）工具包的一部分。
With ever-increasing dataset sizes, subset selection techniques are becoming increasingly important for a plethora of tasks. It is often necessary to guide the subset selection to achieve certain desiderata, which includes focusing or targeting certain data points, while avoiding others. Examples of such problems include: i)targeted learning, where the goal is to find subsets with rare classes or rare attributes on which the model is under performing, and ii)guided summarization, where data (e.g., image collection, text, document or video) is summarized for quicker human consumption with specific additional user intent. Motivated by such applications, we present PRISM, a rich class of PaRameterIzed Submodular information Measures. Through novel functions and their parameterizations, PRISM offers a variety of modeling capabilities that enable a trade-off between desired qualities of a subset like diversity or representation and similarity/dissimilarity with a set of data points. We demonstrate how PRISM can be applied to the two real-world problems mentioned above, which require guided subset selection. In doing so, we show that PRISM interestingly generalizes some past work, therein reinforcing its broad utility. Through extensive experiments on diverse datasets, we demonstrate the superiority of PRISM over the state-of-the-art in targeted learning and in guided image-collection summarization. PRISM is available as a part of the SUBMODLIB (https://github.com/decile-team/submodlib) and TRUST (https://github.com/decile-team/trust) toolkits.