|
学习共识意见：从标签游戏中挖掘数据
Learning Consensus Opinion: Mining Data from a Labeling Game |
|
| |
| 课程网址: |
http://videolectures.net/www09_bennett_lco/
|
| 主讲教师: |
Paul N. Bennett; Anton Mityagin; David Maxwell Chickering
|
| 开课单位: |
微软公司
|
| 开课时间: |
信息不详。欢迎您在右侧留言补充。
|
| 课程语种: |
英语
|
| 中文简介: |
在本文中，我们考虑如何识别一组用户的一致意见，以及如何对查询结果进行排序。一旦为一组查询确定了一致的排名，这些排名就可以用于检索和学习系统的评估和培训。我们提出了一种收集用户对图像搜索结果偏好的新方法:我们使用一种协作游戏，玩家在同意哪种图像结果最适合查询时获得奖励。我们的方法不同于其他标记游戏，因为我们能够直接从查询日志中提取出图像查询的兴趣偏好。作为相关性判断的来源，该数据为点击数据提供了有用的补充。此外，它不存在位置偏差，也不存在用与消除偏误的机制相关的非相关结果让用户感到沮丧的风险。我们描述了从这款游戏的一个已部署版本中收集了超过35天的数据，这些数据相当于1900万对玩家之间表达的喜好。最后，我们提出了几种对这些数据建模的方法，以便从首选项中提取一致的排名，并对目标查询的搜索结果进行更好的排序。
|
| 课程简介: |
In this paper, we consider the challenge of how to identify the consensus opinion of a set of users as to how the results for a query should be ranked. Once consensus rankings are identified for a set of queries, these rankings can serve for both evaluation and training of retrieval and learning systems. We present a novel approach to collecting user preferences over image-search results: we use a collaborative game in which players are rewarded for agreeing on which image result is best for a query. Our approach is distinct from other labeling games because we are able to elicit directly the preferences of interest with respect to image queries extracted from query logs. As a source of relevance judgments, this data provides a useful complement to click data. Furthermore, it is free of positional biases and does not carry the risk of frustrating users with non-relevant results associated with proposed mechanisms for debiasing clicks. We describe data collected over 35 days from a deployed version of this game that amounts to about 19 million expressed preferences between pairs. Finally, we present several approaches to modeling this data in order to extract the consensus rankings from the preferences and better sort the search results for targeted queries.
|
| 关 键 词: |
图像搜索; 数据挖掘; 建模
|
| 课程来源: |
视频讲座网
|
| 最后编审: |
2020-06-08：吴雨秋(课程编辑志愿者)
|
| 阅读次数: |
41
|
|
图 书 馆