用ML和人类在循环中构建Airbnb类别
Airbnb Categories Blog Series — Part II : ML Categorization
Airbnb分类博客系列 - 第二部分：ML分类
By: Mihajlo Grbovic, Pei Xiong, Pratiksha Kadam, Ying Xiao, Aaron Yin, Weiping Peng, Shukun Yang, Chen Qian, Haowei Zhang, Sebastien Dubois, Nate Ney, James Furnary, Mark Giangreco, Nate Rosenthal, Cole Baker, Bill Ulammandakh, Shankar Shetty, Sid Reddy, Egor Pakhomov
作者Mihajlo Grbovic, Pei Xiong, Pratiksha Kadam, Ying Xiao, Aaron Yin, Weiping Peng, Shukun Yang, Chen Qian, Haowei Zhang, Sebastien Dubois, Nate Ney, James Furnary, Mark Giangreco, Nate Rosenthal, Cole Baker, Bill Ulammandakh, Shankar Shetty,Sid Reddy, Egor Pakhomov
Airbnb 2022 release introduced Categories, a browse focused product that allows the user to seek inspiration by browsing collections of homes revolving around a common theme, such as Lakefront, Countryside, Golf, Desert, National Parks, Surfing, etc. In Part I of our Categories Blog Series we covered the high level approach to creating Categories and showcasing them in the product. In this Part II we will describe the ML Categorization work in more detail.
Airbnb 2022年发布了Categories，这是一个以浏览为重点的产品，允许用户通过浏览围绕一个共同主题的房屋系列来寻找灵感，例如湖滨、乡村、高尔夫、沙漠、国家公园、冲浪等。在我们的分类博客系列的第一部分中，我们介绍了创建分类和在产品中展示它们的高层次方法。在第二部分中，我们将更详细地描述ML分类的工作。
Throughout the post we use the Lakefront category as a running example to showcase the ML-powered category development process. Similar process was applied for other categories, with category specific nuances. For example, some categories rely more on points of interests, while others more on structured listing signals, image data, etc.
在整个文章中，我们以湖滨分类为例，展示了以ML为动力的分类开发过程。类似的过程也适用于其他类别，但有类别的细微差别。例如，有些类别更依赖于兴趣点，而其他类别则更依赖于结构化列表信号、图像数据等。
Category Definition
类别定义
Category development starts with a product-driven category definition: “Lakefront category should include listings that are less than 100 meters from the lake”. While this may sound like an easy task at first, it is very delicate and complex as it involves leveraging multiple structured and unstructured listing attributes, points of interest (POI...