课程定位
训练数据的筛选、去重、合成、归因与版权。 本页保留官方档案与入口,具体内容与考核以当季官网为准。
主要内容
本页为延伸课程速览,主要内容见下方官方描述。
先修
见官方描述(延伸课程暂未单独整理先修)。
考核与工作量
以当季官网为准(延伸课程暂未单独整理考核)。
适合谁
对「Data for AI」方向感兴趣的人。
学习建议
先读官方描述和课程官网,再判断是否适合自己。
相关课程
官方课程描述(英文原文)
CS525 surveys the landscape of data for AI with a focus on contemporary learning problems such as training language models or multimodal models. Students will learn about important datasets and common data processing methods including filtering and deduplication. Further topics will include synthetic data, data attribution, and environments for reinforcement learning. The course will also cover ethical and legal aspects of training data such as copyright and privacy. Over the course of the class, students will build a training set for a learning problem of their choice. The class will consist of faculty lectures, student presentations, and guest lectures.