Bytesized AI # 1
Download as PPTX, PDF0 likes27 views
AI refers to building systems that can perform intelligent tasks, while machine learning (ML) refers to systems that can learn from experience. Natural language processing (NLP) involves systems that can understand human language. There are three main types of machine learning: supervised learning uses human-annotated data to train models, semi-supervised learning uses a small amount of annotated data and a large amount of unannotated data, and unsupervised learning uses only unannotated data. The Atticus Project is open-sourcing a labeled legal contracts dataset to improve the accuracy of AI-based contract review tools, where "labeled" means volunteers have marked relevant legal provisions under attorney supervision.
1 of 7
Download to read offline
More Related Content
What's hot (20)
Ad
Recently uploaded (20)
Ad
Bytesized AI # 1
- 1. AI / ML / NLP Basic Terminology © 2020 The Atticus Project
- 2. Terminology AI NLPML Building systems that can do intelligent things. AI that can learn from experience. AI that can understand language. AI that can learn how to understand language.
- 3. All machine learning and natural language processing need data to train. AI and Data Annotated Data Unannotated Data
- 4. Annotated Data Data that is tagged. Also called "labelled data".
- 5. Unannotated Data Data that is not tagged or labelled.
- 6. 3 Main Types of Machine Learning Supervised ● Learns from human-annotated data. Semi-supervised ● Learns from a small amount of annotated data and a large amount of unannotated data. Unsupervised ● Learns from unannotated data.
- 7. Atticus and AI The Atticus Project is open-sourcing a dataset of lLabeled legal contracts for AI training purposes so that contract review tools can improve their accuracy and comprehensiveness. "Labeled" in this context means marking the relevant legal provisions (e.g. most favored nation, exclusivity) in a legal contract by Atticus volunteers under the supervision of experienced attorneys. collection