Chapter
The Importance of Good Software Engineering for Small Machine Learning Systems
1:11:14 - 1:21:36 (10:22)
The machine learning model is only a small part of the entire software system required to build an effective machine learning solution. Good software engineering is crucial for managing the change process and deploying the solution to achieve real impact.
Clips
AI can be used in sectors like manufacturing, agriculture, healthcare, and transportation, which have many untapped opportunities and untapped potential for growth.
1:11:14 - 1:14:29 (03:15)
Summary
AI can be used in sectors like manufacturing, agriculture, healthcare, and transportation, which have many untapped opportunities and untapped potential for growth. McKinsey and PWC have estimated $13-16 trillion of global economic growth using AI.
ChapterThe Importance of Good Software Engineering for Small Machine Learning Systems
Episode#73 – Andrew Ng: Deep Learning, Education, and Real-World AI
PodcastLex Fridman Podcast
Early small-scale machine learning projects can help teams gain faith and learn about the technology.
1:14:29 - 1:17:46 (03:16)
Summary
Early small-scale machine learning projects can help teams gain faith and learn about the technology. Starting small can also lead to successful collaborations with other teams and improve the accuracy of larger systems.
ChapterThe Importance of Good Software Engineering for Small Machine Learning Systems
Episode#73 – Andrew Ng: Deep Learning, Education, and Real-World AI
PodcastLex Fridman Podcast
Good software engineering is crucial to building a successful small machine learning system beyond just the machine learning model, managing the change process, and deploying it effectively.
1:17:46 - 1:21:36 (03:50)
Summary
Good software engineering is crucial to building a successful small machine learning system beyond just the machine learning model, managing the change process, and deploying it effectively. Dealing with test set distributions that are significantly different from the training set distribution in machine learning is still a challenge for academia.