Real-time machine learning: challenges and solutions

[Twitter discussion, LinkedIn]

[ Twitter discussion , LinkedIn ] Updates Jan 3, 2023: Update the online features section to differentiate between real-time features and near real-time features. If you’re interested in this topic, my book Designing Machine Learning Systems (O’Reilly, June 2022) covers online prediction and continual learning in much more detail. Real-time machine learning is the approach of using real-time data to generate more accurate predictions and adapt models to changing environments. A year ago, I wrote a post on how machine learning is going real-time . The post must have captured many data scientist

saved by

• Samuel C
• Jonny Hsu
• Aryan Musharaf

related reading

• Machine learning is going real-timehuyenchip.com
• The Continual Learning Problemjessylin.com
• You can’t imitation-learn how to continual-learn — LessWronglesswrong.com
• Introduction to streaming for data scientistshuyenchip.com
• Scaling is subtler than it seemsberen.io
• Q-learning is not yet scalableseohong.me
• Why We Need Continual Learning | Andreessen Horowitza16z.com
• A Call to Build Models Like We Build Open-Source Softwarecolinraffel.com
• The Unreasonable Difficulty of Time Series Forecastingsuzyahyah.github.io
• Keep the Tokens Flowing: Lessons from 16 Open-Source RL Librarieshuggingface.co
• Recent Advances in Language Model Fine-tuningruder.io
• Low Latency and Model Training at Modalrhea24.github.io