Want a Good Machine Learning Model? Start with Data Processing
summary
The talk argues that ML model quality depends largely on data processing and cleaning, not just algorithms. It compares Batch and Streaming Big Data architectures for data ingestion, augmentation and creating training sets, describes open-source tools to implement those pipelines, and explains how public cloud services (AWS, GCP) can help optimize the workflows.
AI-written summary · about
chapters (15)
- 0:00Welcome
- 1:29What is Big Data
- 6:45Machine learning and data
- 9:13Data lifecycle
- 12:44Key data processing
- 18:27ETL: Extraction and transformation
- 21:54Batch and streaming architectures
- 25:03Automation in processing
- 27:12Integration with ML
- 28:59Data model as core
- 31:06Data quality and testing
- 33:36Data governance
- 34:40Bias in data and ML
- 37:11Less is more in tech
- 38:53Q&A
related talks
41mForget becoming a Data Scientist, become a Machine Learning Engineer instead
06mMachine learning is about your data and deployment
41mInfrastructure for ML Applications
50mIntegrating Machine Learning - from ingestion to persistence
27mA Look Inside Applied Machine Learning
49mTraining Models
11mRecent trends in data and machine learning technologies
37mBig Data Analytics, Machine Learning, Real Time Processing