Practical AI The $1 trillion dollar ML model đź’µ
Jan 11, 2021
Madhurima Khandelwal, Vice President and Head of American Express AI Labs, leads the conversation on their groundbreaking $1 trillion machine learning model. Discover how this model automates over 8 billion decisions and tackles real-time fraud detection in a digital-first world. Madhurima shares insights on integrating AI to enhance customer experience and operational efficiency, while stressing the importance of ethical AI practices. The future of finance with AI promises innovation that will reshape credit decisioning and fraud prevention.
AI Snips
Chapters
Transcript
Episode notes
AI Model Necessity
- Evaluate if a problem truly needs AI/ML.
- Simpler solutions might suffice, so consider data and model explainability.
Data Science vs. AI/ML
- Not all data science problems necessitate AI/ML models.
- Deep learning models have higher deployment costs than other methods.
Model Evaluation Strategy
- Build both a basic segmentation and an AIML model for a new problem.
- Compare their performance to justify the added complexity of AIML if necessary.