AI For All: Enhancing Digital Financial Services using Data Science
summary
Mifos demonstrates a working pipeline: OCR (CNN) extracts three months of bank statements, NLP (bag‑of‑words / TF‑IDF) categorizes line items, and gradient‑boosted trees or random forests produce credit scores exposed as REST microservices on Kubernetes; the integrated system cuts loan onboarding from days to 15–20 minutes. The community publishes a white paper, shipped modules include credit scorecard, chatbot and vision PPI, and next steps prioritize federated learning, explainability and security for cross‑bank training.
AI-written summary · about
chapters (12)
- 0:00Start
- 0:47Al For All Working Group Members
- 1:26Al for All Revisted - Why Data Science Matters
- 4:01Mifos - "Al for ALL" - Proposed Components
- 5:14White Paper Introduction
- 7:20Use Cases Supported by the Ecosystem
- 7:59Current Open Source Al Tooling develop by Community
- 8:11Chatbot
- 11:38Credit Scoring
- 13:25Next Up: Federated Learning
- 14:00Al for Banking - Some of the Use Cases
- 15:30Algorithm by Use Case What and Why
related talks
01h 03mAI Innovation for a Smarter Financial Ecosystem
31mData Science for Financial Inclusion
29mAI - Digital (r)evolution in banking
22mUsing data to innovate credit decisioning for Small-to-Medium Enterprises
57mAI has been achieved internally - The Journey
01h 30mAI in Finance From Insight to Action
34mTrends and Practical Applications of AI & ML in FinTech Industry
39mMachine Learning in Financial Credit Risk Assessment