Several strategies and technologies are directed at addressing the constantly mutating problem of fraud. Most recently, advances in machine learning are being harnessed to detect needles of fraud in the haystack of digital businesses. However, as with all technical advances, fraud analysts are fast realizing that while ML gives them a significant advantage, it still has its bottlenecks, in terms of requiring diverse and significant data to make the models effective, and the complexity of fine tuning the models.
In this session, Kedar Samant, CTO and co-founder of Simility, and Swastik Bihani, VP of Product Management of Simility, will discuss fraud strategies that are evolving beyond machine learning.
You will learn:
- Brief discussion on ML and how it helps with fraud detection
- Limitations of machine learning and unsupervised learning
- Emerging approaches such as adaptive data ingestion and dynamic data ontology
- Steps and approaches to build a strategy beyond machine learning