Today, according to McKinsey, 85% of ML models fail, leading to loss of time, revenue and business impact. Understanding common risk factors for model failure has a direct impact on improving the visibility and predictability of ML development cycles, especially for ADAS & AV applications. In this session Duncan Curtis, Chief Product Officer at Sama (formerly of Zoox, Aptiv, and Google) reveals the criteria executives can use to accelerate their ML models’ path to production — and keep them there.
Duncan will share best practices such as how to:
- Improve predictability of ADAS & AV algorithms
- Add transparency to the process of understanding when and why models fail
- Achieve faster iteration times for ML models
- Lower the Total Cost of Ownership (TCO) of your ML programs
&nsbp;
