| I&T Solution |
AI-driven Auto Inspection Ranking System with Continuous Learning Capability for Lift & Escalator (L&E)
(REF: S-1726) |
| Trial Project |
|
| Solution Feature |
- Data-driven Risk Factor Identification and Weighting: A comprehensive analysis of available L&E data will be conducted to identify the most critical features and their relative importance for L&E inspections. This will empower the creation of a transparent and standardized risk assessment model, facilitating interpretable risk insights and explainable decision-making.
- Standardized Data Pipeline and Transformation: A standardized data input and transformation pipeline will be designed and implemented to streamline the data collection and AI analysis process. This will enable automated risk ranking without human intervention, increasing efficiency and accuracy.
- AI-driven Recommender Engine: A recommender engine will be developed, powered by a trained AI model, to analyze data, rank risks, and prioritize inspections for L&Es. This engine will provide actionable insights to support informed decision-making.
- Continuous Refinement: The recommender engine will be equipped with a reinforcement learning mechanism, allowing it to continuously refine the risk assessment model and adapt to evolving risk patterns. This ensures that the system remains accurate and effective over time, providing a future-proof solution for L&E inspections.
- Seamless Integration with Other Systems: The developed engine with well-defined APIs will ensure seamless integration with other systems, facilitating a unified and efficient data workflow.
|
| Trial Application and Expected Outcome |
- Risk features of L&Es will be identified and quantified to provide a standardized and explainable L&E risk assessment model.
- An efficient and standardized data pipeline and transformation framework will transform input data into the format suitable for the recommender engine.
- An accurate AI-based recommender engine, targeted to achieve an accuracy of 95% or above, will leverage available L&E data to estimate risk scores and prioritize L&E inspections, streamlining the allocation of resources towards higher-risk entities.
- The engine, with the embedded reinforcement learning mechanism, will continuously refine the risk ranking based on the executed inspection results and various big data.
- The recommender engine will seamlessly integrate with other systems for further data workflow and processing.
|
| Additional Solution Information |
EMSD Lift and Escalator_Addtional Information_20240624_v0.2.pdf
|
| Info on I&T Solution Provider |
| Solution Provider | : | Deloitte Innovation Investment (Hong Kong) Limited | | Address | : | 2/F, Harbour View 2, 16E Science Park East Avenue, Hong Kong Science Park, Sha Tin | | Contact Person | : | Wong, Anthony Chun Sang |
| Position | : | Senior Manager | | Tel | : | 22387746 | | Email | : |
anthwong@deloitte.com.hk |
|
