How to Extract Value out of ML in the Process Industry
- Save to Calendar
- 10/03/2021 - 11/03/2021 - 7:00 pm - 12:15 pm IST
Learn How to Start creating Value out of manufacturing data by defining the (achievable) problem statement and objectives. How to select the right ML methods & workflows, which techniques to apply and build the (practical) solution. How to implement ML solution and extract value.
To extract the ROI out of ML-AI implementation, it is important to take a ‘Practical and Simplistic’ approach through Value creation, Value Solution and Value Production.
- Value Creation: Machine learning is not like magic to solve each and every problem, meet the end-objective as desired always. It needs a ‘Design Thinking’ kind of approach to first define the problem statement at hand and derive what is achievable and to what an extend. For example, do we want to monitor the health of the assets or predict failure or look into the performance/efficiency of the assets/processes. The success of the technology depends on its right application and the value it creates at the end of the day. It is like setting the stage for leveraging the benefits of ML-AI techniques. This e-meet will showcase few approaches on how to classify the problem from a data analytics perspective, what are the pre-requisites and how to define the end objectives.
- Value Solution: Data Science as a technology is continuously evolving and we come across hundreds of algorithms, tools, solutions, etc. It starts by selecting the right approach to a ‘Well-defined Problem’ in the Value creation stage. The approach is essentially considering the constraints in data, applying the right process understanding, which leads to the selection of the right methods. The solutioning sometimes become a tedious task and we end up trying and testing various permutation and combination (for e.g. Linear vs non-linear algorithms). What if, we have a matrix of methods/approaches/best practices, which addresses at least 70-80% of the problem statements we encounter in the process industry (descriptive, diagnostic, predictive, prescriptive analytics). What if we have templates/libraries of such methods, approaches, so that solutioning is not a time-consuming or skill-dependent/tool-dependent activity. This e-meet will showcase a matrix of methods, approaches and solutions with Use cases in the process industry.
- Value Production: The success of the technology can only be realized if it can be practically deployed and is operational. It needs the right solution with a simple workflow, which can implement the methods employed during the solutioning stage. It needs a lot of pre-requisites right from data collection to deployment. This e-meet will showcase a few of the ‘Deployable’ methods and solutions with use cases.
Meet our experts and get insights into Approach, Methods, Solutions, and how we enable it through our ‘Pilot Guided Analytics’ framework.
Who should attend:
- Process R&D Heads
- Production / Operations Head
- Maintenance Heads
- Plant heads
- Technology leaders
- Data Science leaders
- Data Analysts