Why Chess Teaches Us A Lot About Forecasting: Deep Learning Analytics and Demand Planning Go Head To Head
Chess tells us a lot about the possibilities for forecasting. Thanks to Deep Learning, algorithms are now able to outsmart even the greatest Chess Grandmasters. We know therefore, that for certain functions, machines are vastly superior to humans. With this in mind, we should all be pretty excited about what they can do for forecast accuracy. Research says that Artificial Intelligence promises to reduce forecasting errors by 30 to 50%, reduce lost sales from product unavailability by up to 65%, and reduce overall inventory by 20 to 50%. But we cannot realize these kinds of gains until demand planners and data scientists compare their strengths and limitations to find alignment. In this thought provoking session, a top MBA demand planner and a globally ranked Deep Learning expert compare their demand forecasting superpowers, and their weaknesses. They will also propose two key areas that Machine Learning can solve with practical takeaways for your own organization.Key learnings:
- A story about chess and how AI can forecast better than humans
- Real demand planner problems where Machine Learning can help
- A framework for ranking Machine Learning forecasting applications based on industry & company characteristics
Vivek is a supply chain professional with forecasting and S&OP experience in CPG, Food Beverage and Oil & Gas industries. After his MBA, Vivek worked in strategy, analytics and reporting where his data analytic solutions drove millions of dollars in cost savings in supply chain operations. Then he moved into S&OP where he has worked in a $30M production plant relocation, SAP APO implementation, new product launches and creating processes for sales to be involved in demand forecasting. He has also worked with a food & beverage startup to set up their demand forecasting, supply planning and MRP/procurement systems.Nima Shahbazi
Nima is co-founder of Deepnify, helping companies reduce inventory with predictive deep learning technology. He has won data mining competitions for Rossmann, Home Depot, ACM and Two Sigma. On the global Kaggle data science competition platform with 1M members, Nima ranked #19 and is one of only eighty-eight data scientists to achieve Grandmaster status. He is a PhD Candidate in the Data Mining and Database Group at York University. He previously worked in big data analytics, specifically on Forex and Stock Market predictions. Nima recently won the prestigious ACM Recommendation Challenge and has presented his winning forecasting techniques at ReWork Deep Learning Summit and ACM WSDM.