New Methods for Generalizable & Robust Time-Series Analysis
We all want to know the future—especially in industry, where every decision hinges on a prediction. Sales forecasts. Financial forecasts. Supply chain forecasts. In many ways, data science is the business of forecasting. Yet, despite our best models, our predictions often fail when we need them most. Why do our forecasts break down, and how can we make them more reliable?
Join us for a talk with Dr. Aditya Prakash (Georgia Tech) as he shares cutting-edge research on AI-driven time series forecasting—techniques already used by Walmart, Facebook, and other industry leaders to improve accuracy, handle uncertainty, and build more robust predictive models. 🚀
🗓 Date: March 26, 2025
⏰ Time: 2:00 – 3:00 PM CST
📍 Location: Online (link provided upon registration)
🎤 Speaker: Dr. B. Aditya Prakash, Georgia Tech
Cost: Nothing, nada, zilch. $0
📦 The Challenge: Forecasting in a Chaotic World
Your company’s supply chain team runs the latest machine learning model to forecast next quarter’s demand. The numbers look promising. But are they right?
Financial analysts face the same problem. Their sales forecasting model suggests a stable revenue trajectory, but how does it handle sudden market shifts?
Predictions drive billions in decision-making across industries—from inventory planning to risk management to financial forecasting—but most models fail to quantify uncertainty. In an unpredictable world, companies need models that don’t just predict but also explain risk, adjust in real-time, and adapt across industries.
🔍 The Breakthrough: Smarter, More Generalizable Forecasting
Dr. B. Aditya Prakash, a leading AI researcher at Georgia Tech, has been tackling this problem head-on. His work bridges deep sequential models with Gaussian processes to build forecasts that aren’t just accurate—they’re calibrated, adaptable, and built for real-world complexity.
In this talk, he’ll reveal:
✅ How to make multi-variate forecasts more reliable using uncertainty-aware models.
✅ How LLMs and multi-domain pre-trained models can enhance generalizability for forecasting.
✅ The power of real-time data revision—ensuring that supply chain, financial, and sales forecasts evolve as new information flows in.
🎙 About the Speaker
Dr. B. Aditya Prakash is an Associate Professor at Georgia Tech and a leading researcher in Data Science, Machine Learning, and AI, focusing on large-scale networks and time-series applications. He has published over 100 research papers, holds two U.S. patents, and has received multiple prestigious awards, including the NSF CAREER Award and Facebook Faculty Awards. His tools have been used by organizations like ORNL, CDC, and Walmart.
We use cookies to improve your experience and to help us understand how you use our site. Please refer to our cookie notice and privacy statement for more information regarding cookies and other third-party tracking that may be enabled.
© 2025 Houston INFORMS