Natasha Balac

About

Natasha Balac’s research lies at the intersection of artificial intelligence, machine learning, and robotics, with a particular focus on enabling autonomous systems to plan and act effectively in unpredictable, real-world environments. Her pioneering work addresses a critical challenge: how can a robot or agent learn the true effects of its actions when the environment itself introduces noise and variability? Balac’s most influential paper, “Using Regression Trees to Learn Action Models” (2002, 11 citations), introduced a novel application of regression trees to model how environmental conditions—like icy roads or rough terrain—alter the outcomes of planned actions. This work built directly on her earlier foundational paper, “Learning Action Models for Navigation in Noisy Environments” (2000, 4 citations), which presented the ERA approach. By explicitly teaching planners to account for environmental interference, Balac’s contributions have helped bridge the gap between theoretical planning models and practical robotic navigation. Her research remains highly relevant for students and engineers working on autonomous vehicles, field robotics, and any system that must make robust decisions under uncertainty.

Research Focus

Key Achievements

Top Papers

1. 1
   Using regression trees to learn action models

   11 citations · 2002
2. 2
   Learning Action Models for Navigation in Noisy Environments

   4 citations · 2000

Key Collaborators

Contact & Links