14 Nov
|
National Research Council Canada
|
Canada
14 Nov
National Research Council Canada
Canada
Apply deep reinforcement learning to driver attention forecasting as an NRC Postdoctoral Fellow in Hamilton. This is a full time onsite appointment.
The project investigates reward structures governing human-driver gaze behavior through graph-based inverse reinforcement learning. You will analyze changing scanpaths across weather scenarios and use spectral graph analysis to quantify attention shifts. Deep learning models will forecast future gaze trajectories and driving actions, creating a complete pipeline from metric extraction to prediction. The two-year fellowship is supervised by an NRC researcher in a laboratory and includes participation in additional multidisciplinary projects. Candidates contribute to research spanning reinforcement learning, vision, and human factors applications.
Key Responsibilities:
- Infer reward structures governing driver gaze behavior
- Quantify attention shifts across weather scenarios
- Build scanpath analysis and prediction pipelines
- Develop models for future driving-action forecasting
- Support additional projects within a multidisciplinary team
Requirements:
- PhD in Computer Engineering, Computer Science, or related discipline
- Two to six years of significant deep learning experience
- Advanced architecture experience including deep reinforcement learning
- Programming experience in Python or C++
- Experience developing maintainable code with listed machine learning libraries
Use reinforcement learning, graph theory, and vision expertise to model driver attention and forecast actions at the NRC.
📌 Deep Reinforcement Learning Postdoctoral Researcher (Canada)
🏢 National Research Council Canada
📍 Canada