Reimagining service delivery in the agentic era with Google Public Sector
cloud.google.com·21h ago
AI Research
arXiv CS.AI·Jul 8·5 min read
MambaGaze: Bidirectional Mamba with Explicit Missing Data Modeling for Cognitive Load Assessment from Eye-Gaze Tracking Data
TL;DR
Cognitive load assessment from eye-gaze tracking data is challenging due to missing data. MambaGaze, a bidirectional model, explicitly addresses this issue by incorporating missing data modeling techniques.
✦ Why It Matters
Implement MambaGaze in your eye-tracking studies to improve cognitive load assessments and data reliability.
Key Takeaways
- Utilize MambaGaze for effective cognitive load monitoring in real-time applications.
- Implement XMD encoding to handle data uncertainty in eye-tracking signals.
- Leverage bidirectional Mamba-2 for efficient modeling of temporal dependencies.
- Deploy on NVIDIA Jetson Orin for low-power, high-performance inference.
How It Works
MambaGaze employs XMD encoding to enhance raw eye-tracking data by adding observation masks and time-deltas, which explicitly model the uncertainty caused by missing data. The bidirectional Mamba-2 model processes this enriched data to capture long-range temporal dependencies efficiently, maintaining linear computational complexity.