← Back to home

Research

I work on brain-inspired computer vision, focusing on recurrent top-down/bottom-up learning, local learning rules, and robust evaluation under domain shift.

Current directions

• Recurrent TD–BU modulation: task-guided refinement during inference.
• Biologically plausible learning: local learning rules (e.g., Counter-Hebbian updates).
• Robust evaluation: multi-seed reliability and stress tests under distribution shift.

Keywords

Top-down / bottom-up vision, recurrent inference, visual attention, biologically plausible learning, domain shift.