FLYNN: Robust Neural Network for Robot Navigation using Fly Brain Topology

2026-06-21

AI Summary

This work addresses the limited robustness of deep neural networks under out-of-distribution data or sensory deprivation, a capability where biological systems excel. To bridge this gap, the authors propose FLYNN—a recurrent neural network grounded in the complete synaptic connectome of the Drosophila melanogaster brain—marking the first effort to directly embed real neural connectivity topology into a trainable visual navigation model. Evaluated in the MuJoCo simulation environment, FLYNN maintains effective navigation performance under extreme conditions such as partial or complete visual loss without any additional training. It matches the performance of comparably sized artificial networks and significantly outperforms specialized baseline models trained explicitly for such scenarios, demonstrating exceptional robustness and modular representational properties.

Abstract

While deep learning models achieve state-of-the-art performance in complex tasks, they remain brittle when faced with new environments or sensory deprivation. In contrast, biological systems exhibit remarkable tolerance to these challenges. We address this vulnerability by developing a recurrent neural network (RNN) whose architecture is directly derived from the synaptic-resolution brain connectome of the fruit fly Drosophila melanogaster. We demonstrate the feasibility of training the fly connectome neural network (FLYNN) to perform vision-based navigation in MuJoCo, achieving performance comparable to modern hand-crafted networks of similar parameter counts. Crucially, FLYNN exhibits superior resistance to out-of-distribution (OOD) data and tolerance to sensory loss without further training. It remained functional even under total vision loss while hand-crafted networks largely failed, even when specifically trained with camera dropout. Principal Component Analysis (PCA) of the internal state of FLYNN suggests that it exhibits a particularly high degree of representational modularity, which might be related to its robustness. Our work provides a new direction for designing resilient artificial agents following the topology of biological brains.

Problem

• robustness
• out-of-distribution
• sensory deprivation
• neural network brittleness
• robot navigation

Innovation

• fly brain connectome
• robust neural network
• out-of-distribution robustness
• sensory deprivation tolerance
• representational modularity