StateLinFormer: Stateful Training Enhancing Long-term Memory in Navigation
This addresses the challenge of persistent memory for navigation intelligence, offering a novel training approach to enhance long-term adaptation, though it appears incremental as it builds on existing linear-attention and Transformer methods.
The paper tackles the problem of limited long-term memory in navigation models by introducing StateLinFormer, a linear-attention model trained with a stateful memory mechanism that preserves recurrent states across training segments, enabling long-horizon memory retention and outperforming stateless and Transformer baselines in MAZE and ProcTHOR environments.
Effective navigation intelligence relies on long-term memory to support both immediate generalization and sustained adaptation. However, existing approaches face a dilemma: modular systems rely on explicit mapping but lack flexibility, while Transformer-based end-to-end models are constrained by fixed context windows, limiting persistent memory across extended interactions. We introduce StateLinFormer, a linear-attention navigation model trained with a stateful memory mechanism that preserves recurrent memory states across consecutive training segments instead of reinitializing them at each batch boundary. This training paradigm effectively approximates learning on infinitely long sequences, enabling the model to achieve long-horizon memory retention. Experiments across both MAZE and ProcTHOR environments demonstrate that StateLinFormer significantly outperforms its stateless linear-attention counterpart and standard Transformer baselines with fixed context windows. Notably, as interaction length increases, persistent stateful training substantially improves context-dependent adaptation, suggesting an enhancement in the model's In-Context Learning (ICL) capabilities for navigation tasks.