Search papers, labs, and topics across Lattice.
This paper tackles the challenge of balancing exploration and stability in self-improving Vision-and-Language Navigation (VLN) agents. They introduce Stability-Diversity Balance (SDB), a method that expands each decision step into multiple latent behavioral hypotheses via controlled shifts in instruction-conditioned hidden states, and then aggregates them based on reliability. By explicitly regularizing hypothesis interactions, SDB stabilizes self-improvement without discarding training signals, leading to improved performance on VLN benchmarks.
Self-improving navigation agents can get lost in their own exploration unless you carefully balance behavioral diversity with learning stability, as shown by a new method that outperforms prior VLN models.
In vision-and-language navigation (VLN), self-improvement from policy-induced experience, using only standard VLN action supervision, critically depends on balancing behavioral diversity and learning stability, which governs whether the agent can extract a reliable learning signal for improvement. Increasing behavioral diversity is necessary to expose alternative action hypotheses but can destabilize policy-induced learning signals, whereas overly conservative stability constraints suppress exploration and induce early commitment, making reliable self-improvement difficult. To address this challenge, we propose Stability-Diversity Balance (SDB), a plug-and-play mechanism for balanced self-improvement in VLN. SDB expands each decision step into multiple latent behavioral hypotheses by applying controlled shifts in the instruction-conditioned hidden states, and then performs reliability-aware soft evaluation and aggregation to retain diverse yet instruction-consistent alternatives during learning. An explicit regularizer further constrains hypothesis interactions, preventing excessive drift or premature collapse of hypothesis diversity and stabilizing self-improvement without discarding training signals. Experiments on R2R, SOON, and REVERIE show consistent improvements; for example, on REVERIE val-unseen, SDB improves SPL from 33.73 to 35.93 and OSR from 51.07 to 54.25.
