OVSegDT: Segmenting Transformer for Open-Vocabulary Object Goal Navigation

TL;DR

OVSegDT introduces a transformer-based policy for open-vocabulary object goal navigation, improving generalization, safety, and efficiency with novel components grounded in segmentation and adaptive loss modulation.

Contribution

The paper proposes OVSegDT, a lightweight transformer with semantic grounding and entropy-adaptive loss, reducing training complexity and enhancing performance on unseen categories.

Findings

Matches performance on unseen categories to seen ones.

Achieves state-of-the-art results: 40.1% SR, 20.9% SPL on val unseen.

Reduces collision count by two times.

Abstract

Open-vocabulary Object Goal Navigation requires an embodied agent to reach objects described by free-form language, including categories never seen during training. Existing end-to-end policies overfit small simulator datasets, achieving high success on training scenes but failing to generalize and exhibiting unsafe behaviour (frequent collisions). We introduce OVSegDT, a lightweight transformer policy that tackles these issues with two synergistic components. The first component is the semantic branch, which includes an encoder for the target binary mask and an auxiliary segmentation loss function, grounding the textual goal and providing precise spatial cues. The second component consists of a proposed Entropy-Adaptive Loss Modulation, a per-sample scheduler that continuously balances imitation and reinforcement signals according to the policy entropy, eliminating brittle manual phase switches. These additions cut the sample complexity of training by 33%, and reduce collision count in two times while keeping inference cost low (130M parameters, RGB-only input). On HM3D-OVON, our model matches the performance on unseen categories to that on seen ones and establishes state-of-the-art results (40.1% SR, 20.9% SPL on val unseen) without depth, odometry, or large vision-language models. Code is available at https://github.com/CognitiveAISystems/OVSegDT.