Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2507.19947 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916870582435840 |
|---|---|
| author | Sitdhipol, Supawich Sukprasongdee, Waritwong Chuangsuwanich, Ekapol Tse, Rina |
| author_facet | Sitdhipol, Supawich Sukprasongdee, Waritwong Chuangsuwanich, Ekapol Tse, Rina |
| contents | Fusing information from human observations can help robots overcome sensing limitations in collaborative tasks. However, an uncertainty-aware fusion framework requires a grounded likelihood representing the uncertainty of human inputs. This paper presents a Feature Pyramid Likelihood Grounding Network (FP-LGN) that grounds spatial language by learning relevant map image features and their relationships with spatial relation semantics. The model is trained as a probability estimator to capture aleatoric uncertainty in human language using three-stage curriculum learning. Results showed that FP-LGN matched expert-designed rules in mean Negative Log-Likelihood (NLL) and demonstrated greater robustness with lower standard deviation. Collaborative sensing results demonstrated that the grounded likelihood successfully enabled uncertainty-aware fusion of heterogeneous human language observations and robot sensor measurements, achieving significant improvements in human-robot collaborative task performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_19947 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Spatial Language Likelihood Grounding Network for Bayesian Fusion of Human-Robot Observations Sitdhipol, Supawich Sukprasongdee, Waritwong Chuangsuwanich, Ekapol Tse, Rina Robotics Computation and Language Information Theory Machine Learning Systems and Control Fusing information from human observations can help robots overcome sensing limitations in collaborative tasks. However, an uncertainty-aware fusion framework requires a grounded likelihood representing the uncertainty of human inputs. This paper presents a Feature Pyramid Likelihood Grounding Network (FP-LGN) that grounds spatial language by learning relevant map image features and their relationships with spatial relation semantics. The model is trained as a probability estimator to capture aleatoric uncertainty in human language using three-stage curriculum learning. Results showed that FP-LGN matched expert-designed rules in mean Negative Log-Likelihood (NLL) and demonstrated greater robustness with lower standard deviation. Collaborative sensing results demonstrated that the grounded likelihood successfully enabled uncertainty-aware fusion of heterogeneous human language observations and robot sensor measurements, achieving significant improvements in human-robot collaborative task performance. |
| title | Spatial Language Likelihood Grounding Network for Bayesian Fusion of Human-Robot Observations |
| topic | Robotics Computation and Language Information Theory Machine Learning Systems and Control |
| url | https://arxiv.org/abs/2507.19947 |