Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional Scenes

This study delves into verifying the applicability of object co-occurrence information generated by a large-scale language model (LLM) to enhance a robot’s spatial ability to understand objects in the real world. Co-occurrence information is crucial in enabling robots to perceive and navi...

Full description

Saved in:
Bibliographic Details
Main Authors: Kenta Gunji, Kazunori Ohno, Shuhei Kurita, Ken Sakurada, Ranulfo Bezerra, Shotaro Kojima, Yoshito Okada, Masashi Konyo, Satoshi Tadokoro
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10786984/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846118313225093120
author Kenta Gunji
Kazunori Ohno
Shuhei Kurita
Ken Sakurada
Ranulfo Bezerra
Shotaro Kojima
Yoshito Okada
Masashi Konyo
Satoshi Tadokoro
author_facet Kenta Gunji
Kazunori Ohno
Shuhei Kurita
Ken Sakurada
Ranulfo Bezerra
Shotaro Kojima
Yoshito Okada
Masashi Konyo
Satoshi Tadokoro
author_sort Kenta Gunji
collection DOAJ
description This study delves into verifying the applicability of object co-occurrence information generated by a large-scale language model (LLM) to enhance a robot’s spatial ability to understand objects in the real world. Co-occurrence information is crucial in enabling robots to perceive and navigate their surroundings. LLM can generate object co-occurrence information based on the learned representations acquired from the learning process. However, the challenge lies in determining whether the co-occurrence gleaned from linguistic data can effectively translate to real-world object relationships, a concept yet to be thoroughly examined. After providing category information about a specific situation, this paper compares and evaluates the co-occurrence degree (co-occurrence coefficient) output by gpt-4-turbo-2024-04-09 (GPT-4) against the object pair data from the ScanNet v2 dataset. The results revealed that GPT-4 achieved a high recall of 0.78 across various situation categories annotated by ScanNet v2, although its precision was relatively low at an average of 0.29. The root mean square error of the co-occurrence coefficient was 0.31. While GPT-4 tends to output slightly higher co-occurrence coefficients, it effectively captures the overall co-occurrence patterns observed in the ScanNet v2 dataset. GPT-4 produced co-occurrence information for more objects than those available in ScanNet v2 while covering co-occurrences among objects within ScanNet v2. These results demonstrate that integrating co-occurrence data from different sources could enhance the ability to recognize real-world objects and potentially strengthen robot intelligence.
format Article
id doaj-art-f6acd24f2ee54513a1f4ba798f767b6b
institution Kabale University
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-f6acd24f2ee54513a1f4ba798f767b6b2024-12-18T00:01:50ZengIEEEIEEE Access2169-35362024-01-011218657318658510.1109/ACCESS.2024.351447310786984Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional ScenesKenta Gunji0https://orcid.org/0000-0002-7924-2437Kazunori Ohno1https://orcid.org/0000-0003-3958-2901Shuhei Kurita2https://orcid.org/0000-0001-7415-3120Ken Sakurada3https://orcid.org/0000-0003-3386-1547Ranulfo Bezerra4https://orcid.org/0000-0001-6017-5724Shotaro Kojima5https://orcid.org/0000-0003-0042-1764Yoshito Okada6https://orcid.org/0000-0003-3830-079XMasashi Konyo7https://orcid.org/0000-0002-6826-9722Satoshi Tadokoro8https://orcid.org/0000-0002-5571-4276Graduate School of Information Sciences, Tohoku University, Sendai, JapanGraduate School of Information Sciences, Tohoku University, Sendai, JapanNational Institute of Informatics, Tokyo, JapanGraduate School of Informatics, Kyoto University, Kyoto, JapanGraduate School of Information Sciences, Tohoku University, Sendai, JapanGraduate School of Information Sciences, Tohoku University, Sendai, JapanGraduate School of Information Sciences, Tohoku University, Sendai, JapanGraduate School of Information Sciences, Tohoku University, Sendai, JapanGraduate School of Information Sciences, Tohoku University, Sendai, JapanThis study delves into verifying the applicability of object co-occurrence information generated by a large-scale language model (LLM) to enhance a robot’s spatial ability to understand objects in the real world. Co-occurrence information is crucial in enabling robots to perceive and navigate their surroundings. LLM can generate object co-occurrence information based on the learned representations acquired from the learning process. However, the challenge lies in determining whether the co-occurrence gleaned from linguistic data can effectively translate to real-world object relationships, a concept yet to be thoroughly examined. After providing category information about a specific situation, this paper compares and evaluates the co-occurrence degree (co-occurrence coefficient) output by gpt-4-turbo-2024-04-09 (GPT-4) against the object pair data from the ScanNet v2 dataset. The results revealed that GPT-4 achieved a high recall of 0.78 across various situation categories annotated by ScanNet v2, although its precision was relatively low at an average of 0.29. The root mean square error of the co-occurrence coefficient was 0.31. While GPT-4 tends to output slightly higher co-occurrence coefficients, it effectively captures the overall co-occurrence patterns observed in the ScanNet v2 dataset. GPT-4 produced co-occurrence information for more objects than those available in ScanNet v2 while covering co-occurrences among objects within ScanNet v2. These results demonstrate that integrating co-occurrence data from different sources could enhance the ability to recognize real-world objects and potentially strengthen robot intelligence.https://ieeexplore.ieee.org/document/10786984/Semantic scene understandinglarge language modelsco-occurrence validationprompt engineering
spellingShingle Kenta Gunji
Kazunori Ohno
Shuhei Kurita
Ken Sakurada
Ranulfo Bezerra
Shotaro Kojima
Yoshito Okada
Masashi Konyo
Satoshi Tadokoro
Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional Scenes
IEEE Access
Semantic scene understanding
large language models
co-occurrence validation
prompt engineering
title Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional Scenes
title_full Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional Scenes
title_fullStr Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional Scenes
title_full_unstemmed Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional Scenes
title_short Validation of LLM-Generated Object Co-Occurrence Information for Understanding Three-Dimensional Scenes
title_sort validation of llm generated object co occurrence information for understanding three dimensional scenes
topic Semantic scene understanding
large language models
co-occurrence validation
prompt engineering
url https://ieeexplore.ieee.org/document/10786984/
work_keys_str_mv AT kentagunji validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT kazunoriohno validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT shuheikurita validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT kensakurada validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT ranulfobezerra validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT shotarokojima validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT yoshitookada validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT masashikonyo validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes
AT satoshitadokoro validationofllmgeneratedobjectcooccurrenceinformationforunderstandingthreedimensionalscenes