BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulation

We present BEHAVIOR-1K, a comprehensive simulation benchmark for human-centered robotics. BEHAVIOR-1K includes two components, guided and motivated by the results of an extensive survey on "what do you want robots to do for you?". The first is the definition of 1,000 everyday activities, g...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, Chengshu, Zhang, Ruohan, Wong, Josiah, Gokmen, Cem, Srivastava, Sanjana, Martín-Martín, Roberto, Wang, Chen, Levine, Gabrael, Ai, Wensi, Martinez, Benjamin, Yin, Hang, Lingelbach, Michael, Hwang, Minjune, Hiranaka, Ayano, Garlanka, Sujay, Aydin, Arman, Lee, Sharon, Sun, Jiankai, Anvari, Mona, Sharma, Manasi, Bansal, Dhruva, Hunter, Samuel, Kim, Kyu-Young, Lou, Alan, Matthews, Caleb R, Villa-Renteria, Ivan, Tang, Jerry Huayang, Tang, Claire, Xia, Fei, Li, Yunzhu, Savarese, Silvio, Gweon, Hyowon, Liu, C. Karen, Wu, Jiajun, Fei-Fei, Li
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Li, Chengshu
Zhang, Ruohan
Wong, Josiah
Gokmen, Cem
Srivastava, Sanjana
Martín-Martín, Roberto
Wang, Chen
Levine, Gabrael
Ai, Wensi
Martinez, Benjamin
Yin, Hang
Lingelbach, Michael
Hwang, Minjune
Hiranaka, Ayano
Garlanka, Sujay
Aydin, Arman
Lee, Sharon
Sun, Jiankai
Anvari, Mona
Sharma, Manasi
Bansal, Dhruva
Hunter, Samuel
Kim, Kyu-Young
Lou, Alan
Matthews, Caleb R
Villa-Renteria, Ivan
Tang, Jerry Huayang
Tang, Claire
Xia, Fei
Li, Yunzhu
Savarese, Silvio
Gweon, Hyowon
Liu, C. Karen
Wu, Jiajun
Fei-Fei, Li
description We present BEHAVIOR-1K, a comprehensive simulation benchmark for human-centered robotics. BEHAVIOR-1K includes two components, guided and motivated by the results of an extensive survey on "what do you want robots to do for you?". The first is the definition of 1,000 everyday activities, grounded in 50 scenes (houses, gardens, restaurants, offices, etc.) with more than 9,000 objects annotated with rich physical and semantic properties. The second is OMNIGIBSON, a novel simulation environment that supports these activities via realistic physics simulation and rendering of rigid bodies, deformable bodies, and liquids. Our experiments indicate that the activities in BEHAVIOR-1K are long-horizon and dependent on complex manipulation skills, both of which remain a challenge for even state-of-the-art robot learning solutions. To calibrate the simulation-to-reality gap of BEHAVIOR-1K, we provide an initial study on transferring solutions learned with a mobile manipulator in a simulated apartment to its real-world counterpart. We hope that BEHAVIOR-1K's human-grounded nature, diversity, and realism make it valuable for embodied AI and robot learning research. Project website: https://behavior.stanford.edu.
doi_str_mv 10.48550/arxiv.2403.09227
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2403_09227</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2403_09227</sourcerecordid><originalsourceid>FETCH-LOGICAL-a677-2fab3003760a7bfbf79385942c395fac7151bd445e7f1695201c3df04b3d20ba3</originalsourceid><addsrcrecordid>eNotz81Og0AYheHZuDDVC3DlXEDBb2YYprijBC2xSZPauCXf_KUTgRqgKHevVldn9-Y8hNwxiJOVlPCA_VeYYp6AiCHjXF2TZl1u8rdqt4_YyyPN6ebcYhcVrhtd7-ySlq0-2eAszSu6dp05tti_088wHilbAgAtJ9fPFmeamzFMYQxuoNhZunfYhGEMhr6G9tzgGE7dDbny2Azu9n8X5PBUHopNtN09V0W-jTBVKuIetQAQKgVU2muvMrGSWcKNyKRHo5hk2iaJdMqzNJMcmBHWQ6KF5aBRLMj9X_bCrT_68HN6rn_Z9YUtvgFgN1BL</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulation</title><source>arXiv.org</source><creator>Li, Chengshu ; Zhang, Ruohan ; Wong, Josiah ; Gokmen, Cem ; Srivastava, Sanjana ; Martín-Martín, Roberto ; Wang, Chen ; Levine, Gabrael ; Ai, Wensi ; Martinez, Benjamin ; Yin, Hang ; Lingelbach, Michael ; Hwang, Minjune ; Hiranaka, Ayano ; Garlanka, Sujay ; Aydin, Arman ; Lee, Sharon ; Sun, Jiankai ; Anvari, Mona ; Sharma, Manasi ; Bansal, Dhruva ; Hunter, Samuel ; Kim, Kyu-Young ; Lou, Alan ; Matthews, Caleb R ; Villa-Renteria, Ivan ; Tang, Jerry Huayang ; Tang, Claire ; Xia, Fei ; Li, Yunzhu ; Savarese, Silvio ; Gweon, Hyowon ; Liu, C. Karen ; Wu, Jiajun ; Fei-Fei, Li</creator><creatorcontrib>Li, Chengshu ; Zhang, Ruohan ; Wong, Josiah ; Gokmen, Cem ; Srivastava, Sanjana ; Martín-Martín, Roberto ; Wang, Chen ; Levine, Gabrael ; Ai, Wensi ; Martinez, Benjamin ; Yin, Hang ; Lingelbach, Michael ; Hwang, Minjune ; Hiranaka, Ayano ; Garlanka, Sujay ; Aydin, Arman ; Lee, Sharon ; Sun, Jiankai ; Anvari, Mona ; Sharma, Manasi ; Bansal, Dhruva ; Hunter, Samuel ; Kim, Kyu-Young ; Lou, Alan ; Matthews, Caleb R ; Villa-Renteria, Ivan ; Tang, Jerry Huayang ; Tang, Claire ; Xia, Fei ; Li, Yunzhu ; Savarese, Silvio ; Gweon, Hyowon ; Liu, C. Karen ; Wu, Jiajun ; Fei-Fei, Li</creatorcontrib><description>We present BEHAVIOR-1K, a comprehensive simulation benchmark for human-centered robotics. BEHAVIOR-1K includes two components, guided and motivated by the results of an extensive survey on "what do you want robots to do for you?". The first is the definition of 1,000 everyday activities, grounded in 50 scenes (houses, gardens, restaurants, offices, etc.) with more than 9,000 objects annotated with rich physical and semantic properties. The second is OMNIGIBSON, a novel simulation environment that supports these activities via realistic physics simulation and rendering of rigid bodies, deformable bodies, and liquids. Our experiments indicate that the activities in BEHAVIOR-1K are long-horizon and dependent on complex manipulation skills, both of which remain a challenge for even state-of-the-art robot learning solutions. To calibrate the simulation-to-reality gap of BEHAVIOR-1K, we provide an initial study on transferring solutions learned with a mobile manipulator in a simulated apartment to its real-world counterpart. We hope that BEHAVIOR-1K's human-grounded nature, diversity, and realism make it valuable for embodied AI and robot learning research. Project website: https://behavior.stanford.edu.</description><identifier>DOI: 10.48550/arxiv.2403.09227</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Robotics</subject><creationdate>2024-03</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2403.09227$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2403.09227$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Li, Chengshu</creatorcontrib><creatorcontrib>Zhang, Ruohan</creatorcontrib><creatorcontrib>Wong, Josiah</creatorcontrib><creatorcontrib>Gokmen, Cem</creatorcontrib><creatorcontrib>Srivastava, Sanjana</creatorcontrib><creatorcontrib>Martín-Martín, Roberto</creatorcontrib><creatorcontrib>Wang, Chen</creatorcontrib><creatorcontrib>Levine, Gabrael</creatorcontrib><creatorcontrib>Ai, Wensi</creatorcontrib><creatorcontrib>Martinez, Benjamin</creatorcontrib><creatorcontrib>Yin, Hang</creatorcontrib><creatorcontrib>Lingelbach, Michael</creatorcontrib><creatorcontrib>Hwang, Minjune</creatorcontrib><creatorcontrib>Hiranaka, Ayano</creatorcontrib><creatorcontrib>Garlanka, Sujay</creatorcontrib><creatorcontrib>Aydin, Arman</creatorcontrib><creatorcontrib>Lee, Sharon</creatorcontrib><creatorcontrib>Sun, Jiankai</creatorcontrib><creatorcontrib>Anvari, Mona</creatorcontrib><creatorcontrib>Sharma, Manasi</creatorcontrib><creatorcontrib>Bansal, Dhruva</creatorcontrib><creatorcontrib>Hunter, Samuel</creatorcontrib><creatorcontrib>Kim, Kyu-Young</creatorcontrib><creatorcontrib>Lou, Alan</creatorcontrib><creatorcontrib>Matthews, Caleb R</creatorcontrib><creatorcontrib>Villa-Renteria, Ivan</creatorcontrib><creatorcontrib>Tang, Jerry Huayang</creatorcontrib><creatorcontrib>Tang, Claire</creatorcontrib><creatorcontrib>Xia, Fei</creatorcontrib><creatorcontrib>Li, Yunzhu</creatorcontrib><creatorcontrib>Savarese, Silvio</creatorcontrib><creatorcontrib>Gweon, Hyowon</creatorcontrib><creatorcontrib>Liu, C. Karen</creatorcontrib><creatorcontrib>Wu, Jiajun</creatorcontrib><creatorcontrib>Fei-Fei, Li</creatorcontrib><title>BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulation</title><description>We present BEHAVIOR-1K, a comprehensive simulation benchmark for human-centered robotics. BEHAVIOR-1K includes two components, guided and motivated by the results of an extensive survey on "what do you want robots to do for you?". The first is the definition of 1,000 everyday activities, grounded in 50 scenes (houses, gardens, restaurants, offices, etc.) with more than 9,000 objects annotated with rich physical and semantic properties. The second is OMNIGIBSON, a novel simulation environment that supports these activities via realistic physics simulation and rendering of rigid bodies, deformable bodies, and liquids. Our experiments indicate that the activities in BEHAVIOR-1K are long-horizon and dependent on complex manipulation skills, both of which remain a challenge for even state-of-the-art robot learning solutions. To calibrate the simulation-to-reality gap of BEHAVIOR-1K, we provide an initial study on transferring solutions learned with a mobile manipulator in a simulated apartment to its real-world counterpart. We hope that BEHAVIOR-1K's human-grounded nature, diversity, and realism make it valuable for embodied AI and robot learning research. Project website: https://behavior.stanford.edu.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Robotics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz81Og0AYheHZuDDVC3DlXEDBb2YYprijBC2xSZPauCXf_KUTgRqgKHevVldn9-Y8hNwxiJOVlPCA_VeYYp6AiCHjXF2TZl1u8rdqt4_YyyPN6ebcYhcVrhtd7-ySlq0-2eAszSu6dp05tti_088wHilbAgAtJ9fPFmeamzFMYQxuoNhZunfYhGEMhr6G9tzgGE7dDbny2Azu9n8X5PBUHopNtN09V0W-jTBVKuIetQAQKgVU2muvMrGSWcKNyKRHo5hk2iaJdMqzNJMcmBHWQ6KF5aBRLMj9X_bCrT_68HN6rn_Z9YUtvgFgN1BL</recordid><startdate>20240314</startdate><enddate>20240314</enddate><creator>Li, Chengshu</creator><creator>Zhang, Ruohan</creator><creator>Wong, Josiah</creator><creator>Gokmen, Cem</creator><creator>Srivastava, Sanjana</creator><creator>Martín-Martín, Roberto</creator><creator>Wang, Chen</creator><creator>Levine, Gabrael</creator><creator>Ai, Wensi</creator><creator>Martinez, Benjamin</creator><creator>Yin, Hang</creator><creator>Lingelbach, Michael</creator><creator>Hwang, Minjune</creator><creator>Hiranaka, Ayano</creator><creator>Garlanka, Sujay</creator><creator>Aydin, Arman</creator><creator>Lee, Sharon</creator><creator>Sun, Jiankai</creator><creator>Anvari, Mona</creator><creator>Sharma, Manasi</creator><creator>Bansal, Dhruva</creator><creator>Hunter, Samuel</creator><creator>Kim, Kyu-Young</creator><creator>Lou, Alan</creator><creator>Matthews, Caleb R</creator><creator>Villa-Renteria, Ivan</creator><creator>Tang, Jerry Huayang</creator><creator>Tang, Claire</creator><creator>Xia, Fei</creator><creator>Li, Yunzhu</creator><creator>Savarese, Silvio</creator><creator>Gweon, Hyowon</creator><creator>Liu, C. Karen</creator><creator>Wu, Jiajun</creator><creator>Fei-Fei, Li</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240314</creationdate><title>BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulation</title><author>Li, Chengshu ; Zhang, Ruohan ; Wong, Josiah ; Gokmen, Cem ; Srivastava, Sanjana ; Martín-Martín, Roberto ; Wang, Chen ; Levine, Gabrael ; Ai, Wensi ; Martinez, Benjamin ; Yin, Hang ; Lingelbach, Michael ; Hwang, Minjune ; Hiranaka, Ayano ; Garlanka, Sujay ; Aydin, Arman ; Lee, Sharon ; Sun, Jiankai ; Anvari, Mona ; Sharma, Manasi ; Bansal, Dhruva ; Hunter, Samuel ; Kim, Kyu-Young ; Lou, Alan ; Matthews, Caleb R ; Villa-Renteria, Ivan ; Tang, Jerry Huayang ; Tang, Claire ; Xia, Fei ; Li, Yunzhu ; Savarese, Silvio ; Gweon, Hyowon ; Liu, C. Karen ; Wu, Jiajun ; Fei-Fei, Li</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a677-2fab3003760a7bfbf79385942c395fac7151bd445e7f1695201c3df04b3d20ba3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Robotics</topic><toplevel>online_resources</toplevel><creatorcontrib>Li, Chengshu</creatorcontrib><creatorcontrib>Zhang, Ruohan</creatorcontrib><creatorcontrib>Wong, Josiah</creatorcontrib><creatorcontrib>Gokmen, Cem</creatorcontrib><creatorcontrib>Srivastava, Sanjana</creatorcontrib><creatorcontrib>Martín-Martín, Roberto</creatorcontrib><creatorcontrib>Wang, Chen</creatorcontrib><creatorcontrib>Levine, Gabrael</creatorcontrib><creatorcontrib>Ai, Wensi</creatorcontrib><creatorcontrib>Martinez, Benjamin</creatorcontrib><creatorcontrib>Yin, Hang</creatorcontrib><creatorcontrib>Lingelbach, Michael</creatorcontrib><creatorcontrib>Hwang, Minjune</creatorcontrib><creatorcontrib>Hiranaka, Ayano</creatorcontrib><creatorcontrib>Garlanka, Sujay</creatorcontrib><creatorcontrib>Aydin, Arman</creatorcontrib><creatorcontrib>Lee, Sharon</creatorcontrib><creatorcontrib>Sun, Jiankai</creatorcontrib><creatorcontrib>Anvari, Mona</creatorcontrib><creatorcontrib>Sharma, Manasi</creatorcontrib><creatorcontrib>Bansal, Dhruva</creatorcontrib><creatorcontrib>Hunter, Samuel</creatorcontrib><creatorcontrib>Kim, Kyu-Young</creatorcontrib><creatorcontrib>Lou, Alan</creatorcontrib><creatorcontrib>Matthews, Caleb R</creatorcontrib><creatorcontrib>Villa-Renteria, Ivan</creatorcontrib><creatorcontrib>Tang, Jerry Huayang</creatorcontrib><creatorcontrib>Tang, Claire</creatorcontrib><creatorcontrib>Xia, Fei</creatorcontrib><creatorcontrib>Li, Yunzhu</creatorcontrib><creatorcontrib>Savarese, Silvio</creatorcontrib><creatorcontrib>Gweon, Hyowon</creatorcontrib><creatorcontrib>Liu, C. Karen</creatorcontrib><creatorcontrib>Wu, Jiajun</creatorcontrib><creatorcontrib>Fei-Fei, Li</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Li, Chengshu</au><au>Zhang, Ruohan</au><au>Wong, Josiah</au><au>Gokmen, Cem</au><au>Srivastava, Sanjana</au><au>Martín-Martín, Roberto</au><au>Wang, Chen</au><au>Levine, Gabrael</au><au>Ai, Wensi</au><au>Martinez, Benjamin</au><au>Yin, Hang</au><au>Lingelbach, Michael</au><au>Hwang, Minjune</au><au>Hiranaka, Ayano</au><au>Garlanka, Sujay</au><au>Aydin, Arman</au><au>Lee, Sharon</au><au>Sun, Jiankai</au><au>Anvari, Mona</au><au>Sharma, Manasi</au><au>Bansal, Dhruva</au><au>Hunter, Samuel</au><au>Kim, Kyu-Young</au><au>Lou, Alan</au><au>Matthews, Caleb R</au><au>Villa-Renteria, Ivan</au><au>Tang, Jerry Huayang</au><au>Tang, Claire</au><au>Xia, Fei</au><au>Li, Yunzhu</au><au>Savarese, Silvio</au><au>Gweon, Hyowon</au><au>Liu, C. Karen</au><au>Wu, Jiajun</au><au>Fei-Fei, Li</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulation</atitle><date>2024-03-14</date><risdate>2024</risdate><abstract>We present BEHAVIOR-1K, a comprehensive simulation benchmark for human-centered robotics. BEHAVIOR-1K includes two components, guided and motivated by the results of an extensive survey on "what do you want robots to do for you?". The first is the definition of 1,000 everyday activities, grounded in 50 scenes (houses, gardens, restaurants, offices, etc.) with more than 9,000 objects annotated with rich physical and semantic properties. The second is OMNIGIBSON, a novel simulation environment that supports these activities via realistic physics simulation and rendering of rigid bodies, deformable bodies, and liquids. Our experiments indicate that the activities in BEHAVIOR-1K are long-horizon and dependent on complex manipulation skills, both of which remain a challenge for even state-of-the-art robot learning solutions. To calibrate the simulation-to-reality gap of BEHAVIOR-1K, we provide an initial study on transferring solutions learned with a mobile manipulator in a simulated apartment to its real-world counterpart. We hope that BEHAVIOR-1K's human-grounded nature, diversity, and realism make it valuable for embodied AI and robot learning research. Project website: https://behavior.stanford.edu.</abstract><doi>10.48550/arxiv.2403.09227</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2403.09227
ispartof
issn
language eng
recordid cdi_arxiv_primary_2403_09227
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Robotics
title BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulation
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-01T04%3A57%3A31IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=BEHAVIOR-1K:%20A%20Human-Centered,%20Embodied%20AI%20Benchmark%20with%201,000%20Everyday%20Activities%20and%20Realistic%20Simulation&rft.au=Li,%20Chengshu&rft.date=2024-03-14&rft_id=info:doi/10.48550/arxiv.2403.09227&rft_dat=%3Carxiv_GOX%3E2403_09227%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true