Prompt-based Context- and Domain-aware Pretraining for Vision and Language Navigation

Pretrained visual-language models have extensive world knowledge and are widely used in visual and language navigation (VLN). However, they are not sensitive to indoor scenarios for VLN tasks. Another challenge for VLN is how the agent understands the contextual relations between actions on a path a...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2023-12
Hauptverfasser: Liu, Ting, Hu, Yue, Wu, Wansen, Wang, Youkai, Xu, Kai, Yin, Quanjun
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Liu, Ting
Hu, Yue
Wu, Wansen
Wang, Youkai
Xu, Kai
Yin, Quanjun
description Pretrained visual-language models have extensive world knowledge and are widely used in visual and language navigation (VLN). However, they are not sensitive to indoor scenarios for VLN tasks. Another challenge for VLN is how the agent understands the contextual relations between actions on a path and performs cross-modal alignment sequentially. In this paper, we propose a novel Prompt-bAsed coNtext- and inDoor-Aware (PANDA) pretraining framework to address these problems. It performs prompting in two stages. In the indoor-aware stage, we apply an efficient tuning paradigm to learn deep visual prompts from an indoor dataset, in order to augment pretrained models with inductive biases towards indoor environments. This can enable more sample-efficient adaptation for VLN agents. Furthermore, in the context-aware stage, we design a set of hard context prompts to capture the sequence-level semantics in the instruction. They enable further tuning of the pretrained models via contrastive learning. Experimental results on both R2R and REVERIE show the superiority of PANDA compared to existing state-of-the-art methods.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2862630521</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2862630521</sourcerecordid><originalsourceid>FETCH-proquest_journals_28626305213</originalsourceid><addsrcrecordid>eNqNitEKgjAYRkcQJOU7DLoezH9q3lvRRYQX1a384RyT3Gyb1eMn0QN0dfjOd2YkAiESVqQACxJ733HOId9AlomIXCpn-yGwG3rZ0NKaIN-BUTQN3doetWH4Qidp5WRw09RG0dY6etVeW_PtjmjUiErSEz61wjD5FZm3ePcy_nFJ1vvduTywwdnHKH2oOzs6M101FDnkgmeQiP-qD5AhQHE</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2862630521</pqid></control><display><type>article</type><title>Prompt-based Context- and Domain-aware Pretraining for Vision and Language Navigation</title><source>Free E- Journals</source><creator>Liu, Ting ; Hu, Yue ; Wu, Wansen ; Wang, Youkai ; Xu, Kai ; Yin, Quanjun</creator><creatorcontrib>Liu, Ting ; Hu, Yue ; Wu, Wansen ; Wang, Youkai ; Xu, Kai ; Yin, Quanjun</creatorcontrib><description>Pretrained visual-language models have extensive world knowledge and are widely used in visual and language navigation (VLN). However, they are not sensitive to indoor scenarios for VLN tasks. Another challenge for VLN is how the agent understands the contextual relations between actions on a path and performs cross-modal alignment sequentially. In this paper, we propose a novel Prompt-bAsed coNtext- and inDoor-Aware (PANDA) pretraining framework to address these problems. It performs prompting in two stages. In the indoor-aware stage, we apply an efficient tuning paradigm to learn deep visual prompts from an indoor dataset, in order to augment pretrained models with inductive biases towards indoor environments. This can enable more sample-efficient adaptation for VLN agents. Furthermore, in the context-aware stage, we design a set of hard context prompts to capture the sequence-level semantics in the instruction. They enable further tuning of the pretrained models via contrastive learning. Experimental results on both R2R and REVERIE show the superiority of PANDA compared to existing state-of-the-art methods.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Alignment ; Context ; Datasets ; Navigation ; Representations ; Semantics ; Tuning ; Vision</subject><ispartof>arXiv.org, 2023-12</ispartof><rights>2023. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</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>780,784</link.rule.ids></links><search><creatorcontrib>Liu, Ting</creatorcontrib><creatorcontrib>Hu, Yue</creatorcontrib><creatorcontrib>Wu, Wansen</creatorcontrib><creatorcontrib>Wang, Youkai</creatorcontrib><creatorcontrib>Xu, Kai</creatorcontrib><creatorcontrib>Yin, Quanjun</creatorcontrib><title>Prompt-based Context- and Domain-aware Pretraining for Vision and Language Navigation</title><title>arXiv.org</title><description>Pretrained visual-language models have extensive world knowledge and are widely used in visual and language navigation (VLN). However, they are not sensitive to indoor scenarios for VLN tasks. Another challenge for VLN is how the agent understands the contextual relations between actions on a path and performs cross-modal alignment sequentially. In this paper, we propose a novel Prompt-bAsed coNtext- and inDoor-Aware (PANDA) pretraining framework to address these problems. It performs prompting in two stages. In the indoor-aware stage, we apply an efficient tuning paradigm to learn deep visual prompts from an indoor dataset, in order to augment pretrained models with inductive biases towards indoor environments. This can enable more sample-efficient adaptation for VLN agents. Furthermore, in the context-aware stage, we design a set of hard context prompts to capture the sequence-level semantics in the instruction. They enable further tuning of the pretrained models via contrastive learning. Experimental results on both R2R and REVERIE show the superiority of PANDA compared to existing state-of-the-art methods.</description><subject>Alignment</subject><subject>Context</subject><subject>Datasets</subject><subject>Navigation</subject><subject>Representations</subject><subject>Semantics</subject><subject>Tuning</subject><subject>Vision</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNqNitEKgjAYRkcQJOU7DLoezH9q3lvRRYQX1a384RyT3Gyb1eMn0QN0dfjOd2YkAiESVqQACxJ733HOId9AlomIXCpn-yGwG3rZ0NKaIN-BUTQN3doetWH4Qidp5WRw09RG0dY6etVeW_PtjmjUiErSEz61wjD5FZm3ePcy_nFJ1vvduTywwdnHKH2oOzs6M101FDnkgmeQiP-qD5AhQHE</recordid><startdate>20231214</startdate><enddate>20231214</enddate><creator>Liu, Ting</creator><creator>Hu, Yue</creator><creator>Wu, Wansen</creator><creator>Wang, Youkai</creator><creator>Xu, Kai</creator><creator>Yin, Quanjun</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20231214</creationdate><title>Prompt-based Context- and Domain-aware Pretraining for Vision and Language Navigation</title><author>Liu, Ting ; Hu, Yue ; Wu, Wansen ; Wang, Youkai ; Xu, Kai ; Yin, Quanjun</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_28626305213</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Alignment</topic><topic>Context</topic><topic>Datasets</topic><topic>Navigation</topic><topic>Representations</topic><topic>Semantics</topic><topic>Tuning</topic><topic>Vision</topic><toplevel>online_resources</toplevel><creatorcontrib>Liu, Ting</creatorcontrib><creatorcontrib>Hu, Yue</creatorcontrib><creatorcontrib>Wu, Wansen</creatorcontrib><creatorcontrib>Wang, Youkai</creatorcontrib><creatorcontrib>Xu, Kai</creatorcontrib><creatorcontrib>Yin, Quanjun</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Liu, Ting</au><au>Hu, Yue</au><au>Wu, Wansen</au><au>Wang, Youkai</au><au>Xu, Kai</au><au>Yin, Quanjun</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Prompt-based Context- and Domain-aware Pretraining for Vision and Language Navigation</atitle><jtitle>arXiv.org</jtitle><date>2023-12-14</date><risdate>2023</risdate><eissn>2331-8422</eissn><abstract>Pretrained visual-language models have extensive world knowledge and are widely used in visual and language navigation (VLN). However, they are not sensitive to indoor scenarios for VLN tasks. Another challenge for VLN is how the agent understands the contextual relations between actions on a path and performs cross-modal alignment sequentially. In this paper, we propose a novel Prompt-bAsed coNtext- and inDoor-Aware (PANDA) pretraining framework to address these problems. It performs prompting in two stages. In the indoor-aware stage, we apply an efficient tuning paradigm to learn deep visual prompts from an indoor dataset, in order to augment pretrained models with inductive biases towards indoor environments. This can enable more sample-efficient adaptation for VLN agents. Furthermore, in the context-aware stage, we design a set of hard context prompts to capture the sequence-level semantics in the instruction. They enable further tuning of the pretrained models via contrastive learning. Experimental results on both R2R and REVERIE show the superiority of PANDA compared to existing state-of-the-art methods.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2023-12
issn 2331-8422
language eng
recordid cdi_proquest_journals_2862630521
source Free E- Journals
subjects Alignment
Context
Datasets
Navigation
Representations
Semantics
Tuning
Vision
title Prompt-based Context- and Domain-aware Pretraining for Vision and Language Navigation
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T19%3A36%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Prompt-based%20Context-%20and%20Domain-aware%20Pretraining%20for%20Vision%20and%20Language%20Navigation&rft.jtitle=arXiv.org&rft.au=Liu,%20Ting&rft.date=2023-12-14&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2862630521%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2862630521&rft_id=info:pmid/&rfr_iscdi=true