helpResearch question人机协同标注与示例选择动态推荐关键帧和实时主动学习在音视频标注中表现如何?UIST '23PEANUT: A Human-AI Collaborative Tool for Annotating Audio-Visual Data
lightbulbPractical problem人机协同标注与示例选择音视频数据标注耗时费力,现有工具无法高效处理跨模态内容。UIST '23PEANUT: A Human-AI Collaborative Tool for Annotating Audio-Visual Data
helpResearch question人机协同标注与示例选择如何利用图像处理和深度学习技术自动从移动应用教程视频中提取用户操作?UIST '23Video2Action: Reducing Human Interactions in Action Annotation of App Tutorial Videos
helpResearch question人机协同标注与示例选择针对移动UI教程视频,哪些图像处理方法和深度学习模型能实现场景分割和动作位置预测的高精度?UIST '23Video2Action: Reducing Human Interactions in Action Annotation of App Tutorial Videos
helpResearch question人机协同标注与示例选择设计什么样的交互界面能辅助视频创作者高效完成操作标注?UIST '23Video2Action: Reducing Human Interactions in Action Annotation of App Tutorial Videos
lightbulbPractical problem人机协同标注与示例选择视频创作者标注移动应用教程视频的用户操作费时费力。UIST '23Video2Action: Reducing Human Interactions in Action Annotation of App Tutorial Videos
helpResearch question人机协同标注与示例选择在人体活动识别(HAR)中,如何通过微调编码器和分类层来优化自监督学习(SSL)的表现?UbiComp '23Don't freeze: Finetune encoders for better Self-Supervised HAR
helpResearch question人机协同标注与示例选择不同的自监督学习预置任务(如CPC、SimCLR和多任务学习)在人体活动识别中有什么性能差异?UbiComp '23Don't freeze: Finetune encoders for better Self-Supervised HAR
helpResearch question人机协同标注与示例选择在标签数据稀缺的情况下,联合微调方法如何影响模型的表现?UbiComp '23Don't freeze: Finetune encoders for better Self-Supervised HAR
lightbulbPractical problem人机协同标注与示例选择人体活动识别中缺少标注数据限制了模型性能提升。UbiComp '23Don't freeze: Finetune encoders for better Self-Supervised HAR
helpResearch question人机协同标注与示例选择如何系统地开发一个既满足工程需求又兼顾伦理和法律规定的数据标注协议?UbiComp '23ARDUOUS: Tutorial on Annotation of useR Data for UbiquitOUs Systems – Developing a Data Annotation Protocol
helpResearch question人机协同标注与示例选择在数据标注协议开发的各阶段中,哪些方法和工具能够提高标注效率和质量?UbiComp '23ARDUOUS: Tutorial on Annotation of useR Data for UbiquitOUs Systems – Developing a Data Annotation Protocol
helpResearch question人机协同标注与示例选择生成式AI和弱监督学习如何在数据标注中发挥作用,同时满足高风险AI系统的监管要求?UbiComp '23ARDUOUS: Tutorial on Annotation of useR Data for UbiquitOUs Systems – Developing a Data Annotation Protocol
lightbulbPractical problem人机协同标注与示例选择数据标注耗时高昂且易出错,不满足高风险AI系统的监管要求。UbiComp '23ARDUOUS: Tutorial on Annotation of useR Data for UbiquitOUs Systems – Developing a Data Annotation Protocol
helpResearch question人机协同标注与示例选择如何设计一个既高效又响应迅速的多目标跟踪标注工具?CHI '23TmoTA: Simple, Highly Responsive Tool for Multiple Object Tracking Annotation
helpResearch question人机协同标注与示例选择该工具能否在多目标跟踪标注任务中显著减少用户的操作时间?CHI '23TmoTA: Simple, Highly Responsive Tool for Multiple Object Tracking Annotation
helpResearch question人机协同标注与示例选择该工具的用户体验是否优于现有的手动标注工具?CHI '23TmoTA: Simple, Highly Responsive Tool for Multiple Object Tracking Annotation
lightbulbPractical problem人机协同标注与示例选择标注多目标视频数据费时且低效,现有工具响应慢。CHI '23TmoTA: Simple, Highly Responsive Tool for Multiple Object Tracking Annotation
helpResearch question人机协同标注与示例选择如何通过界面设计集成概念层级信息以提高多标签文本标注任务的质量和效率?CHI '23Interface Design for Crowdsourcing Hierarchical Multi-Label Text Annotations
helpResearch question人机协同标注与示例选择分层多标签标注方案是否在处理复杂和高度主观的任务中具有显著优势?CHI '23Interface Design for Crowdsourcing Hierarchical Multi-Label Text Annotations
helpResearch question人机协同标注与示例选择相较于随机分组,基于层级结构的分组能否显著提升标注准确率和一致性?CHI '23Interface Design for Crowdsourcing Hierarchical Multi-Label Text Annotations
lightbulbPractical problem人机协同标注与示例选择文本标注任务复杂、成本高且容易出现主观不一致的问题。CHI '23Interface Design for Crowdsourcing Hierarchical Multi-Label Text Annotations
helpResearch question人机协同标注与示例选择现有机器学习模型评估方法在特定应用上下文中存在哪些不足?CHI '23Kaleidoscope: Semantically-grounded, Context-specific ML Model Evaluation
helpResearch question人机协同标注与示例选择如何利用用户的实际经验定义上下文特定的语义概念集来评估机器学习模型?CHI '23Kaleidoscope: Semantically-grounded, Context-specific ML Model Evaluation
helpResearch question人机协同标注与示例选择基于语义的、自适应的机器学习模型评估框架如何提升评估的灵活性和相关性?CHI '23Kaleidoscope: Semantically-grounded, Context-specific ML Model Evaluation
lightbulbPractical problem人机协同标注与示例选择内容审核中,模型对社区特定语义内容的表现难以评估和优化。CHI '23Kaleidoscope: Semantically-grounded, Context-specific ML Model Evaluation
helpResearch question人机协同标注与示例选择如何在概念图编辑中有效分配人机任务,从而减轻用户的交互负担?CHI '23A Human-Computer Collaborative Editing Tool for Conceptual Diagrams
helpResearch question人机协同标注与示例选择多模态交互(如语音和手势)如何帮助用户更自然地编辑概念图?CHI '23A Human-Computer Collaborative Editing Tool for Conceptual Diagrams
helpResearch question人机协同标注与示例选择系统如何处理用户在概念图编辑中提供的模糊指令并推荐最佳解决方案?CHI '23A Human-Computer Collaborative Editing Tool for Conceptual Diagrams
lightbulbPractical problem人机协同标注与示例选择用户在移动设备上编辑概念图操作复杂,效率低。CHI '23A Human-Computer Collaborative Editing Tool for Conceptual Diagrams
helpResearch question人机协同标注与示例选择在人机协作训练中,样本排序对图像分类模型的训练效率有何影响?IUI '22Efficiently correcting machine learning: considering the role of example ordering in human-in-the-loop training of image classification models
helpResearch question人机协同标注与示例选择不同的样本排序策略如何影响人工纠错的工作量和模型性能?IUI '22Efficiently correcting machine learning: considering the role of example ordering in human-in-the-loop training of image classification models
helpResearch question人机协同标注与示例选择在主动学习和先验排序方法中,哪种策略能够更好地平衡精度和人工干预成本?IUI '22Efficiently correcting machine learning: considering the role of example ordering in human-in-the-loop training of image classification models
lightbulbPractical problem人机协同标注与示例选择人工标注图像分类数据集的成本高,尤其在高精度领域如医疗影像。IUI '22Efficiently correcting machine learning: considering the role of example ordering in human-in-the-loop training of image classification models
helpResearch question人机协同标注与示例选择在无须传统标注的情况下,如何通过试错互动有效地构建弱监督资源?UbiComp '22Let’s Embed Your Knowledge into AI by Trial and Error Instead of Annotation
helpResearch question人机协同标注与示例选择如何测量和识别最具破坏性的弱监督资源以优化模型性能?UbiComp '22Let’s Embed Your Knowledge into AI by Trial and Error Instead of Annotation
helpResearch question人机协同标注与示例选择弱监督中加入交互式修改流程是否能显著减少噪声数据并提升分类精度?UbiComp '22Let’s Embed Your Knowledge into AI by Trial and Error Instead of Annotation
lightbulbPractical problem人机协同标注与示例选择机器学习应用中,人工标注成本高,弱监督资源构建效率低。UbiComp '22Let’s Embed Your Knowledge into AI by Trial and Error Instead of Annotation
helpResearch question人机协同标注与示例选择如何实现无需依赖手动标注或特定设备的室内楼层定位?UbiComp '22TransFloor: Transparent Floor Localization for Crowdsourcing Instant Delivery
helpResearch question人机协同标注与示例选择能否通过自监督学习结合气压和Wi-Fi信号提高楼层定位的准确性?UbiComp '22TransFloor: Transparent Floor Localization for Crowdsourcing Instant Delivery
helpResearch question人机协同标注与示例选择在复杂的多楼层场景中,如何设计轻量灵活的楼层定位系统以支持广泛部署?UbiComp '22TransFloor: Transparent Floor Localization for Crowdsourcing Instant Delivery
lightbulbPractical problem人机协同标注与示例选择外卖场景中,快递员难以准确定位并找到指定楼层。UbiComp '22TransFloor: Transparent Floor Localization for Crowdsourcing Instant Delivery
helpResearch question人机协同标注与示例选择如何设计一个灵活的系统,以支持构建多样化和可扩展的数据标注工具?CHI '22OneLabeler: A Flexible System for Building Data Labeling Tools
helpResearch question人机协同标注与示例选择哪些模块和工作流程可以概括数据标注任务的需求,并实现高效的工具开发?CHI '22OneLabeler: A Flexible System for Building Data Labeling Tools
helpResearch question人机协同标注与示例选择可视化编程和模块化设计如何降低数据标注工具的开发门槛?CHI '22OneLabeler: A Flexible System for Building Data Labeling Tools
lightbulbPractical problem人机协同标注与示例选择开发数据标注工具耗时且需要复杂的跨学科技能。CHI '22OneLabeler: A Flexible System for Building Data Labeling Tools
helpResearch question人机协同标注与示例选择如何设计一个系统来自动检测和标注折线图中的潜在误导设计?CHI '22Annotating Line Charts for Addressing Deception
helpResearch question人机协同标注与示例选择这些标注能在多大程度上帮助普通读者识别和理解常见的设计问题?CHI '22Annotating Line Charts for Addressing Deception
helpResearch question人机协同标注与示例选择在何种程度上,自动标注工具可以减少低视觉素养用户对数据失真的感知?CHI '22Annotating Line Charts for Addressing Deception
lightbulbPractical problem人机协同标注与示例选择普通用户难以检测线上折线图中的误导设计,易被虚假数据影响。CHI '22Annotating Line Charts for Addressing Deception