help研究问题HAR模型鲁棒性、泛化与数据增强如何通过自监督学习(SSL)从未标注的加速度计数据中学习到有效的人类活动识别(HAR)表示?UbiComp '24Self-supervised Learning for Accelerometer-based Human Activity Recognition: A Survey
help研究问题HAR模型鲁棒性、泛化与数据增强现有的SSL方法在HAR中的训练目标和方法上存在哪些优缺点?UbiComp '24Self-supervised Learning for Accelerometer-based Human Activity Recognition: A Survey
help研究问题HAR模型鲁棒性、泛化与数据增强如何改进SSL在HAR中的通用化能力,并实现标准化评估?UbiComp '24Self-supervised Learning for Accelerometer-based Human Activity Recognition: A Survey
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强标注加速度计数据的成本高,现有算法无法很好适应不同场景。UbiComp '24Self-supervised Learning for Accelerometer-based Human Activity Recognition: A Survey
help研究问题HAR模型鲁棒性、泛化与数据增强如何利用智能手表的被动振动,识别用户日常使用的物体?UbiComp '24ViObject: Harness Passive Vibrations for Daily Object Recognition with Commodity Smartwatches
help研究问题HAR模型鲁棒性、泛化与数据增强在日常物体识别中,被动振动技术相比主动振动技术和标签技术有哪些优势?UbiComp '24ViObject: Harness Passive Vibrations for Daily Object Recognition with Commodity Smartwatches
help研究问题HAR模型鲁棒性、泛化与数据增强怎样通过少量训练样本快速适配新的用户和新的物体?UbiComp '24ViObject: Harness Passive Vibrations for Daily Object Recognition with Commodity Smartwatches
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强基于标签或图像的物体识别技术需要额外设备,干扰自然使用。UbiComp '24ViObject: Harness Passive Vibrations for Daily Object Recognition with Commodity Smartwatches
help研究问题HAR模型鲁棒性、泛化与数据增强不同深度学习模型在IMU传感器安装位置变化下的鲁棒性表现如何?UbiComp '24DNN Model Comparison for Sensor Location Robustness
help研究问题HAR模型鲁棒性、泛化与数据增强动态3D旋转技术能否提升模型应对IMU位置偏移的能力?UbiComp '24DNN Model Comparison for Sensor Location Robustness
help研究问题HAR模型鲁棒性、泛化与数据增强如何系统性评估IMU位置变化对分类性能的影响?UbiComp '24DNN Model Comparison for Sensor Location Robustness
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强穿戴设备的IMU传感器位置微小变化会显著降低活动识别效果。UbiComp '24DNN Model Comparison for Sensor Location Robustness
help研究问题HAR模型鲁棒性、泛化与数据增强深度学习模型在面对传感器佩戴方向变化时,其性能会受到多大影响?UbiComp '23Investigating the Effect of Orientation Variability in Deep Learning-based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强人类活动识别中,传感器方向变化导致的数据分布偏移对模型性能的影响机制是什么?UbiComp '23Investigating the Effect of Orientation Variability in Deep Learning-based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强深度学习架构在应对方向变化的鲁棒性上存在多大的差异?UbiComp '23Investigating the Effect of Orientation Variability in Deep Learning-based Human Activity Recognition
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强可穿戴设备因佩戴方式不同,导致运动监测准确性下降。UbiComp '23Investigating the Effect of Orientation Variability in Deep Learning-based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强如何在无人监督学习模式下,为基于传感器的人类活动识别生成多样高质量的时间序列数据?UbiComp '23Unsupervised Diffusion Model for Sensor-based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强统计特征(如均值、标准差、Z分数等)是否可以有效引导扩散模型生成更具代表性的数据?UbiComp '23Unsupervised Diffusion Model for Sensor-based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强与传统过采样方法和GAN生成方法相比,扩散模型能否显著提高活动识别性能?UbiComp '23Unsupervised Diffusion Model for Sensor-based Human Activity Recognition
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强传感器数据匮乏且标注成本高,导致活动识别模型效果不佳UbiComp '23Unsupervised Diffusion Model for Sensor-based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强如何在联合学习(FL)中,解决人类活动识别(HAR)客户端缺乏标注数据的挑战?UbiComp '23Personalized Federated Human Activity Recognition through Semi-supervised Learning and Enhanced Representation
help研究问题HAR模型鲁棒性、泛化与数据增强霍桑效应(知道被观察时行为改变)是否会导致传感器收集的人类活动数据存在偏差?UbiComp '23A Data-Driven Study on the Hawthorne Effect in Sensor-Based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强实验室收集的传感器数据在多大程度上能够泛化到真实环境中?UbiComp '23A Data-Driven Study on the Hawthorne Effect in Sensor-Based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强深度学习分类算法是否能检测并适应观察环境的变化?UbiComp '23A Data-Driven Study on the Hawthorne Effect in Sensor-Based Human Activity Recognition
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强可穿戴设备的活动识别系统在真实环境中可能存在预测不准确的问题。UbiComp '23A Data-Driven Study on the Hawthorne Effect in Sensor-Based Human Activity Recognition
help研究问题HAR模型鲁棒性、泛化与数据增强如何利用联邦学习(保护用户隐私的分布式算法)实现智能眼镜的人类活动识别?UbiComp '23Privacy-aware Human Activity Recognition with Smart Glasses for Digital Therapeutics
help研究问题HAR模型鲁棒性、泛化与数据增强在隐私保护的前提下,复杂网络(如STResNet)和简单网络(如DANA)在活动识别中的性能差异是什么?UbiComp '23Privacy-aware Human Activity Recognition with Smart Glasses for Digital Therapeutics
help研究问题HAR模型鲁棒性、泛化与数据增强使用智能眼镜的惯性数据,联邦学习能否达到接近集中式训练的精度表现?UbiComp '23Privacy-aware Human Activity Recognition with Smart Glasses for Digital Therapeutics
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强用户担心使用健康监测设备会导致隐私泄露。UbiComp '23Privacy-aware Human Activity Recognition with Smart Glasses for Digital Therapeutics
help研究问题HAR模型鲁棒性、泛化与数据增强如何将基于Transformer的大型语言模型扩展应用于一维传感器信号数据的自监督学习?UbiComp '23Transformer-based LLMs for Sensor Data
help研究问题HAR模型鲁棒性、泛化与数据增强基于多任务预训练和对比学习机制,Transformer模型在活动识别和情感检测等应用中能否超越传统ResNet架构?UbiComp '23Transformer-based LLMs for Sensor Data
help研究问题HAR模型鲁棒性、泛化与数据增强设计什么样的预训练任务可以最有效增强传感器数据的特征提取能力?UbiComp '23Transformer-based LLMs for Sensor Data
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强可穿戴设备中传感器数据的分析难以满足健康监测和情感识别需求。UbiComp '23Transformer-based LLMs for Sensor Data
help研究问题HAR模型鲁棒性、泛化与数据增强自监督学习如何克服人类活动识别中标签数据匮乏的问题?UbiComp '23Solving the Sensor-Based Activity Recognition Problem (SOAR): Self-Supervised, Multi-Modal Recognition of Activities from Wearable Sensors
help研究问题HAR模型鲁棒性、泛化与数据增强多模态数据如何增强可穿戴传感器在活动识别中的性能?UbiComp '23Solving the Sensor-Based Activity Recognition Problem (SOAR): Self-Supervised, Multi-Modal Recognition of Activities from Wearable Sensors
help研究问题HAR模型鲁棒性、泛化与数据增强现代表示学习技术(如对比学习)在传感器数据中的潜力是什么?UbiComp '23Solving the Sensor-Based Activity Recognition Problem (SOAR): Self-Supervised, Multi-Modal Recognition of Activities from Wearable Sensors
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强穿戴设备在无标签数据场景下的活动识别性能不足。UbiComp '23Solving the Sensor-Based Activity Recognition Problem (SOAR): Self-Supervised, Multi-Modal Recognition of Activities from Wearable Sensors
help研究问题HAR模型鲁棒性、泛化与数据增强IMU传感器在进行人体活动识别(HAR)时的最佳位置是哪里?UbiComp '23Where Are the Best Positions of IMU Sensors for HAR?
help研究问题HAR模型鲁棒性、泛化与数据增强单个传感器与多个传感器组合的位置如何影响活动分类的准确率?UbiComp '23Where Are the Best Positions of IMU Sensors for HAR?
help研究问题HAR模型鲁棒性、泛化与数据增强不同活动(例如基础活动和日常活动)在最佳传感器位置上是否存在差异?UbiComp '23Where Are the Best Positions of IMU Sensors for HAR?
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强现有IMU传感器位置选择粗糙,导致活动识别效果受限。UbiComp '23Where Are the Best Positions of IMU Sensors for HAR?
help研究问题HAR模型鲁棒性、泛化与数据增强如何通过结合低采样率音频和低功耗惯性测量单元(IMU)实现高效的活动识别?UbiComp '22SAMoSA: Sensing Activities with Motion and Subsampled Audio
help研究问题HAR模型鲁棒性、泛化与数据增强低采样率音频(≤1 kHz)在活动识别中如何在保证隐私的同时维持性能?UbiComp '22SAMoSA: Sensing Activities with Motion and Subsampled Audio
help研究问题HAR模型鲁棒性、泛化与数据增强如何优化资源受限设备(如智能手表)的隐私保护型人类活动识别系统?UbiComp '22SAMoSA: Sensing Activities with Motion and Subsampled Audio
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强智能手表等设备的活动识别系统耗能高且存在隐私泄露风险UbiComp '22SAMoSA: Sensing Activities with Motion and Subsampled Audio
help研究问题HAR模型鲁棒性、泛化与数据增强在传感器数据部分缺失的情况下,人类活动识别模型如何提升对不同传感器组合的泛化能力?UbiComp '22Augmented Adversarial Learning for Human Activity Recognition with Partial Sensor Sets
help研究问题HAR模型鲁棒性、泛化与数据增强混合缺失策略和多域混合策略如何影响模型对高传感器丢失率的鲁棒性?UbiComp '22Augmented Adversarial Learning for Human Activity Recognition with Partial Sensor Sets
help研究问题HAR模型鲁棒性、泛化与数据增强强化对抗学习框架能否在处理多源异质数据时提升模型性能及适应性?UbiComp '22Augmented Adversarial Learning for Human Activity Recognition with Partial Sensor Sets
lightbulb现实痛点HAR模型鲁棒性、泛化与数据增强真实场景中传感器数据常有丢失,活动识别模型准确率低。UbiComp '22Augmented Adversarial Learning for Human Activity Recognition with Partial Sensor Sets
help研究问题HAR模型鲁棒性、泛化与数据增强如何改进人类活动识别模型的跨域泛化能力,尤其是在未知目标域数据不可访问的情况下?UbiComp '22Semantic-Discriminative Mixup for Generalizable Sensor-based Cross-domain Activity Recognition