Tag
该论文提出了一种接触式多模态筛查框架，利用与 Amazon Alexa 交互过程中采集的视频、音频和文本信号来检测轻度认知障碍（MCI），实验表明多模态融合模型在 Command-Reading 和 Question-Answering 任务上均达到 0.76 的准确率，优于单一模态模型。
A physics education paper showing how 3D trajectories of roller coasters and pendulum-ship rides can be reconstructed from smartphone accelerometer and gyroscope data using quaternion-based orientation integration with drift correction, enabling Newtonian mechanics and energy analysis for teaching.
DyRAD presents a radar novel view synthesis method for dynamic driving scenes that models static background and motion-tracked dynamic point reflectors to render complete range-azimuth-Doppler tensors, using a fixed analytic point-spread function to enable zero-shot sensor-configuration transfer and outperforming baselines by a wide margin on the RADIal benchmark.
This paper develops two resource-efficient distributed recursive Gaussian process algorithms for multi-output regression in multi-agent systems, reducing communication overhead while maintaining estimation accuracy.
GaitVista is a reliability-aware AI measurement layer that uses lightweight gates to assign joint-specific visual contributions from camera and inertial sensor data, reducing error and improving accuracy for longitudinal gait assessment in accessible settings.
This paper introduces Dynamic Influence Weighting (DIW), a knowledge distillation method that improves single-IMU activity recognition by dynamically weighting teacher targets from multiple IMUs during training, achieving significant performance gains.
The author built a tiny robot named Growbot using 3D printing, a Raspberry Pi Pico 2, servos, and a smartphone, combined with modern AI. It can see, hear, remember, move, and interact. The author shares their building journey from failure to a robot that feels alive, highlighting the transformative effect of giving AI a physical body.
This paper systematically compares seven state-of-the-art sensor fusion methods for multi-modal human activity recognition on the HARMES dataset, showing that Gated Multi-modal Fusion achieves the highest macro F1-score of 0.82, outperforming the baseline by 6 percentage points.
This paper discovers 'Authority Inversion' in LLM-mediated ubiquitous systems, where models favor natural-language user claims over numerical sensor data. It proposes geometric audit metrics (CIR, AAI) and an inference-time calibration method (GAC) to correct this bias, showing significant improvements in accuracy.
FusionSense introduces a tri-stage near-sensor learning framework for multimodal edge intelligence that jointly reduces compute and communication by using fusion-aware filtering, achieving up to 33× energy savings and significant data-reduction gains on RGB-Depth/LiDAR tasks.
Proposes PDRNN, a modular hybrid AI-assisted pedestrian dead reckoning system that combines a recurrent neural network with separate ML models for orientation, velocity, and distance estimation, with optional radio-based stabilization. Experiments on dynamic sports movement data show superior accuracy and precision compared to classic and ML-based methods.
A developer built a fully offline suitcase robot named Sparky using a Jetson Orin NX and Gemma 4 E4B model, achieving ~200ms cached TTFT and 14-15 tok/s with 30+ sensors feeding into the prompt as natural language, all without network connectivity.
A web-based guitar tuner that utilizes the phone's accelerometer to detect string vibrations and calculate pitch.