Tag
Oura withdrew its $2.2 billion IPO after completing its roadshow, just hours before pricing, officially citing market uncertainty. However, the prospectus showed that 73% of the offering consisted of existing shareholders cashing out, while 93% of the proceeds were earmarked to pay taxes on employees’ behalf—making it, in substance, an exit rather than a fundraising exercise. The article provides an in-depth analysis of the controversy surrounding Oura’s valuation, the tensions between its hardware and software business models, and the competitive threats posed by giants such as Apple, Samsung, and Google.
The paper introduces ProMemAgent, the first proactive LLM agent incorporating context-aware prospective memory over sensory perceptions (egocentric video and audio) for long-horizon assistance, along with CPMBench, a 1,000-instance benchmark evaluating delayed-intention assistance across eight everyday categories.
This paper introduces a streamlined pipeline for training and deploying machine learning models on the WeBe Band wearable device, focusing on system-level automation and hardware-aware optimization to enable rapid iteration and on-device evaluation for edge AI applications.
This paper proposes a multimodal framework called PART for predicting injury risk in tennis players using machine learning on data from wearables, questionnaires, and video analysis, showing strong performance in holistic athlete assessment.
This paper investigates reducing the computational complexity of deep neural networks for EEG analysis on wearable devices by applying parameter quantization and electrode reduction techniques, demonstrating significant complexity reduction with minimal accuracy loss for epileptic seizure detection.
VCR is a self-supervised framework that learns robust representations from incomplete wearable signals using orthogonal tokenization and missing-aware mixture-of-experts, improving performance under modality missingness.
Researchers at MIT developed artificial muscles from woven fibers with embedded pumps that use electric fields to move fluid, enabling smooth, quiet motion for soft robotic arms and wearable devices.
This paper introduces LifeDialBench, a novel benchmark for evaluating memory capabilities in continuous lifelog scenarios using wearable devices, and proposes an online evaluation protocol that enforces temporal causality. Key finding: sophisticated memory systems underperform simple RAG baselines, highlighting the importance of high-fidelity context preservation over lossy compression.