Smartphones provide an attractive yet challenging platform for human activity recognition (HAR). They are ubiquitous, but also limit the input of HAR systems to a single IMU. These systems are also challenged by the inherent diversity of human activities and varying phone placement on the user’s body. This results in traditional smartphone HAR systems having limited personalization potential or imposing a high user burden. We propose ActivitySeeker, a personalized smartphone HAR system that combines self-supervised activity discovery and low-burden user interaction to collaboratively label IMU data and adapt HAR models to individual users on-device through transfer learning. We evaluated ActivitySeeker through simulated online learning and in-the-wild user experiments, where it discovered 95.5% of personal activity types and achieved high recognition accuracy (93.3%) while maintaining a positive user experience. Leveraging the synergy between user and smartphone, ActivitySeeker opens up new possibilities for HAR-based applications like fitness, health and personalized recommendation.

Heterogeneous transfer learning is an effective approach for medical imaging problems with limited data and scarce public homogeneous resources, yet selecting the optimal domain for feature extraction remains an open, often intuition-driven challenge. This study proposes and validates a set of quantitative domain similarity measurements to a priori identify the most suitable intermediate domain for early laryngeal cancer detection within a two-step heterogeneous transfer learning (THTL) framework, thereby avoiding computationally expensive trial-and-error training. We introduce eight domain similarity measurements to access the similarity between intermediate domains and the target domain. Multiple common medical imaging modalities, including angiography, chest radiographs, lung computed tomography (CT), brain magnetic resonance imaging (MRI), pathological section images, diabetic retinopathy fundus images, skin lesion images, and gastroenteroscopy, are served as candidate intermediate domains. The resulting similarity scores are ranked and compared with actual THTL performance rankings. Finally, we employ normalized discounted cumulative gain (NDCG) to determine the most predictive measurement. Our findings reveal that Earth Mover’s Distance (EMD) is the most effective domain similarity measurement for grayscale images, while cosine similarity based on global features extracted from a convolutional neural network (CNN) is optimal for RGB images. Using these measurements, angiography and skin lesion images are identified as the most beneficial intermediate domains. This work establishes a validated, data-driven methodology that enables future researchers to replace subjective intuition in domain selection, thereby saving substantial computational resources while improving model performance.

Activation sparsity denotes the existence of substantial weakly-contributed neurons within feed-forward networks of large language models (LLMs), providing wide potential benefits such as computation acceleration. However, existing works lack thorough quantitative studies on this useful property, in terms of both its measurement and influential factors. In this paper, we address three underexplored research questions: (1) How can activation sparsity be measured more accurately? (2) How is activation sparsity affected by the model architecture and training process? (3) How can we build a more sparsely activated and efficient LLM? Specifically, we develop a generalizable and performance-friendly metric, named CETT-PPL-1%, to measure activation sparsity. Based on CETT-PPL-1%, we quantitatively study the influence of various factors and observe several important phenomena, such as the convergent power-law relationship between sparsity and training data amount, the higher competence of ReLU activation than mainstream SiLU activation, the potential sparsity merit of a small width-depth ratio, and the scale insensitivity of activation sparsity. Finally, we provide implications for building sparse and effective LLMs, and demonstrate the reliability of our findings by training a 2.4B model with a sparsity ratio of 93.52%, showing 4.1$\times$ speedup compared with its dense version. The codes and checkpoints are available at https://github.com/thunlp/SparsingLaw/.

In the digital world, gesture recognition plays a crucial role in human-computer interaction (HCI). In this paper, we propose an innovative wearable air-writing system that allows users to write the English alphabet and Arabic numerals in free space without using any predefined gestures or rules. Based on an Inertial Measurement Unit (IMU), the proposed air-writing wearable device uses the constrained dynamic time warping (cDTW) algorithm for the distance measure and K-nearest neighbors (KNN) as the classifier. In addition, to increase the recognition accuracy and meet HCI requirements, we develop a novel method that allows users to rapidly switch to correct recognition results when the initial results are erroneous. In the experiment, the accuracy rate is 88.9% for the alphabet and 10 decimal digits in the user-dependent condition, and the recognition is in real-time, consuming only 0.427s for each character, which is superior to many other approaches that employ classic DTW or FastDTW. With the proposed HCI design, character input accuracy of over 95% can be obtained in about 1 second. We also simulated the application scenarios of Parkinson’s disease patients and obtained a high accuracy rate of 85.4%. Besides, we explored the variety of K values in KNN and w values in cDTW, and propose a multi-template system that gives new optimization directions for the KNN-cDTW algorithm.

Xu Han, Yuqi Luo, Weize Chen, Zhiyuan Liu, Maosong Sun, Zhou Botong, Hao Fei, Suncong Zheng. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022.

Gesture recognition has been a popular research field under the trend of IoT and intelligent devices. Air-writing is the most challenging and crucial topic in the gesture recognition field. In this paper, we propose a wearable air-writing system that makes users can write the English alphabet in the three-dimensional space without any write rules. The proposed system is based on the Inertial Measurement Unit (IMU), and it uses dynamic time warping (DTW) as the main recognition algorithm. In addition, to improve the recognition accuracy and take a better advantage of the DTW algorithm, we present an adjustment system that gives some new optimization methods to the application of IMU and DTW. In the experiment, the accuracy of recognition is 84.6% for the uppercase alphabet (from `A' to `Z') in user-dependent case. And we also confirmed that the recognition method only based on the DTW algorithm is one kind of user-dependent methods, which means this method is heavily dependent on personalization.