Qin Lv

Yawen Zhang, Michael Hannigan, Qin Lv

Air pollution is a major global environmental health threat, in particular for people who live or work near pollution sources. Areas adjacent to pollution sources often have high ambient pollution concentrations, and those areas are commonly referred to as air pollution hotspots. Detecting and characterizing pollution hotspots are of great importance for air quality management, but are challenging due to the high spatial and temporal variability of air pollutants. In this work, we explore the use of mobile sensing data (i.e., air quality sensors installed on vehicles) to detect pollution hotspots. One major challenge with mobile sensing data is uneven sampling, i.e., data collection can vary by both space and time. To address this challenge, we propose a two-step approach to detect hotspots from mobile sensing data, which includes local spike detection and sample-weighted clustering. Essentially, this approach tackles the uneven sampling issue by weighting samples based on their spatial frequency and temporal hit rate, so as to identify robust and persistent hotspots. To contextualize the hotspots and discover potential pollution source characteristics, we explore a variety of cross-domain urban data and extract features from them. As a soft-validation of the extracted features, we build hotspot inference models for cities with and without mobile sensing data. Evaluation results using real-world mobile sensing air quality data as well as cross-domain urban data demonstrate the effectiveness of our approach in detecting and inferring pollution hotspots. Furthermore, the empirical analysis of hotspots and source features yields useful insights regarding neighborhood pollution sources.

Chenyu Wang, Shuo Yan, Yixuan Chen et al.

Video generation using diffusion-based models is constrained by high computational costs due to the frame-wise iterative diffusion process. This work presents a Diffusion Reuse MOtion (Dr. Mo) network to accelerate latent video generation. Our key discovery is that coarse-grained noises in earlier denoising steps have demonstrated high motion consistency across consecutive video frames. Following this observation, Dr. Mo propagates those coarse-grained noises onto the next frame by incorporating carefully designed, lightweight inter-frame motions, eliminating massive computational redundancy in frame-wise diffusion models. The more sensitive and fine-grained noises are still acquired via later denoising steps, which can be essential to retain visual qualities. As such, deciding which intermediate steps should switch from motion-based propagations to denoising can be a crucial problem and a key tradeoff between efficiency and quality. Dr. Mo employs a meta-network named Denoising Step Selector (DSS) to dynamically determine desirable intermediate steps across video frames. Extensive evaluations on video generation and editing tasks have shown that Dr. Mo can substantially accelerate diffusion models in video tasks with improved visual qualities.

Anrui Chen, Ruijun Huang, Xin Zhang et al.

Mixture-of-Experts (MoE) architectures are often considered a natural fit for continual learning because sparse routing should localize updates and reduce interference, yet MoE Transformers still forget substantially even with sparse, well-balanced expert utilization. We attribute this gap to a pre-routing bottleneck: multi-head attention concatenates head-specific signals into a single post-attention router input, forcing routing to act on co-occurring feature compositions rather than separable head channels. We show that this router input simultaneously encodes multiple separately decodable semantic and structural factors with uneven head support, and that different feature compositions induce weakly aligned parameter-gradient directions; as a result, routing maps many distinct compositions to the same route. We quantify this collision effect via a route-wise effective composition number $N_{eff}$ and find that higher $N_{eff}$ is associated with larger old-task loss increases after continual training. Motivated by these findings, we propose MH-MoE, which performs head-wise routing over sub-representations to increase routing granularity and reduce composition collisions. On TRACE with Qwen3-0.6B/8B, MH-MoE effectively mitigates forgetting, reducing BWT on Qwen3-0.6B from 11.2% (LoRAMoE) to 4.5%.

Ruijun Huang, Fang Dong, Xin Zhang et al.

Mixture-of-Experts (MoE) architectures scale Large Language Models via expert specialization induced by conditional computation. In practice, however, expert specialization often fails: some experts become functionally similar, while others functioning as de facto shared experts, limiting the effective capacity and model performance. In this work, we analysis from a spectral perspective on parameter and gradient spaces, uncover that (1) experts share highly overlapping dominant spectral components in their parameters, (2) dominant gradient subspaces are strongly aligned across experts, driven by ubiquitous low-rank structure in human corpus, and (3) gating mechanisms preferentially route inputs along these dominant directions, further limiting specialization. To address this, we propose Spectral-Decoupled MoE (SD-MoE), which decomposes both parameter and gradient in the spectral space. SD-MoE improves performance across downstream tasks, enables effective expert specialization, incurring minimal additional computation, and can be seamlessly integrated into a wide range of existing MoE architectures, including Qwen and DeepSeek.

Zhendong Huang, Hengjie Cao, Fang Dong et al.

Gradient signals in LLM training are highly anisotropic: recurrent linguistic structure concentrates energy into a small set of dominant spectral directions, while context specific information resides in a long tail. We show that this spike tail separation persists throughout training, with the spike occupying only about 1.5% of directions yet dominating optimizer statistics. This dominance suppresses tail learning by contracting tail updates through second moment normalization and tightening the globally stable learning rate bound. Motivated by this analysis, we propose Spectra, a spike aware optimizer that suppresses the dominant low rank spike subspace without amplifying the noise sensitive spectral tail. Spectra tracks the spike subspace via cached, warm started power iteration and applies low rank spectral shaping with negligible overhead and substantially reduced optimizer state memory. On LLaMA3 8B trained on 50B tokens, Spectra reaches the same target loss 30% faster than AdamW, reduces per step end to end overhead by 0.7%, cuts optimizer state memory by 49.25%, and improves average downstream accuracy by 1.62%. Compared to Muon, Spectra is 5.1x faster in optimizer processing time, achieves a lower final loss, and improves average accuracy by 0.66%.

Hengjie Cao, Zhendong Huang, Mengyi Chen et al.

Large language models trained on natural language exhibit pronounced anisotropy: a small number of directions concentrate disproportionate energy, while the remaining dimensions form a broad semantic tail. In low-bit training regimes, this geometry becomes numerically unstable. Because blockwise quantization scales are determined by extreme elementwise magnitudes, dominant directions stretch the dynamic range, compressing long-tail semantic variation into narrow numerical bins. We show that this instability is primarily driven by a coherent rank-one mean bias, which constitutes the dominant component of spectral anisotropy in LLM representations. This mean component emerges systematically across layers and training stages and accounts for the majority of extreme activation magnitudes, making it the principal driver of dynamic-range inflation under low precision. Crucially, because the dominant instability is rank-one, it can be eliminated through a simple source-level mean-subtraction operation. This bias-centric conditioning recovers most of the stability benefits of SVD-based spectral methods while requiring only reduction operations and standard quantization kernels. Empirical results on FP4 (W4A4G4) training show that mean removal substantially narrows the loss gap to BF16 and restores downstream performance, providing a hardware-efficient path to stable low-bit LLM training.

Yubin Shi, Yixuan Chen, Mingzhi Dong et al.

Despite their prevalence in deep-learning communities, over-parameterized models convey high demands of computational costs for proper training. This work studies the fine-grained, modular-level learning dynamics of over-parameterized models to attain a more efficient and fruitful training strategy. Empirical evidence reveals that when scaling down into network modules, such as heads in self-attention models, we can observe varying learning patterns implicitly associated with each module's trainability. To describe such modular-level learning capabilities, we introduce a novel concept dubbed modular neural tangent kernel (mNTK), and we demonstrate that the quality of a module's learning is tightly associated with its mNTK's principal eigenvalue $λ_{\max}$. A large $λ_{\max}$ indicates that the module learns features with better convergence, while those miniature ones may impact generalization negatively. Inspired by the discovery, we propose a novel training strategy termed Modular Adaptive Training (MAT) to update those modules with their $λ_{\max}$ exceeding a dynamic threshold selectively, concentrating the model on learning common features and ignoring those inconsistent ones. Unlike most existing training schemes with a complete BP cycle across all network modules, MAT can significantly save computations by its partially-updating strategy and can further improve performance. Experiments show that MAT nearly halves the computational cost of model training and outperforms the accuracy of baselines.

Alexander Hoelzemann, Julia Lee Romero, Marius Bock et al.

We present a benchmark dataset for evaluating physical human activity recognition methods from wrist-worn sensors, for the specific setting of basketball training, drills, and games. Basketball activities lend themselves well for measurement by wrist-worn inertial sensors, and systems that are able to detect such sport-relevant activities could be used in applications toward game analysis, guided training, and personal physical activity tracking. The dataset was recorded for two teams from separate countries (USA and Germany) with a total of 24 players who wore an inertial sensor on their wrist, during both repetitive basketball training sessions and full games. Particular features of this dataset include an inherent variance through cultural differences in game rules and styles as the data was recorded in two countries, as well as different sport skill levels, since the participants were heterogeneous in terms of prior basketball experience. We illustrate the dataset's features in several time-series analyses and report on a baseline classification performance study with two state-of-the-art deep learning architectures.

Yingying Zhao, Yuhu Chang, Yutian Lu et al.

Emotion recognition in smart eyewear devices is highly valuable but challenging. One key limitation of previous works is that the expression-related information like facial or eye images is considered as the only emotional evidence. However, emotional status is not isolated; it is tightly associated with people's visual perceptions, especially those sentimental ones. However, little work has examined such associations to better illustrate the cause of different emotions. In this paper, we study the emotionship analysis problem in eyewear systems, an ambitious task that requires not only classifying the user's emotions but also semantically understanding the potential cause of such emotions. To this end, we devise EMOShip, a deep-learning-based eyewear system that can automatically detect the wearer's emotional status and simultaneously analyze its associations with semantic-level visual perceptions. Experimental studies with 20 participants demonstrate that, thanks to the emotionship awareness, EMOShip not only achieves superior emotion recognition accuracy over existing methods (80.2% vs. 69.4%), but also provides a valuable understanding of the cause of emotions. Pilot studies with 20 participants further motivate the potential use of EMOShip to empower emotion-aware applications, such as emotionship self-reflection and emotionship life-logging.

Yuhu Chang, Yingying Zhao, Mingzhi Dong et al.

This work presents MemX: a biologically-inspired attention-aware eyewear system developed with the goal of pursuing the long-awaited vision of a personalized visual Memex. MemX captures human visual attention on the fly, analyzes the salient visual content, and records moments of personal interest in the form of compact video snippets. Accurate attentive scene detection and analysis on resource-constrained platforms is challenging because these tasks are computation and energy intensive. We propose a new temporal visual attention network that unifies human visual attention tracking and salient visual content analysis. Attention tracking focuses computation-intensive video analysis on salient regions, while video analysis makes human attention detection and tracking more accurate. Using the YouTube-VIS dataset and 30 participants, we experimentally show that MemX significantly improves the attention tracking accuracy over the eye-tracking-alone method, while maintaining high system energy efficiency. We have also conducted 11 in-field pilot studies across a range of daily usage scenarios, which demonstrate the feasibility and potential benefits of MemX.

Yingying Zhao, Mingzhi Dong, Yujiang Wang et al.

Deep-learning-based video processing has yielded transformative results in recent years. However, the video analytics pipeline is energy-intensive due to high data rates and reliance on complex inference algorithms, which limits its adoption in energy-constrained applications. Motivated by the observation of high and variable spatial redundancy and temporal dynamics in video data streams, we design and evaluate an adaptive-resolution optimization framework to minimize the energy use of multi-task video analytics pipelines. Instead of heuristically tuning the input data resolution of individual tasks, our framework utilizes deep reinforcement learning to dynamically govern the input resolution and computation of the entire video analytics pipeline. By monitoring the impact of varying resolution on the quality of high-dimensional video analytics features, hence the accuracy of video analytics results, the proposed end-to-end optimization framework learns the best non-myopic policy for dynamically controlling the resolution of input video streams to globally optimize energy efficiency. Governed by reinforcement learning, optical flow is incorporated into the framework to minimize unnecessary spatio-temporal redundancy that leads to re-computation, while preserving accuracy. The proposed framework is applied to video instance segmentation which is one of the most challenging computer vision tasks, and achieves better energy efficiency than all baseline methods of similar accuracy on the YouTube-VIS dataset.

Upasana Dutta, Rhett Hanscom, Jason Shuo Zhang et al.

Social media platforms have been exploited to conduct election interference in recent years. In particular, the Russian-backed Internet Research Agency (IRA) has been identified as a key source of misinformation spread on Twitter prior to the 2016 U.S. presidential election. The goal of this research is to understand whether general Twitter users changed their behavior in the year following first contact from an IRA account. We compare the before and after behavior of contacted users to determine whether there were differences in their mean tweet count, the sentiment of their tweets, and the frequency and sentiment of tweets mentioning @realDonaldTrump or @HillaryClinton. Our results indicate that users overall exhibited statistically significant changes in behavior across most of these metrics, and that those users that engaged with the IRA generally showed greater changes in behavior.

Jason Shuo Zhang, Chenhao Tan, Qin Lv

Intergroup contact has long been considered as an effective strategy to reduce prejudice between groups. However, recent studies suggest that exposure to opposing groups in online platforms can exacerbate polarization. To further understand the behavior of individuals who actively engage in intergroup contact in practice, we provide a large-scale observational study of intragroup behavioral differences between members with and without intergroup contact. We leverage the existing structure of NBA-related discussion forums on Reddit to study the context of professional sports. We identify fans of each NBA team as members of a group and trace whether they have intergroup contact. Our results show that members with intergroup contact use more negative and abusive language in their affiliated group than those without such contact, after controlling for activity levels. We further quantify different levels of intergroup contact and show that there may exist nonlinear mechanisms regarding how intergroup contact relates to intragroup behavior. Our findings provide complementary evidence to experimental studies in a novel context and also shed light on possible reasons for the different outcomes in prior studies.

Jason Shuo Zhang, Mike Gartrell, Richard Han et al.

In this paper, we present GEVR, the first Group Event Venue Recommendation system that incorporates mobility via individual location traces and context information into a "social-based" group decision model to provide venue recommendations for groups of mobile users. Our study leverages a real-world dataset collected using the OutWithFriendz mobile app for group event planning, which contains 625 users and over 500 group events. We first develop a novel "social-based" group location prediction model, which adaptively applies different group decision strategies to groups with different social relationship strength to aggregate each group member's location preference, to predict where groups will meet. Evaluation results show that our prediction model not only outperforms commonly used and state-of-the-art group decision strategies with over 80% accuracy for predicting groups' final meeting location clusters, but also provides promising qualities in cold-start scenarios. We then integrate our prediction model with the Foursquare Venue Recommendation API to construct an event venue recommendation framework for groups of mobile users. Evaluation results show that GEVR outperforms the comparative models by a significant margin.

Yingying Zhao, Dongsheng Li, Qin Lv et al.

Recommender systems have become an indispensable component in online services during recent years. Effective recommendation is essential for improving the services of various online business applications. However, serious privacy concerns have been raised on recommender systems requiring the collection of users' private information for recommendation. At the same time, the success of e-commerce has generated massive amounts of information, making scalability a key challenge in the design of recommender systems. As such, it is desirable for recommender systems to protect users' privacy while achieving high-quality recommendations with low-complexity computations. This paper proposes a scalable privacy-preserving item-based top-N recommendation solution, which can achieve high-quality recommendations with reduced computation complexity while ensuring that users' private information is protected. Furthermore, the computation complexity of the proposed method increases slowly as the number of users increases, thus providing high scalability for privacy-preserving recommender systems. More specifically, the proposed approach consists of two key components: (1) MinHash-based similarity estimation and (2) client-side privacy-preserving prediction generation. Our theoretical and experimental analysis using real-world data demonstrates the efficiency and effectiveness of the proposed approach.

Dongsheng Li, Chao Chen, Qin Lv et al.

Collaborative filtering (CF) is a popular technique in today's recommender systems, and matrix approximation-based CF methods have achieved great success in both rating prediction and top-N recommendation tasks. However, real-world user-item rating matrices are typically sparse, incomplete and noisy, which introduce challenges to the algorithm stability of matrix approximation, i.e., small changes in the training data may significantly change the models. As a result, existing matrix approximation solutions yield low generalization performance, exhibiting high error variance on the training data, and minimizing the training error may not guarantee error reduction on the test data. This paper investigates the algorithm stability problem of matrix approximation methods and how to achieve stable collaborative filtering via stable matrix approximation. We present a new algorithm design framework, which (1) introduces new optimization objectives to guide stable matrix approximation algorithm design, and (2) solves the optimization problem to obtain stable approximation solutions with good generalization performance. Experimental results on real-world datasets demonstrate that the proposed method can achieve better accuracy compared with state-of-the-art matrix approximation methods and ensemble methods in both rating prediction and top-N recommendation tasks.

Jiangtao Wang, Feng Wang, Yasha Wang et al.

Task allocation is a major challenge in Mobile Crowd Sensing (MCS). While previous task allocation approaches follow either the opportunistic or participatory mode, this paper proposes to integrate these two complementary modes in a two-phased hybrid framework called HyTasker. In the offline phase, a group of workers (called opportunistic workers) are selected, and they complete MCS tasks during their daily routines (i.e., opportunistic mode). In the online phase, we assign another set of workers (called participatory workers) and require them to move specifically to perform tasks that are not completed by the opportunistic workers (i.e., participatory mode). Instead of considering these two phases separately, HyTasker jointly optimizes them with a total incentive budget constraint. In particular, when selecting opportunistic workers in the offline phase of HyTasker, we propose a novel algorithm that simultaneously considers the predicted task assignment for the participatory workers, in which the density and mobility of participatory workers are taken into account. Experiments on a real-world mobility dataset demonstrate that HyTasker outperforms other methods with more completed tasks under the same budget constraint.

Jiangtao Wang, Yasha Wang, Daqing Zhang et al.

With the advent of seamless connection of human, machine, and smart things, there is an emerging trend to leverage the power of crowds (e.g., citizens, mobile devices, and smart things) to monitor what is happening in a city, understand how the city is evolving, and further take actions to enable better quality of life, which is referred to as Crowd-Powered Smart City (CPSC). In this article, we provide a literature review for CPSC and identify future research opportunities. Specifically, we first define the concepts with typical CPSC applications. Then, we present the main characteristics of CPSC and further highlight the research issues. In the end, we point out existing limitations which can inform and guide future research directions.

Shuo Zhang, Khaled Alanezi, Mike Gartrell et al.

The wide adoption of smartphones and mobile applications has brought significant changes to not only how individuals behave in the real world, but also how groups of users interact with each other when organizing group events. Understanding how users make event decisions as a group and identifying the contributing factors can offer important insights for social group studies and more effective system and application design for group event scheduling. In this work, we have designed a new mobile application called OutWithFriendz, which enables users of our mobile app to organize group events, invite friends, suggest and vote on event time and venue. We have deployed OutWithFriendz at both Apple App Store and Google Play, and conducted a large-scale user study spanning over 500 users and 300 group events. Our analysis has revealed several important observations regarding group event planning process including the importance of user mobility, individual preferences, host preferences, and group voting process.

Qi Liu, Yawen Zhang, Qin Lv et al.

During summer, melt ponds have a significant influence on Arctic sea-ice albedo. The melt pond fraction (MPF) also has the ability to forecast the Arctic sea-ice in a certain period. It is important to retrieve accurate melt pond fraction (MPF) from satellite data for Arctic research. This paper proposes a satellite MPF retrieval model based on the multi-layer neural network, named MPF-NN. Our model uses multi-spectral satellite data as model input and MPF information from multi-site and multi-period visible imagery as prior knowledge for modeling. It can effectively model melt ponds evolution of different regions and periods over the Arctic. Evaluation results show that the MPF retrieved from MODIS data using the proposed model has an RMSE of 3.91% and a correlation coefficient of 0.73. The seasonal distribution of MPF is also consistent with previous results.

Homa Hosseinmardi, Sabrina Arredondo Mattson, Rahat Ibn Rafiq et al.

Cyberbullying is a growing problem affecting more than half of all American teens. The main goal of this paper is to investigate fundamentally new approaches to understand and automatically detect and predict incidents of cyberbullying in Instagram, a media-based mobile social network. In this work, we have collected a sample data set consisting of Instagram images and their associated comments. We then designed a labeling study and employed human contributors at the crowd-sourced CrowdFlower website to label these media sessions for cyberbullying. A detailed analysis of the labeled data is then presented, including a study of relationships between cyberbullying and a host of features such as cyberaggression, profanity, social graph features, temporal commenting behavior, linguistic content, and image content. Using the labeled data, we further design and evaluate the performance of classifiers to automatically detect and pre- dict incidents of cyberbullying and cyberaggression.

Alireza Mahdian, Richard Han, Qin Lv et al.

This paper presents MyZone, a private online social network for relatively small, closely-knit communities. MyZone has three important distinguishing features. First, users keep the ownership of their data and have complete control over maintaining their privacy. Second, MyZone is free from any possibility of content censorship and is highly resilient to any single point of disconnection. Finally, MyZone minimizes deployment cost by minimizing its computation, storage and network bandwidth requirements. It incorporates both a P2P architecture and a centralized architecture in its design ensuring high availability, security and privacy. A prototype of MyZone was deployed over a period of 40 days with a membership of more than one hundred users. The paper provides a detailed evaluation of the results obtained from this deployment.