Search for: All records

Online Anomaly Detection (OAD) is critical for identifying rare yet important data points in large, dynamic, and complex data streams. A key challenge lies in achieving accurate and consistent detection of anomalies while maintaining computational and memory efficiency. Conventional OAD approaches, which depend on distributional deviations and static thresholds, struggle with model update delays and catastrophic forgetting, leading to missed detections and high false positive rates. To address these limitations, we propose a novel Streaming Anomaly Detection (SAD) method, grounded in a sparse active online learning framework. Our approach uniquely integrates ℓ1,2-norm sparse online learning with CUR decomposition-based active learning, enabling simultaneous fast feature selection and dynamic instance selection. The efficient CUR decomposition further supports real-time residual analysis for anomaly scoring, eliminating the need for manual threshold settings about temporal data distributions. Extensive experiments on diverse streaming datasets demonstrate SAD's superiority, achieving a 14.06% reduction in detection error rates compared to five state-of-the-art competitors. 

Group Fairness-aware Continual Learning (GFCL) aims to eradicate discriminatory predictions against certain demographic groups in a sequence of diverse learning tasks.This paper explores an even more challenging GFCL problem – how to sustain a fair classifier across a sequence of tasks with covariate shifts and unlabeled data. We propose the MacFRL solution, with its key idea to optimizethe sequence of learning tasks. We hypothesize that high-confident learning can be enabled in the optimized task sequence, where the classifier learns from a set of prioritized tasks to glean knowledge, thereby becoming more capable to handle the tasks with substantial distribution shifts that were originally deferred. Theoretical and empirical studies substantiate that MacFRL excels among its GFCL competitors in terms of prediction accuracy and group fair-ness metrics. 

Abstract Analogous of pixels to two-dimensional pictures, voxels—in the form of either small cubes or spheres—are the basic building blocks of three-dimensional objects. However, precise manipulation of viscoelastic bio-ink voxels in three-dimensional space represents a grand challenge in both soft matter science and biomanufacturing. Here, we present a voxelated bioprinting technology that enables the digital assembly of interpenetrating double-network hydrogel droplets made of polyacrylamide/alginate-based or hyaluronic acid/alginate-based polymers. The hydrogels are crosslinked via additive-free and biofriendly click reaction between a pair of stoichiometrically matched polymers carrying norbornene and tetrazine groups, respectively. We develop theoretical frameworks to describe the crosslinking kinetics and stiffness of the hydrogels, and construct a diagram-of-state to delineate their mechanical properties. Multi-channel print nozzles are developed to allow on-demand mixing of highly viscoelastic bio-inks without significantly impairing cell viability. Further, we showcase the distinctive capability of voxelated bioprinting by creating highly complex three-dimensional structures such as a hollow sphere composed of interconnected yet distinguishable hydrogel particles. Finally, we validate the cytocompatibility and in vivo stability of the printed double-network scaffolds through cell encapsulation and animal transplantation.