CurvAGN: Curvature-based Adaptive Graph Neural Networks for Predicting Protein-Ligand Binding Affinity.

Accurately predicting the binding affinity between proteins and ligands is crucial for drug discovery. Recent advances in graph neural networks (GNNs) have made significant progress in learning representations of protein-ligand complexes to estimate binding affinities. To improve the performance of GNNs, there frequently needs to look into protein-ligand complexes from geometric perspectives. While the "off-the-shelf" GNNs could incorporate some basic geometric structures of molecules, such as distances and angles, through modeling the complexes as homophilic graphs, these solutions seldom take into account the higher-level geometric attributes like curvatures and homology, and also heterophilic interactions.To address these limitations, we introduce the Curvature-based Adaptive Graph Neural Network (CurvAGN). This GNN comprises two components: a curvature block and an adaptive attention guided neural block (AGN). The curvature block encodes multiscale curvature informaton, then the AGN, based on an adaptive graph attention mechanism, incorporates geometry structure including angle, distance, and multiscale curvature, long-range molecular interactions, and heterophily of the graph into the protein-ligand complex representation. We demonstrate the superiority of our proposed model through experiments conducted on the PDBbind-V2016 core dataset.

Using Mobile Sensing to Test Clinical Models of Depression, Social Anxiety, State Affect, and Social Isolation Among College Students.

BACKGROUND: Research in psychology demonstrates a strong link between state affect (moment-to-moment experiences of positive or negative emotionality) and trait affect (eg, relatively enduring depression and social anxiety symptoms), and a tendency to withdraw (eg, spending time at home). However, existing work is based almost exclusively on static, self-reported descriptions of emotions and behavior that limit generalizability. Despite adoption of increasingly sophisticated research designs and technology (eg, mobile sensing using a global positioning system [GPS]), little research has integrated these seemingly disparate forms of data to improve understanding of how emotional experiences in everyday life are associated with time spent at home, and whether this is influenced by depression or social anxiety symptoms. OBJECTIVE: We hypothesized that more time spent at home would be associated with more negative and less positive affect. METHODS: We recruited 72 undergraduate participants from a southeast university in the United States. We assessed depression and social anxiety symptoms using self-report instruments at baseline. An app (Sensus) installed on participants' personal mobile phones repeatedly collected in situ self-reported state affect and GPS location data for up to 2 weeks. Time spent at home was a proxy for social isolation. RESULTS: We tested separate models examining the relations between state affect and time spent at home, with levels of depression and social anxiety as moderators. Models differed only in the temporal links examined. One model focused on associations between changes in affect and time spent at home within short, 4-hour time windows. The other 3 models focused on associations between mean-level affect within a day and time spent at home (1) the same day, (2) the following day, and (3) the previous day. Overall, we obtained many of the expected main effects (although there were some null effects), in which higher social anxiety was associated with more time or greater likelihood of spending time at home, and more negative or less positive affect was linked to longer homestay. Interactions indicated that, among individuals higher in social anxiety, higher negative affect and lower positive affect within a day was associated with greater likelihood of spending time at home the following day. CONCLUSIONS: Results demonstrate the feasibility and utility of modeling the relationship between affect and homestay using fine-grained GPS data. Although these findings must be replicated in a larger study and with clinical samples, they suggest that integrating repeated state affect assessments in situ with continuous GPS data can increase understanding of how actual homestay is related to affect in everyday life and to symptoms of anxiety and depression.

Learning Discriminative Features for Crowd Counting.

Crowd counting models in highly congested areas confront two main challenges: weak localization ability and difficulty in differentiating between foreground and background, leading to inaccurate estimations. The reason is that objects in highly congested areas are normally small and high-level features extracted by convolutional neural networks are less discriminative to represent small objects. To address these problems, we propose a learning discriminative features framework for crowd counting, which is composed of a masked feature prediction module (MPM) and a supervised pixel-level contrastive learning module (CLM). The MPM randomly masks feature vectors in the feature map and then reconstructs them, allowing the model to learn about what is present in the masked regions and improving the model's ability to localize objects in high-density regions. The CLM pulls targets close to each other and pushes them far away from background in the feature space, enabling the model to discriminate foreground objects from background. Additionally, the proposed modules can be beneficial in various computer vision tasks, such as crowd counting and object detection, where dense scenes or cluttered environments pose challenges to accurate localization. The proposed two modules are plug-and-play, incorporating the proposed modules into existing models can potentially boost their performance in these scenarios.

AA-forecast: anomaly-aware forecast for extreme events.

Time series models often are impacted by extreme events and anomalies, both prevalent in real-world datasets. Such models require careful probabilistic forecasts, which is vital in risk management for extreme events such as hurricanes and pandemics. However, it's challenging to automatically detect and learn from extreme events and anomalies for large-scale datasets which often results in extra manual efforts. Here, we propose an anomaly-aware forecast framework that leverages the effects of anomalies to improve its prediction accuracy during the presence of extreme events. Our model has trained to extract anomalies automatically and incorporates them through an attention mechanism to increase the accuracy of forecasts during extreme events. Moreover, the framework employs a dynamic uncertainty optimization algorithm that reduces the uncertainty of forecasts in an online manner. The proposed framework demonstrated consistent superior accuracy with less uncertainty on three datasets with different varieties of anomalies over the current prediction models.

Temporal Output Discrepancy for Loss Estimation-Based Active Learning.

While deep learning succeeds in a wide range of tasks, it highly depends on the massive collection of annotated data which is expensive and time-consuming. To lower the cost of data annotation, active learning has been proposed to interactively query an oracle to annotate a small proportion of informative samples in an unlabeled dataset. Inspired by the fact that the samples with higher loss are usually more informative to the model than the samples with lower loss, in this article we present a novel deep active learning approach that queries the oracle for data annotation when the unlabeled sample is believed to incorporate high loss. The core of our approach is a measurement temporal output discrepancy (TOD) that estimates the sample loss by evaluating the discrepancy of outputs given by models at different optimization steps. Our theoretical investigation shows that TOD lower-bounds the accumulated sample loss thus it can be used to select informative unlabeled samples. On basis of TOD, we further develop an effective unlabeled data sampling strategy as well as an unsupervised learning criterion for active learning. Due to the simplicity of TOD, our methods are efficient, flexible, and task-agnostic. Extensive experimental results demonstrate that our approach achieves superior performances than the state-of-the-art active learning methods on image classification and semantic segmentation tasks. In addition, we show that TOD can be utilized to select the best model of potentially the highest testing accuracy from a pool of candidate models.

Mobile Sensing in the COVID-19 Era: A Review.

Background: During the COVID-19 pandemic, mobile sensing and data analytics techniques have demonstrated their capabilities in monitoring the trajectories of the pandemic, by collecting behavioral, physiological, and mobility data on individual, neighborhood, city, and national scales. Notably, mobile sensing has become a promising way to detect individuals' infectious status, track the change in long-term health, trace the epidemics in communities, and monitor the evolution of viruses and subspecies. Methods: We followed the PRISMA practice and reviewed 60 eligible papers on mobile sensing for monitoring COVID-19. We proposed a taxonomy system to summarize literature by the time duration and population scale under mobile sensing studies. Results: We found that existing literature can be naturally grouped in four clusters, including remote detection, long-term tracking, contact tracing, and epidemiological study. We summarized each group and analyzed representative works with regard to the system design, health outcomes, and limitations on techniques and societal factors. We further discussed the implications and future directions of mobile sensing in communicable diseases from the perspectives of technology and applications. Conclusion: Mobile sensing techniques are effective, efficient, and flexible to surveil COVID-19 in scales of time and populations. In the post-COVID era, technical and societal issues in mobile sensing are expected to be addressed to improve healthcare and social outcomes.

Data-Driven Robust Control of Discrete-Time Uncertain Linear Systems via Off-Policy Reinforcement Learning.

This paper presents a model-free solution to the robust stabilization problem of discrete-time linear dynamical systems with bounded and mismatched uncertainty. An optimal controller design method is derived to solve the robust control problem, which results in solving an algebraic Riccati equation (ARE). It is shown that the optimal controller obtained by solving the ARE can robustly stabilize the uncertain system. To develop a model-free solution to the translated ARE, off-policy reinforcement learning (RL) is employed to solve the problem in hand without the requirement of system dynamics. In addition, the comparisons between on- and off-policy RL methods are presented regarding the robustness to probing noise and the dependence on system dynamics. Finally, a simulation example is carried out to validate the efficacy of the presented off-policy RL approach.

DBSDA : Lowering the Bound of Misclassification Rate for Sparse Linear Discriminant Analysis via Model Debiasing.

Linear discriminant analysis (LDA) is a well-known technique for linear classification, feature extraction, and dimension reduction. To improve the accuracy of LDA under the high dimension low sample size (HDLSS) settings, shrunken estimators, such as Graphical Lasso, can be used to strike a balance between biases and variances. Although the estimator with induced sparsity obtains a faster convergence rate, however, the introduced bias may also degrade the performance. In this paper, we theoretically analyze how the sparsity and the convergence rate of the precision matrix (also known as inverse covariance matrix) estimator would affect the classification accuracy by proposing an analytic model on the upper bound of an LDA misclassification rate. Guided by the model, we propose a novel classifier, DBSDA , which improves classification accuracy through debiasing. Theoretical analysis shows that DBSDA possesses a reduced upper bound of misclassification rate and better asymptotic properties than sparse LDA (SDA). We conduct experiments on both synthetic datasets and real application datasets to confirm the correctness of our theoretical analysis and demonstrate the superiority of DBSDA over LDA, SDA, and other downstream competitors under HDLSS settings.