Online peer-support communities are vital for mental health, but their therapeutic benefit hinges on receiving a timely and helpful first reply. Posts that languish unanswered can exacerbate feelings of distress and abandonment. This paper develops and validates an integrated framework to predict, explain, and reduce this ``reply gap" on Reddit. First, using survival analysis on over 91,000 posts (2018–2025), we show that a deep learning model (DySurv) can accurately predict reply times (C-Index = 0.742), with a post's lexico-semantic content being a far stronger predictor than author history. Second, moving from correlation to causation, we use a causal inference framework on 48,612 posts to estimate the effect of different support types. We find that initial replies providing emotional support are most effective, increasing the odds of a positive user response by 49% (OR=1.49), an effect most pronounced for high-risk users. Third, we operationalize these insights in RiskMatch, a recommender system that routes at-risk posts to historically effective helpers. Rigorous counterfactual evaluation using inverse propensity scoring (IPS)—a method that corrects for biases in historical data—demonstrates that our system reduces the median wait time by 26 minutes for the highest-risk quintile. This work provides a validated, data-driven methodology to build more responsive and effective peer-support ecosystems, offering a concrete pathway to ensure fewer calls for help go unanswered.

When asked explicitly, a Large language model (LLM) may validate your anger-but implicitly, it may still judge that anger as inappropriate.We call this divergence the endorsementexposure gap, and it reveals that LLMs encode hidden norms about which emotions are acceptable in which contexts.To measure these norms systematically, we introduce FEELING RULES ATLAS, a benchmark of 1,320 vignettes spanning 6 institutional settings, 12 roles, 7 emotions, and 5 intensity levels.We pair the benchmark with two probes: explicit norm judgments (APPROPRIATE/INAPPROPRIATE/DEPENDS) and implicit acceptability scored by log-likelihood contrast.Across six model families, we find large cross-model variation in sanctioning thresholds and institutional "norm signatures" not reducible to overall strictness; models that appear similarly lenient explicitly can diverge sharply in implicit judgments.These results establish normative affect: contextconditioned judgments of emotional appropriateness, as a distinct alignment axis, and motivate transparent profiling of feeling rules for emotionally sensitive deployments.

In intimate messaging, how a difficult note is produced signals personal effort and authorship. We examine how AI assistance level (light tone rewrite versus heavy full draft) and a brief sender‑voiced co‑sign disclosure shape receiver attributions and outcomes in two scenarios: apology/repair and boundary requests. Study 1 (N = 152) instrumented authoring to create a curated message corpus with authoring telemetry. Study 2 (N = 704) used those messages in a randomized 2 × 2 receiver experiment (Help by Disclosure) with crossed random effects. Across scenarios, heavier drafting reliably reduced perceived ownership and authenticity; improvements in competence or clarity did not compensate. In apologies, co‑signing a tone rewrite increased authenticity and forgiveness; co‑signing a full draft slightly decreased both. In boundary requests, co‑signing was neutral to mildly negative. Stimulus‑level analyses linked concrete authoring traces to receiver judgments: more idiosyncratic “voice” cues and higher human contribution predicted higher ownership and authenticity; larger drafting distance predicted the reverse. Senders also overestimated the relational benefits of light help in boundary requests relative to receivers. We contribute: (i) scenario‑aware causal estimates of help and disclosure on ownership, authenticity, trust facets, and outcomes; (ii) an empirically grounded attributional account that clarifies why competence gains rarely offset authenticity losses; (iii) evidence of sender–receiver miscalibration under realistic disclosure mixtures; and (iv) scenario‑sensitive design guidance for voice‑preserving defaults, ownership‑restoring scaffolds, and CPM‑aligned disclosure.

Digital eco‑nudges are widely deployed, yet their long‑term efficacy, ethical acceptability, and net environmental impact remain unclear. We report two field studies targeting routine online behaviors under strict parity of message content and delivery budgets. Study 1 shows that minimal, factual tailoring improves compliance over generic prompts when opportunities are defined independently of delivery. Study 2 introduces a privacy‑preserving, on‑device contextual bandit that learns when to act and when to DoNothing, achieving higher compliance at comparable prompt intensity while maintaining autonomy. We operationalize an Ethical–Efficacy Frontier (EEF) to visualize autonomy–effectiveness trade‑offs, and compute an energy Return on Investment (ROI) that nets behavior‑driven savings against measured system overhead. Energy savings are estimated using literature‑calibrated proxies with sensitivity bands, and we discuss the energy trade‑offs of on‑device learning relative to a stylized cloud alternative. We probe short‑term persistence via withdrawal and a brief follow‑up; long‑term habit formation and rebound remain out of scope. We contribute design and reporting practices—ablation parity, opportunity denominators, EEF, and net‑impact accounting—that make digital sustainability interventions more rigorous, transparent, and respectful, advancing sustainable HCI beyond “small changes.”

AI coding assistants help, but developers still spend effort verifying model output. We isolate interface effects by holding a single LLM fixed while N = 60 participants solve three Python tasks with Inline, Chat, or Structured prompting, plus a no‑AI control. AI reduced workload by − 18.2 TLX points and time by 22% (25.0 vs. 32.1 min) and improved correctness (OR = 1.71). Within AI, Inline is fastest and lowest‑load on simple work; Chat yields higher correctness beyond a per‑observation complexity threshold (z ≈ + 0.41) without a time cost; Structured benefits novices at mid complexity. We introduce a mode‑agnostic verification‑load index (failures, time‑to‑first‑compile, churn, pauses, switches) that partially mediates rising stress/fatigue across tasks. We translate these findings into design guidance: adaptive mode orchestration, transparency on demand, and verification aware packaging, and propose reporting verification load alongside outcomes to evaluate interfaces as models evolve.

Federated multitask optimization (FMTO) allows private collaboration for distributed black-box optimization, but often suffers from negative transfer under strong task heterogeneity. We propose FMDLES, a personalized FMTO framework that couples sequence-based relatedness estimation with inverse-guide transfer and local refinement. Each client trains an auxiliary inverse mapping guide that predicts decision vectors from target objective values, providing coarse basin-level proposals without sharing landscapes or solutions. To support collaboration across heterogeneous tasks without exchanging solver-specific parameters, we introduce a solver-parameter-agnostic relatedness signature based on improvement-rate sequences that privately capture optimization dynamics. The server performs personalized Top-K aggregation using this sequence similarity to produce client-specific inverse guides, which are then exploited locally via NN-guided candidate generation followed by local surrogate-assisted differential evolution (LSA-DE). Experiments on Arxiv2017, CEC2022, and a heterogeneous kinematic arm control task show competitive and often improved performance over representative FMTO baselines across the studied heterogeneity settings.

For web platforms facing regulatory scrutiny---from content moderation to ad delivery and recommendations---fairness audits routinely disagree due to metric choice, subgroup granularity, sampling variance, and dataset shift. Point estimates yield brittle pass/fail narratives that are hard to defend in governance contexts. We propose a Bayesian audit-of-audits that pools heterogeneous audits---count-based and metric-only---into interval-valued fairness claims with explicit uncertainty and policy-risk tables aligned to practitioner thresholds. The framework unifies classification and exposure metrics, enforces consistency across coarse and intersectional group definitions via soft coherence constraints, and quantifies the Value-of-Information of prospective audits. We also provide heterogeneity diagnostics and leave-one-audit-out sensitivities. Across a synthetic Audit Zoo, a content-moderation case study on CivilComments--WILDS, and an ad-delivery simulation, our meta-evaluator attains near-nominal coverage with narrower intervals and fewer decision flips than per-audit baselines, while integrating partial-information audits.

Existing Video Anomaly Detection (VAD) methods predominantly rely on a single visual modality and fail to fully leverage multimodal information due to limited available datasets. However, recent advances in video understanding models have enabled multimodal integration into VAD. In this paper, we propose a novel weakly supervised video anomaly detection network based on adaptive fusion of multimodal features (AFM-WSVAD). This network introduces textual information corresponding to the video through the caption generation model SwinBERT. In VAD, effectively modeling long-term temporal dynamics is crucial. Therefore, we designed a two-stage approach to capture this dynamic information. First, the External Attention mechanism is employed to capture long-range dependencies between samples. Then, the Temporal Context Aggregation (TCA) module and the Multi-scale Temporal Network (MTN) are utilized to model the short-term temporal dynamics of visual and textual features. This design enables the model to handle long-range dependencies and complex dynamics. During multimodal fusion, an adaptive fusion strategy and a Multi-scale Convolutional Attention (MSCA) module are employed to highlight key features and minimize noise, thereby enhancing the model’s detection accuracy. AFM-WSVAD achieves significant performance improvements, with an AUC of 85.4% on UCF-Crime ([Formula: see text]%), 98.1% on ShanghaiTech ([Formula: see text]%), and an AP of 81.8% on XD-Violence ([Formula: see text]%).

We present Conformal@K , a model-agnostic calibration layer that provides distribution-free, finite-sample control of Top- \( K \) miss-risk in two-stage recommendation and retrieval. Given arbitrary retrieval and re-ranking scorers, Conformal@K selects monotone budgets so that, at level \(1-\alpha\) , the probability that no relevant item appears in the returned Top- \( K \) is controlled whenever a feasible parameter exists; otherwise, a transparent best-effort mode reports the residual gap with actionable diagnostics. Beyond marginal validity, we introduce overlapping-group guarantees via smoothed, self-normalized estimates, joint two-stage calibration controlling both retrieve-miss and final miss@K, and importance-weighted and windowed variants for covariate shift and temporal dependence. Empirically, on MSLR-WEB10K , Conformal@K tracks target risks across \(\alpha\) and meets global and cohort targets while preserving ranking quality. On POI recommendation ( Gowalla, Foursquare ) under an all-ranking protocol with a display cap ( \(K_{\max}=50\) ), small \(\alpha\) can be infeasible; our method still reduces global and worst-group miss-risk and improves HR@K, explicitly reporting infeasibility gaps. We compare against four fairness-of-exposure baselines, showing that Conformal@K and exposure-fair methods target complementary objectives and compose in practice. Shift-aware and streaming variants stabilize miss-risk under drift. The method drops into existing stacks with audit-friendly diagnostics.

Human–AI collaboration is increasingly integral to a variety of domains where creative ideation unfolds in iterative cycles, yet most existing studies evaluate AI-generated concepts in a single step. This paper addresses the gap by investigating “Creative Momentum Transfer”—how the timing (early vs. late) and labeling (AI-labeled vs. unlabeled) of AI prompts shape multi-round human ideation. In a between-subjects experiment (N = 247), participants proposed solutions for plastic pollution over two rounds, with AI suggestions introduced either at the outset or mid-process and labeled explicitly or not. Results reveal that early AI prompts increase overall creativity but induce stronger anchoring, whereas late AI prompts trigger a mid-round pivot that fosters more divergent thinking yet still boosts final outcomes compared to a no-AI control. Labeling amplifies both subjective and objective adoption of AI ideas, although most participants could detect AI sources even when unlabeled. Furthermore, qualitative interviews highlight nuanced perspectives on perceived ownership, authenticity, and the ways in which labeling triggers deeper scrutiny of the AI’s style. By demonstrating that baseline creativity moderates these effects more robustly than trust in AI, this study advances our theoretical understanding of multi-round human–AI synergy while offering design guidelines for next-generation creativity support systems. We discuss how user-centered design can balance rapid convergence (via early AI) with strategic pivot opportunities (via late AI) and weigh transparent labeling against ethical considerations of authorship and user autonomy.

Generative AI (GenAI) tools have rapidly transformed creative workflows by providing diverse visual outputs, yet a critical and understudied phenomenon Involves the moment-to-moment decisions that artists make to accept or dismiss machine-generated suggestions. In this mixed-methods investigation, we focus on these subtle "override" behaviors within AI-assisted illustration tasks. We recruited 57 digital artists (novice, intermediate, expert) and collected data from multiple sources: automated logs of override actions, concise in-task rationales, post-task surveys, and semi-structured interviews. Our findings indicate that stylistic discordance and technical flaws drive the highest number of rejections, though novices and experts differ markedly in both the frequency and speed of overriding. Regression analyses show that frequent dismissals strengthen perceptions of creative agency but do not consistently improve overall satisfaction; instead, the quality of final outputs emerges as the primary determinant of satisfaction. Interviews reveal how experienced artists swiftly discard suboptimal suggestions to maintain personal style, while novices oscillate between curiosity and fatigue. By mapping these decision points and their outcomes, we illuminate critical aspects of human-AI co-creation and propose design strategies—such as adaptive recommendation filtering and stage-aware interventions—to ease "override fatigue." Although this study centers on digital illustration, our approach and insights can extend to other creative or even scientific domains where iterative human-AI collaboration is essential.

To address the deficiencies of existing weakly supervised video anomaly detection methods in long video anomaly frame localization, feature representation distinction, and feature magnitude consistency, this paper proposes a method based on spatiotemporal dual-memory units and contrastive measurement. The core of this method lies in accurately locating anomaly frames in long videos. Firstly, a panoramic spatiotemporal reasoning module is designed, which performs global context modeling through global attention pooling and feature fusion techniques, and employs a triple attention mechanism to capture deep spatial features, thereby obtaining accurate spatiotemporal feature representations. Secondly, a feature enhancement mechanism is constructed to highlight anomaly features, and two memory units are introduced to store normal and anomaly prototypes respectively, optimizing the processing of ambiguous features. Finally, a contrastive measurement loss is proposed to enhance the similarity of feature magnitudes among similar videos and the separability of feature magnitudes between normal and abnormal videos, learning an adaptive feature magnitude distribution across scenes to solve the inconsistency of feature magnitudes between different scenes. Experimental results demonstrate that this method can accurately locate the position of anomaly frames in various scenarios and achieves higher accuracy on the UCF-Crime and XD-Violence datasets.

ABSTRACT Video anomaly detection (VAD), a critical task in intelligent surveillance systems, faces two key challenges: Dynamic behavioral characterization under complex scenarios and robust spatiotemporal context modeling. Existing methods face limitations, such as inadequate cross‐scale feature fusion, weak channel‐wise dependency modeling, and sensitivity to background noise. To address these issues, we propose a novel multi‐scale spatiotemporal feature augmentation framework. Our approach introduces three core innovations: Hierarchical feature pyramid architecture for multi‐granularity representation learning, capturing both local motion patterns and global scene semantics; A channel‐adaptive attention mechanism that dynamically models long‐range spatiotemporal dependencies; A spatiotemporal Gaussian difference module to enhance anomaly response through frequency‐domain feature reconstruction, effectively suppressing noise interference. Extensive experiments on UCSD Ped1/2, CUHK Avenue, and ShanghaiTech benchmarks demonstrate that our method achieves state‐of‐the‐art performance, outperforming existing approaches in both accuracy and robustness.

Multi-label text classification (MLTC) holds significant importance in the field of data management and information retrieval, where the distribution of label samples often exhibits a long-tail pattern. This poses a formidable challenge for effectively summarizing test data, particularly for classes with limited sample representation. Existing approaches aimed at addressing this challenge often involve altering the original data distribution, resulting in reduced generalization performance on real-world data and inadequate mitigation of the long-tail problem in samples. To tackle this challenge, we propose a novel method for multi-label text classification called MLTC-LD (an enhanced Multi-Label Text Classification model by incorporating Label Dependency). The primary objective of MLTC-LD is to enhance the performance of classes with scarce data by leveraging knowledge acquired from classes with abundant data. MLTC-LD leverages Graph Attention Networks (GAT) to incorporate prior information associated with labels. It calculates a relationship matrix between label samples and integrates this information into the classifier. We conducted a comprehensive evaluation of MLTC-LD on three datasets and compared its performance against eight baseline models. The experimental results validate the superiority of MLTC-LD in effectively mitigating the challenges posed by long-tail distributions.

Video anomaly detection has always been a challenging task in computer vision due to data imbalance and susceptibility to scene variations such as lighting and occlusions. In response to this challenge, this paper proposes an unsupervised video anomaly detection method based on an attention-enhanced memory network. The method utilizes a dual-stream network structure of autoencoders, enhancing the model’s learning ability for important features in appearance and motion by introducing coordinate attention mechanisms and variance attention mechanisms, emphasizing significant characteristics of static objects and rapidly moving regions. By adding memory modules to both the appearance and motion branches, the network structure’s memory information is reinforced, enabling it to capture long-term spatiotemporal dependencies in videos and thereby improving the accuracy of anomaly detection. Furthermore, by optimizing the network structure’s activation functions to handle negative inputs, it enhances its nonlinear modeling capabilities, enabling better adaptation to complex environments, including variations in lighting and occlusions, further improving the effectiveness of anomaly detection. The paper conducts comparative experiments and ablation studies using three public available datasets and various models. The results demonstrate that compared to baseline models, the AUC performance is improved by 3.9%, 4.7%, and 1.7% on UCSD Ped2, CHUK Avenue, and ShanghaiTech datasets, respectively. When compared with the other models, the average AUC performance is improved by 4.3%, 5.4%, and 6.2%, with an average improvement of 8.75% in the ERR metric, validating the effectiveness and adaptability of the proposed method. The code can be obtained at the following URL: https://github.com/AcademicWhite/AEMNet .

Abstract Pedestrian detection is crucial for various applications, including intelligent transportation and video surveillance systems. Although recent research has advanced pedestrian detection models like the YOLO series, they still face limitations in handling diverse pedestrian scales, leading to performance challenges. To address these issues, we propose HF-YOLO, an advanced pedestrian detection model. HF-YOLO tackles the complexities of pedestrian detection in complex scenes by addressing scale variations and occlusions among pedestrians. In the feature fusion stage, our algorithm leverages both shallow localization information and deep semantic information. This involves fusing P2 layer features and adding a high-resolution detection layer, significantly improving the detection of small-scale pedestrians and occluded instances. To enhance feature representation, HF-YOLO incorporates the HardSwish activation function, introducing more non-linear factors and strengthening the model’s ability to represent complex and discriminative features. Additionally, to address regression imbalance, a balance factor is introduced to the CIoU loss function. This modification effectively resolves the imbalance problem and enhances pedestrian localization accuracy. Experimental results demonstrate the effectiveness of our proposed algorithm. HF-YOLO achieves notable improvements, including a 3.52% increase in average precision, a 1.35% boost in accuracy, and a 4.83% enhancement in recall. Moreover, the algorithm maintains real-time performance with a detection time of 8.5ms, meeting the stringent requirements of real-time applications.

Abstract We propose a simple image high-perception super-resolution reconstruction method based on multi-layer feature fusion model with adaptive compression and parameter tuning. The aim is to further balance the high and low-frequency information of an image, enrich the detailed texture to improve perceptual quality, and improve the adaptive optimization and generalization of the model in the process of super-resolution reconstruction. First, an effective multi-layer fusion super-resolution (MFSR) basic model is constructed by the design of edge enhancement, refine layering, enhanced super-resolution generative adversarial network and other sub-models, and effective multi-layer fusion. This further enriches the image representation of features of different scales and depths and improves the feature representation of high and low-frequency information in a balanced way. Next, a total loss function of the generator is constructed with adaptive parameter tuning performance. The overall adaptability of the model is improved through adaptive weight distribution and fusion of content loss, perceptual loss, and adversarial loss, and improving the error while reducing the edge enhancement model. Finally, a fitness function with the evaluation perceptual function as the optimization strategy is constructed, and the model compression and adaptive tuning of MFSR are carried out based on the multi-mechanism fusion strategy. Consequently, the construction of the adaptive MFSR model is realized. Adaptive MFSR can maintain high peak signal to noise ratio and structural similarity on the test sets Set5, Set14, and BSD100, and achieve high-quality reconstructed images with low learned perceptual image patch similarity and perceptual index, while having good generalization capabilities.

Open Relation Extraction (OpenRE) aims at clustering relation instances to extract relation types. By learning relation patterns between named entities, it clusters semantically equivalent patterns into a unified relation cluster. Existing clustering-based OpenRE methods only consider the information of the instance itself, ignoring knowledge of any relations between instances. Therefore, a Cross-Attention Fusion based Graph Convolution Auto-Encoder (CAGCE) method for Open Relation Extraction is proposed. The Auto-Encoder learns the semantic information of the sentence instance itself, and the Graph Convolution Network learns the relational similarity information between sentences. Then, the two heterogeneous representations are crossed and fused layer-by-layer through a cross-attention fusion mechanism. Finally, the fused features are used for clustering to form the relation types. A comparison with baseline models using the FewRel and NYT-FB datasets shows the effectiveness and superiority of the proposed method.

The Corona Virus Disease 2019 (COVID-19) epidemic is a sudden public health crisis, known as an "International Emergency of Public Health Event". This study uses the bottom-up characteristics of multi-agents to construct multi-agent simulation models for COVID-19 prevention and control. The development trend of the epidemic situation under the condition that the government adopts different prevention and control measures is studied, and on this basis, the influence of temperature on the spread of the virus is discussed. The simulation results show that the multi-agent modeling method can effectively capture the emergence of complex systems. The evaluation of the effects of single measures and multiple interventions will help determine key prevention and control strategies and provide important experience and scientific basis for future epidemic prevention and control.