In this study, the influence of polyol type and molecular weight on the performance of Waterborne polyurethanes (WPUs) was systematically investigated. WPUs systems modified with long fatty side chains were successfully prepared by innovatively introducing terminal hydroxyl silicone oils to construct a heterostructure and using glycerol monostearate as a functional chain extender. Characterization via FTIR, TGA, SEM, and contact angle testing confirmed that when polytetrahydrofuran ether diol (PTMG) with a molecular weight of 2000 g/mol was selected as the soft segment, the resulting material exhibited significant synergistic effects. Its water contact angle reached 108.7°, and the 72 h water absorption rate was only 10.78% markedly lower than that of WPUs derived from other polyols. Additionally, when the film was heated to 250℃, only 5 wt% decomposed. This material combines excellent hydrophobicity, low water absorption, good thermal stability, and mechanical properties. The significantly enhanced water resistance effectively addresses the application limitations of traditional waterborne coatings in humid environments. It provides a theoretical foundation and process reference for developing high-performance eco-friendly protective coatings, particularly suitable for ship anticorrosion, outdoor building materials, and other harsh weathering applications.

Accurate protein-ligand binding affinity prediction is central to computational drug discovery, yet modern docking engines frequently disagree without indicating which prediction to trust. Consensus scoring and ensemble methods improve mean accuracy but treat all predictions identically without interpretable confidence measures or uncertainty decomposition, ignoring the chemical context of each protein-ligand pair. To address this limitation, we introduce RELIABLE-BA (RELIABiLity-aware Evidential fusion for Binding Affinity), an evidential framework for multi-engine binding affinity prediction. Our model comprises three steps: (1) modeling each engine as an evidential expert via Normal-Inverse-Gamma distributions, (2) scaling epistemic uncertainty through learned reliability from molecular context while preserving each expert's predictive mean, and (3) fusing experts through closed-form aggregation that captures both individual uncertainty and inter-engine disagreement. Experiments on the PDBBind and BDB2020+ benchmarks demonstrate competitive point prediction with substantially improved uncertainty calibration, and additional validation on the SARS-CoV-2 Mpro dataset and 5HT2A receptor demonstrates applicability to clinically relevant drug targets. Crucially, these uncertainty estimates enable reliable filtering of protein-ligand pairs, reducing prediction error by up to 25

Compared to conventional image content analysis tasks, visual emotion analysis is perceived as a complex, abstract, and potentially culturally dependent endeavor. The accuracy of automatic image-based emotion recognition remains a challenge, and the most significant obstacles are the affective gap and scarcity of data, particularly labeled data. To address these challenge, this paper proposes a saliency-guided masked image modeling approach. Specifically, the proposed framework employs multi-modal large model to generate more emotional images, thereby reducing the impact of data scarcity on model performance. Subsequently, neuroimaging and behavioral studies have demonstrated that human visual attention is attracted by the emotional relevance of a stimulus. In light of this, our model employs a saliency-guided masking strategy to identify emotion-related regions for masking sampling to fit the affective gap. In contrast to the conventional approach of using the original pixel values for the reconstruction target, our model eliminates high-frequency components from the pixels, thus enhancing the generalizability of the model. The use of this unsupervised representation learning approach enables the model to exhibit outstanding recognition performance in downstream emotion recognition tasks on three standard emotion datasets. Furthermore, ablation experiments, robustness test, and visualization experiments corroborate the effectiveness of the proposed method.

Time series analysis underpins scientific advances. While specialized models have advanced various time series tasks, scientific domains face unique challenges: limited samples with complex physical dynamics, missing observations, multi-resolution sampling, and requirements for physical consistency. With the increasing demands on generative modeling capabilities, we introduce PINFDiT, a diffusion transformer-based model with physics injection during inference. Our approach combines a transformer backbone for capturing temporal dependencies with a comprehensive masking strategy that addresses imperfect data. The diffusion framework enables high-quality sample generation with inherent generative capability. In addition, our model-free physics-guided correction steers generated samples toward physically consistent solutions using calibrated Langevin dynamics, which balances distribution fidelity and physical law adherence without architectural modifications or retraining. Our evaluation demonstrates PINFDiT's effectiveness across multivariate forecasting with imperfect data, physics knowledge incorporation in data-limited scenarios, zero-shot and fine-tuning performance across diverse domains, establishing it as a proto-foundation model that bridges the gap between general-purpose and domain-specific models.

Multi-omics profiling characterizes cancer biology and supports biomarker discovery for prognosis and therapy selection. Although numerous computational multi-omics biomarker identification methods have been proposed, their ability to identify clinically relevant biomarkers has not been systematically evaluated, leaving it unclear whether the resulting biomarker nominations are reliable for downstream validation. Here, we systematically benchmark 20 representative statistical, machine learning and deep learning methods using curated gold-standard prognostic and therapeutic biomarkers across five real-world datasets. We evaluate performance in terms of both biomarker identification accuracy and stability. Overall, DeePathNet and DeepKEGG achieve the best performance. Across methods, effective biomarker recovery is associated with the integration of biological knowledge, global feature interactions, multivariate feature attribution, and effective regularization. Analysis of omics type contributions reveals method- and modality-specific biases, highlighting the importance of broader omics integration. We further evaluate methods on simulated datasets to probe sensitivity with controlled signal and noise. By aggregating results from top-performing methods, we construct consensus biomarker panels that nominate candidates for potential investigations. Finally, we provide user-friendly interfaces to allow researchers to benchmark new methods against the 20 baselines or apply selected methods for biomarker identification on custom multi-omics datasets. Our benchmark is publicly available at https://github.com/athanzli/CancerMOBI-Bench.

Graph topology is a fundamental determinant of memory leakage in multi-agent LLM systems, yet its effects remain poorly quantified. We introduce MAMA (Multi-Agent Memory Attack), a framework that measures how network structure shapes leakage. MAMA operates on synthetic documents containing labeled Personally Identifiable Information (PII) entities, from which we generate sanitized task instructions. We execute a two-phase protocol: Engram (seeding private information into a target agent's memory) and Resonance (multi-round interaction where an attacker attempts extraction). Over 10 rounds, we measure leakage as exact-match recovery of ground-truth PII from attacker outputs. We evaluate six canonical topologies (complete, ring, chain, tree, star, star-ring) across n∈{4,5,6}, attacker-target placements, and base models. Results are consistent: denser connectivity, shorter attacker-target distance, and higher target centrality increase leakage; most leakage occurs in early rounds and then plateaus; model choice shifts absolute rates but preserves topology ordering; spatiotemporal/location attributes leak more readily than identity credentials or regulated identifiers. We distill practical guidance for system design: favor sparse or hierarchical connectivity, maximize attacker-target separation, and restrict hub/shortcut pathways via topology-aware access control.

Worldwide incidence and prevalence of ulcerative colitis (UC) has been rising in recent years, which can occur at any age, with a high frequency seen in young children and people aged 40 to 50. The aryl hydrocarbon receptor (AhR) activation axis is well known for its important role in the regulation of intestinal inflammation, intestinal homeostasis, intestinal immune system and improvement of colitis outcomes. This study investigated the therapeutic efficacy of the thiophene-based styrene derivative (TBSD), a novel AhR agonist against UC in vitro and in vivo. TBSD decreased FITC-dextran hyperpermeability, upregulated the tight junction (TJ)-related protein expression levels and regulated the inflammatory mediators including tumor necrosis factor-alpha (TNF-α), interleukin (IL)-6, IL-22 and cyclooxygenase 2 (COX-2) in the Caco-2/RAW264.7 co-culture system and in DSS-induced UC-like mice. Overall, TBSD may be considered as a promising therapeutic agent to improve UC severity through mitigating inflammation, maintaining intestinal mucosal homeostasis and enhancing the intestinal barrier integrity.

Large Language Models (LLMs) have become central to modern AI, with interpretability serving as a critical means of investigating the opaque and highly nonlinear mechanisms encoded within billions of parameters and ensuring trustworthy deployment. However, descriptive interpretability approaches for LLMs remain largely post-hoc, illuminating model behavior without providing the actionable leverage needed to influence or adapt model behavior, thereby limiting their practical utility. Recent work has therefore reframed interpretability as an actionable paradigm, shifting the focus from explanation alone toward methods that connect internal mechanisms to model refinement. This survey reviews LLM interpretability through the lens of actionability, presenting a taxonomy of attributional, concept-based, and mechanistic approaches, along with emerging methods tailored to vision–language models (VLMs). We further examine how interpretability supports downstream objectives such as hallucination mitigation, model editing, fairness, and safety. By positioning interpretability as a pathway to better-guided LLM design and practice, this survey outlines key challenges and future directions toward trustworthy and controllable foundation models.

Spatiotemporal models have drawn significant interest recently due to their widespread applicability across many domains. These models are often made more practically useful by incorporating beneficial inductive biases, such as laws or symmetries from domain-relevant physics equations. This "physics-awareness" provides an interpretable means of grounding otherwise purely data-driven models, improving robustness and boosting performance in settings with limited data. In this work, we view physical dynamics as domain knowledge that captures fundamental causal relationships across space and time, and can be effectively leveraged by our proposed physics-aware spatiotemporal causal graph network (P-STCGN). We firstly describe a means of deriving causal relationships from spatiotemporal data, serving as physics-aware labels to learn a causal structure via a dedicated neural module. We then formulate a forecasting module that can operate under this causal structure, producing predictions that are guided by physics-aware cause-effect relationships among modeled variables. Extensive experimentation demonstrates that our method is robust to noisy and limited data, outperforming existing models across a variety of challenging synthetic tasks and benchmark datasets. We further evaluate our method on real-world graph signals and observe superior forecasting performance, achieved by effectively utilizing causal signals from prior physics knowledge.

Large language models (LLMs) have been serving as effective backbones for retrieval systems, including Retrieval-Augmentation-Generation (RAG), Dense Information Retriever (IR), and Agent Memory Retrieval. Recent studies have demonstrated that such LLM-based Retrieval (LLMR) is vulnerable to adversarial attacks, which manipulates documents by token-level injections and enables adversaries to either boost or diminish these documents in retrieval tasks. However, existing attack studies mainly (1) presume a known query is given to the attacker, and (2) highly rely on access to the victim model's parameters or interactions, which are hardly accessible in real-world scenarios, leading to limited validity. To further explore the secure risks of LLMR, we propose a practical black-box attack method that generates transferable injection tokens based on zero-shot surrogate LLMs without need of victim queries or victim models knowledge. The effectiveness of our attack raises such a robustness issue that similar effects may arise from benign or unintended document edits in the real world. To achieve our attack, we first establish a theoretical framework of LLMR and empirically verify it. Under the framework, we simulate the transferable attack as a min-max problem, and propose an adversarial learning mechanism that finds optimal adversarial tokens with learnable query samples. Our attack is validated to be effective on benchmark datasets across popular LLM retrievers.

In this study, waterborne polyurethane (WPU) microcapsules encapsulating bendiocarb were successfully prepared by interfacial polymerization for sustainable mosquito control. A high encapsulation efficiency of 94

Autocorrelation is a defining characteristic of time-series data, where each observation is statistically dependent on its predecessors. In the context of deep time-series forecasting, autocorrelation arises in both the input history and the label sequences, presenting two central research challenges: (1) designing neural architectures that model autocorrelation in history sequences, and (2) devising learning objectives that model autocorrelation in label sequences. Recent studies have made strides in tackling these challenges, but a systematic survey examining both aspects remains lacking. To bridge this gap, this paper provides a comprehensive review of deep time-series forecasting from the perspective of autocorrelation modeling. In contrast to existing surveys, this work makes two distinctive contributions. First, it proposes a novel taxonomy that encompasses recent literature on both model architectures and learning objectives – whereas prior surveys neglect or inadequately discuss the latter aspect. Second, it offers a thorough analysis of the motivations, insights, and progression of the surveyed literature from a unified, autocorrelation-centric perspective, providing a holistic overview of the evolution of deep time-series forecasting. The full list of papers and resources is available at https://github.com/Master-PLC/Awesome-TSF-Papers.

With the increasing concerns about privacy and data regulations, federated learning (FL) has been emerging as a solution to train machine learning models collaboratively with non-exchangeable data from multiple clients. As a result of data locality, data is usually not identically or independently (non-IID) distributed across clients, and the non-IID property has long been the key challenge in FL. Furthermore, in real-world cross-silo scenarios, it is ubiquitous that clients are organizations owning private data from multiple domains internally, which exacerbates the non-IID issue. For example, in healthcare applications, each client (hospital) gathers data from patients with heterogeneous demographics. While previous works have made efforts to address the non-IID challenge across clients by assuming various relations among client-level data distributions and enabling personalized models at the client level, they ignore the internal data heterogeneity within each client or require explicit data domain indicators, which are hardly accessible in real-world data. Here, we propose Sample-Level Prototypical Federated Learning (SL-PFL) to bridge the gap. SL-PFL incorporates prototypical learning under the FL framework and provides a fine-grained personalized model for each data sample instead of learning one uniform model for all samples of each client. Meanwhile, it can be trained using data without ground-truth domain indicators. Experimental results demonstrate that our proposed method with sample-level personalized models outperforms existing FL methods with a global model or client-level personalized models on various real-world regression and classification tasks from weather, computer vision, and healthcare applications.

The accumulating body of research indicates that Motor Imagery (MI)-BCIs have the potential to enhance the quality of life for individuals with disabilities and to advance our understanding of brain function and rehabilitation strategies. Among these diseases, stroke is the leading cause of long-term motor disability across the globe, thereby underscoring the need for innovative rehabilitation strategies, such as MI-BCI technologies. In contrast with these expectations, the majority of existing research is built upon data obtained from healthy subjects. The construction of effective classification models for Motor Imagery tasks in patients with brain diseases, particularly stroke, remains a significant challenge. The lateralization of the left and right hemispheres is more pronounced in patients who have suffered a stroke than in healthy individuals. Moreover, the specific locations of lesions and the regions of influence result in significant variations in the electroencephalogram (EEG) data of patients with different hemiplegic sides. This paper explores the potential of generative models in addressing the issue of domain differences arising from different hemiplegic sides EEG data. Furthermore, this paper circumvents the potential adverse effects of rigorous optimization of low-quality samples on model performance through the utilization of label softening algorithm. Two MI-EEG datasets of stroke patients performing Motor Imagery tasks are used to validate our method. In comparison to both classical machine learning methods and those state-of-the-art models for MI classification, the classification model in this paper achieves a noticeable performance improvement in different data partitioning strategies, including subject-dependent and subject-independent scenarios. Each sub-module, and each designed loss function, contributes to the final performance growth. In addition, this paper also investigates the potential of the proposed framework for predicting a patient's level of functional recovery. Our findings indicate that the addition of a prediction layer to the proposed model enables the accurate prediction of functional recovery level in stroke patients.

Retrieval-Augmented Generation (RAG) systems deployed over proprietary knowledge bases face growing threats from reconstruction attacks that aggregate model responses to replicate knowledge bases. Such attacks exploit both intra-class and inter-class paths—progressively extracting fine-grained knowledge within topics and diffusing it across semantically related ones, thereby enabling comprehensive extraction of the original knowledge base. However, existing defenses target only one path, leaving the other unprotected. We conduct a systematic exploration to assess the impact of protecting each path independently and find that joint protection is essential for effective defense. Based on this, we propose RAGFort, a structure-aware dual-module defense combining contrastive reindexing for inter-class isolation and constrained cascade generation for intra-class protection. Experiments across security, performance, and robustness confirm that RAGFort significantly reduces reconstruction success while preserving answer quality, offering the first comprehensive defense against knowledge base extraction attacks.

In response to increasing concerns over particulate matter (PM) pollution and the environmental impact of disposable air filters, this study reports a nanofibrous filter with a vertically oriented multiscale interwoven network structure. Such uniquely structured filters are fabricated via a facile one-step electrospinning process, avoiding additional processing steps. The resulting hierarchical network consists of microfibers (≈ 1.07 μm), nanofibers (≈ 154 nm), and ultrafine nanofibers (≈ 34 nm), forming interconnected pores with an average size of 2.31 μm and a low areal density of 2.62 g m⁻². This multiscale architecture provides a unique combination of mechanical robustness (0.076 MPa, 6.9%) and optimized airflow pathways, resulting in highly efficient PM₀.₃ removal (99.28 – 99.52%) with an ultralow pressure drop (48 – 65 Pa). In addition, the filters exhibit dual-surface wettability, integrating hydrophobicity and wettability without compromising filtration performance. These results demonstrate that structural engineering of vertically oriented multiscale networks enables simultaneous enhancement of filtration efficiency and air permeability, effectively overcoming the conventional trade-off between these properties, and provides a design principle for next-generation high-performance air filtration materia

With economic development, the treatment of complex wastewater systems imposes increasingly stringent requirements on the performance of separation membrane materials. Perfluoroalkoxy (PFA), also known as fusible polytetrafluoroethylene, has excellent resistance to organic solvents, acid and alkali, and high temperature resistance, along with melt-processing capabilities, making it the ideal material for treating complex wastewater. In this study, PFA was used as the membrane-forming polymer to fabricate membranes via the method of melt hot-pressing, with deliberate design of the pore structure. The composite pore-forming agent consisting of watersoluble inorganic particles (NaCl) and soluble organic polymer (PVDF) was employed to prepare the separation membranes. The influence of the composite pore-forming agent on the membrane structure and performance was investigated, and the formation mechanism of the multi-level pore structure was elucidated. The results indicate that using the organic polymer PVDF alone as the pore-forming agent tended to produce streamlined large pores, whereas the inorganic particles NaCl alone as the pore-forming agent led to the formation of small pores with poor connectivity. In contrast, the composite pore-forming agent composed of organic polymer and inorganic particles was more conducive to forming an ideal pore structure, which could avoid the disadvantages of using single-component pore-forming agents to form structure that have excessively large pores or inadequate permeability. Furthermore, the content of the composite pore-forming agent had a significant regulatory effect on the pore structure of the separation membranes. When the composite pore-forming agent content was 20 %, the PFA membrane exhibited a pore size of 0.038 mu m, while increasing the content to 50 % led to a pore size of 10.814 mu m. Meanwhile, the prepared PFA membranes showed good chemical stability and thermal stability, and demonstrated effective treatment capabilities for acids/alkalis/organic solvent-containing liquids and hightemperature organic liquids. Therefore, the prepared PFA membrane had remarkable application potential for efficient separation and purification of complex chemical liquids.

Large language models (LLMs) are attractive for context-aware time series forecasting because they can integrate heterogeneous textual signals, yet their discrete, language-oriented tokenization and embedding interfaces are misaligned with continuous numerical values, often harming numerical ordering and forecasting reliability. We propose TempoWave, a plug-and-play temporal wavelet digit interface that maps each scalar observation into digit-wise embeddings constructed from multi-wavelet, multi-scale coefficients. By directly overriding standard token representations, TempoWave seamlessly exposes both fine-grained local fluctuations and macro global structures in a transformer-compatible form, ensuring that precise numerical formatting, distinct digit identity, and robustness to common normalization operations are maintained throughout the LLM pipeline. Experiments across five context-enriched forecasting benchmarks demonstrate that TempoWave consistently improves LLM-based forecasters over standard numeric tokenization and alternative embedding interfaces, achieving a new state-of-the-art. These results highlight the numeric interface as a key bottleneck and suggest that principled multi-resolution embeddings can better couple LLMs' contextual reasoning with precise forecasting. Our code is available at https://github.com/DC-research/TempoWAVE.

Understanding human mobility is critical for a wide range of urban applications, including traffic management, epidemic control, and urban planning. However, due to privacy concerns, the availability of large-scale public trajectory data remains limited, posing challenges for downstream mobility analysis. Existing methods for synthetic trajectory generation primarily focus on matching global distribution similarity, while often overlooking mobility patterns across different spatial and temporal resolutions that are essential for practical utility. To address these challenges, we propose a novel multi-resolution diffusion framework, MR-Traj, for large-scale trajectory generation. MR-Traj explicitly models trajectories as compositions of coarse-grained milestones and fine-grained segments, enabling the capture of complex spatial-temporal dependencies at multiple resolutions. Experimental results demonstrate that MR-Traj achieves comparable performance to state-of-the-art methods in terms of global distribution similarity, while consistently outperforming them in modeling fine-resolution mobility patterns and supporting downstream urban mobility tasks. In addition, by introducing stochasticity at multiple resolution levels, MR-Traj generates more diverse trajectories, which empirically reduces trajectory linkage risk under a seed-guided data release setting. Our code is available at https://github.com/Ray0202/MR-Traj.