Nat Commun. 2025 Apr 5;16(1):3251. doi: 10.1038/s41467-025-58568-x.

ABSTRACT

Inhaled therapeutics have high potential for the treatment of chronic respiratory diseases of high unmet medical need, such as idiopathic pulmonary fibrosis (IPF). Preclinical and early clinical evidence show that cellular communication network factor 2 (CCN2), previously called connective tissue growth factor (CTGF), is a promising target for the treatment of IPF. In recent phase 3 clinical trials, however, systemic CCN2 inhibition failed to demonstrate a clinically meaningful benefit. Here, we present the preclinical profile of the inhaled anti-CCN2 Anticalin® protein PRS-220. Our study demonstrates that efficient pulmonary delivery directly translates into superior efficacy in relevant models of pulmonary fibrosis when compared to systemic CCN2 inhibition. Moreover, we present a holistic approach for the preclinical characterization of inhaled PRS-220 from state-of-the art in vitro and in vivo models to novel human ex vivo and in silico models, highlighting the advantage of inhaled drug delivery for treatment of respiratory disease.

PMID:40185752 | DOI:10.1038/s41467-025-58568-x

Radiother Oncol. 2025 Apr 2:110837. doi: 10.1016/j.radonc.2025.110837. Online ahead of print.

ABSTRACT

The incidence of radiation pneumonitis (RP) has decreased significantly compared to historical series, mainly due to improved radiotherapy techniques and patient selection. Nevertheless, some patients still develop RP. This guideline provides user-friendly flowcharts to address common clinical practice questions regarding RP. We summarize the current state of the art regarding the mechanisms, risk factors, diagnosis and treatment of RP. Dosimetric constraints to minimize the incidence of RP, as well as risk factors for developing RP, such as idiopathic pulmonary fibrosis (IPF) were identified. The combination of radiotherapy and medication as a risk factor for the development of RP was reviewed. RP remains a diagnosis of exclusion, but an algorithm for reaching the diagnosis has been proposed. Finally, practical approaches to the treatment of RP are outlined.

PMID:40185160 | DOI:10.1016/j.radonc.2025.110837

Colloids Surf B Biointerfaces. 2025 Mar 26;252:114652. doi: 10.1016/j.colsurfb.2025.114652. Online ahead of print.

ABSTRACT

Idiopathic pulmonary fibrosis (IPF) is a fatal, progressive lung disease characterized by extensive scarring and thickening of lung tissue that leads to respiratory failure. Early and accurate diagnosis is crucial for monitoring disease progression and assessing therapeutic efficacy. While imaging modalities such as radiological X-rays and high-resolution computed tomography (HRCT) are commonly employed, magnetic resonance imaging (MRI) offers significant advantages, including superior soft tissue contrast and the absence of ionizing radiation. However, in lung MRIs are hindered by short transverse relaxation times, low proton density, and motion artifacts which is addressed herein by developing theranostic platform combining MRI imaging with targeted drug delivery using melanin nanoparticles conjugated with nintedanib (MNP-NIN). Chelation with ferric ions (MNP-NIN-Fe³⁺) enhanced MRI visibility enabling non-invasive imaging and drug tracking. MNP-NIN and MNP-NIN-Fe³ ⁺ nanoparticles were built with mean diameters of 189 ± 44 nm and 182 ± 35 nm, respectively and demonstrating successful NIN conjugation. Controlled NIN release followed zero-order kinetics over 36 days. MNP conjugation reduced cytotoxicity in BEAS-2B and A549 cells improving the drug's safety. MRI experiments conducted with a 7.0 T animal scanner demonstrated enhanced imaging contrast with MNP-NIN-Fe solutions compared to saline underscoring their potential for localized visualization and tracking within lung tissues. By integrating MRI diagnostics and targeted drug delivery, the MNP-NIN-Fe³ ⁺ system offers a promising solution to overcome current challenges in IPF management. This theranostic platform addresses the limitations of conventional imaging techniques while providing an innovative strategy for reducing drug-related systemic side effects and improving therapeutic efficacy.

PMID:40184721 | DOI:10.1016/j.colsurfb.2025.114652

EMBO J. 2025 Apr 4. doi: 10.1038/s44318-025-00424-1. Online ahead of print.

ABSTRACT

The mammary epithelium derives from multipotent mammary stem cells (MaSCs) that engage into differentiation during embryonic development. However, adult MaSCs maintain the ability to reactivate multipotency in non-physiological contexts. We previously reported that Notch1 activation in committed basal cells triggers a basal-to-luminal cell fate switch in the mouse mammary gland. Here, we report conservation of this mechanism and found that in addition to the mammary gland, constitutive Notch1 signaling induces a basal-to-luminal cell fate switch in adult cells of the lacrimal gland, the salivary gland, and the prostate. Since the lineage transition is progressive in time, we performed single-cell transcriptomic analysis on index-sorted mammary cells at different stages of lineage conversion, generating a temporal map of changes in cell identity. Combining single-cell analyses with organoid assays, we demonstrate that cell proliferation is indispensable for this lineage conversion. We also reveal the individual transcriptional landscapes underlying the cellular plasticity switching of committed mammary cells in vivo with spatial and temporal resolution. Given the roles of Notch signaling in cancer, these results may help to better understand the mechanisms that drive cellular transformation.

PMID:40186028 | DOI:10.1038/s44318-025-00424-1

EMBO Rep. 2025 Apr 4. doi: 10.1038/s44319-025-00437-1. Online ahead of print.

ABSTRACT

Genes regulating the finger-like cellular protrusions-filopodia have long been implicated in cancer metastasis. However, depleting the flat lamellipodia but retaining filopodia drastically hampers cell migration on spread surface, obscuring the role of filopodia in cell motility. It has been noticed recently that cells under confinement may employ distinct migratory machineries. However, the regulating factors have mainly been focused on cell blebbing, nuclear deformation and cell rear contractility, without much emphasis on cell protrusions and even less on filopodia. Here, by micropore-based screening, we identified espin as an active regulator for confined migration and that its overexpression was associated with metastasis. In comparison to fascin, espin showed stronger actin bundling in vitro and induced shorter and thicker filopodia in cells. Combining the imaging-compatible microchannels and DNA-based tension probes, we uncovered that espin overexpression induced excessive filopodia at the leading edge and along the sides, exerting force for confined migration. Our results demonstrate an important role for filopodia and the regulating protein-espin in confined cell migration and shed new light on cytoskeletal mechanisms underlying metastasis.

PMID:40185977 | DOI:10.1038/s44319-025-00437-1

Sci Rep. 2025 Apr 4;15(1):11544. doi: 10.1038/s41598-025-90592-1.

ABSTRACT

Understanding the multicellular organization of stem cells is vital for determining the mechanisms that coordinate cell fate decision-making during differentiation; these mechanisms range from neighbor-to-neighbor communication to tissue-level biochemical gradients. Current methods for quantifying multicellular patterning tend to capture the spatial properties of cell colonies at a fixed scale and typically rely on human annotation. We present a computational pipeline that utilizes topological data analysis to generate quantitative, multiscale descriptors which capture the shape of data extracted from 2D multichannel microscopy images. By applying our pipeline to certain stem cell colonies, we detected subtle differences in patterning that reflect distinct spatial organization associated with loss of pluripotency. These results yield insight into putative directed cellular organization and morphogen-mediated, neighbor-to-neighbor signaling. Because of its broad applicability to immunofluorescence microscopy images, our pipeline is well-positioned to serve as a general-purpose tool for the quantitative study of multicellular pattern formation.

PMID:40185811 | DOI:10.1038/s41598-025-90592-1

Nat Commun. 2025 Apr 4;16(1):3106. doi: 10.1038/s41467-025-58272-w.

ABSTRACT

"To act or not to act" is a fundamental decision made in daily life. However, it is unknown how the relevant signals are transmitted to the secondary motor cortex (M2), which is the cortical origin of motor initiation. Here, we found that in a decision-making task in male mice, inputs from the thalamus to M2 positively regulated the action while inputs from the lateral part of the orbitofrontal cortex (LO) negatively regulated it. The motor thalamus that received the basal ganglia outputs transmitted action value-related signals to M2 regardless of whether the animal acted or not. By contrast, a large subpopulation of LO inputs showed decreased activity before and during the action, regardless of the action value. These results suggest that M2 integrates the positive signal of the action value from the motor thalamus with the negative action-biased signal from the LO to finally determine whether to act or not.

PMID:40185746 | DOI:10.1038/s41467-025-58272-w

Comput Biol Med. 2025 Apr 3;190:110064. doi: 10.1016/j.compbiomed.2025.110064. Online ahead of print.

ABSTRACT

The rapidly advancing field of artificial intelligence (AI) has transformed numerous scientific domains, including biology, where a vast and complex volume of data is available for analysis. This paper provides a comprehensive overview of the current state of AI-driven methodologies in genomics, proteomics, and systems biology. We discuss how machine learning algorithms, particularly deep learning models, have enhanced the accuracy and efficiency of embedding sequences, motif discovery, and the prediction of gene expression and protein structure. Additionally, we explore the integration of AI in the embedding and analysis of biological networks, including protein-protein interaction networks and multi-layered networks. By leveraging large-scale biological data, AI techniques have enabled unprecedented insights into complex biological processes and disease mechanisms. This work underlines the potential of applying AI to complex biological data, highlighting current applications and suggesting directions for future research to further explore AI in this rapidly evolving field.

PMID:40184941 | DOI:10.1016/j.compbiomed.2025.110064

Int Immunopharmacol. 2025 Apr 3;154:114515. doi: 10.1016/j.intimp.2025.114515. Online ahead of print.

ABSTRACT

BACKGROUND: Sulfur mustard (SM) is a highly lethal chemical warfare agent that induces severe health complications in exposed individuals. Gaining insights into the metabolic changes caused by SM exposure is essential for understanding its underlying mechanisms and developing effective diagnostic and therapeutic interventions.

METHODS: In this investigation, we utilized proton nuclear magnetic resonance (H-NMR) spectroscopy to conduct metabolomic analysis in patients diagnosed with mustard lung disease (MLD) using a non-targeted approach. Metabolite measurements were conducted on plasma and urine samples collected from a total of 54 individuals, including 20 individuals with mild MLD, 20 individuals with moderate MLD, and 14 healthy individuals. Multivariate and univariate analyses were applied to identify metabolites that distinguish between the different groups, and enrichment analysis was performed to unveil the underlying biochemical pathways involved.

RESULTS: The obtained metabolic profile had the potential to differentiate moderate from healthy plasma, but not from mild patients using multivariate analysis. Sixteen metabolites from plasma were considered significantly different between the moderate and control groups (VIP > 1 and p < 0.05) that these metabolites involved in fatty acid and amino acid metabolism. Utilizing all 16 metabolites as a combined panel, we were able to distinguish between the moderate and control groups, achieving an area under the curve (AUC) of 0.854. Moreover, 6 and 8 urinary metabolites were detected between mild vs. control and moderate vs. control groups, respectively. Fourteen metabolites exhibited significant fold changes (FC) (FC < 0.66 or FC > 1.5; p < 0.05). These metabolites are involved in amino acid and nicotinate metabolism.

CONCLUSION: Our study provides novel insights into the metabolic changes associated with MLD and highlights potential pathways involved in the disease progression. These findings have implications for the development of targeted diagnostic and therapeutic strategies for MLD.

PMID:40184812 | DOI:10.1016/j.intimp.2025.114515

Food Chem. 2025 Apr 1;482:144174. doi: 10.1016/j.foodchem.2025.144174. Online ahead of print.

ABSTRACT

A key barrier in transitioning to plant-based, more satiating diets, is the bitter taste of plant proteins. We hypothesize that both, a more bitter tasting (MBT) and a less bitter tasting (LBT) pea protein hydrolysate (PPH) can be digested in the stomach into bitter tasting peptides that stimulate proton secretion (PS) and serotonin release, as two of the key gastric satiety signals, via the functional involvement of bitter taste receptors (TAS2Rs). Using a sensory-guided LC-MS approach, we identified six bitter peptides that were released from LBT-PPH and MBT-PPH during gastric digestion in vitro. TAS2R4 and TAS2R43 involvement in PS and serotonin release was confirmed via CRISPR-Cas9 knockout experiments. Our hypothesis was proven with all six peptides equally stimulating PS in immortalized human gastric HGT-1 cells, and LBT-PPH-derived peptides eliciting a higher serotonin release in HGT-1 cells than MBT-PPH peptides, indicating a satiating potential of less bitter tasting protein hydrolysates.

PMID:40184744 | DOI:10.1016/j.foodchem.2025.144174

Sci Rep. 2025 Apr 4;15(1):11628. doi: 10.1038/s41598-025-96427-3.

ABSTRACT

Proteasome inhibitor analogs (PIs) have significantly improved the degree of remission and survival rate of patients with multiple myeloma. However, serious adverse events (AEs) have hindered their clinical application. This study analyzed the AEs reported in the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS) database to determine the safety profile and differences for the PI drugs bortezomib, carfilzomib, and ixazomib. The reporting odds ratio (ROR) was used to detect safety signals. Significant safety signals were detected based on system-organ classification (SOC). For bortezomib, the most significant SOC signal was "blood and lymphatic system disorders" (ROR = 3.47, 95% CI 3.37-3.57), while the most significant PT signal was "enteric neuropathy" (ROR = 134.96, 95% CI 45.67-398.79). For carfilzomib, the most significant SOC signal being "blood and lymphatic system disorders" (ROR = 4.34, 95% CI 4.17-4.53), while the most significant PT signal was "light chain analysis increased" (ROR = 76.65, 95% CI 57.07-102.96). For ixazomib, the most significant SOC signal was "gastrointestinal disorders" (ROR = 2.04, 95% CI 1.96-2.12), while the most significant PT signal was "light chain analysis increased" (ROR = 67.15, 95% CI 45.36-99.42). For bortezomib and carfilzomib, the top 20 reported PTs were consistent with AEs listed in the drug information. For ixazomib, six unexpected AEs were observed: asthenia, malaise, pyrexia, decreased appetite, dehydration, and falls. The PIs were consistent with the early failure model based on time-series analysis of the occurrence of adverse reactions to the drug. The data mined from FAERS generates new AE signals, and further clinical studies are needed to validate these findings.

PMID:40185858 | DOI:10.1038/s41598-025-96427-3

Colloids Surf B Biointerfaces. 2025 Mar 26;252:114652. doi: 10.1016/j.colsurfb.2025.114652. Online ahead of print.

ABSTRACT

Idiopathic pulmonary fibrosis (IPF) is a fatal, progressive lung disease characterized by extensive scarring and thickening of lung tissue that leads to respiratory failure. Early and accurate diagnosis is crucial for monitoring disease progression and assessing therapeutic efficacy. While imaging modalities such as radiological X-rays and high-resolution computed tomography (HRCT) are commonly employed, magnetic resonance imaging (MRI) offers significant advantages, including superior soft tissue contrast and the absence of ionizing radiation. However, in lung MRIs are hindered by short transverse relaxation times, low proton density, and motion artifacts which is addressed herein by developing theranostic platform combining MRI imaging with targeted drug delivery using melanin nanoparticles conjugated with nintedanib (MNP-NIN). Chelation with ferric ions (MNP-NIN-Fe³⁺) enhanced MRI visibility enabling non-invasive imaging and drug tracking. MNP-NIN and MNP-NIN-Fe³ ⁺ nanoparticles were built with mean diameters of 189 ± 44 nm and 182 ± 35 nm, respectively and demonstrating successful NIN conjugation. Controlled NIN release followed zero-order kinetics over 36 days. MNP conjugation reduced cytotoxicity in BEAS-2B and A549 cells improving the drug's safety. MRI experiments conducted with a 7.0 T animal scanner demonstrated enhanced imaging contrast with MNP-NIN-Fe solutions compared to saline underscoring their potential for localized visualization and tracking within lung tissues. By integrating MRI diagnostics and targeted drug delivery, the MNP-NIN-Fe³ ⁺ system offers a promising solution to overcome current challenges in IPF management. This theranostic platform addresses the limitations of conventional imaging techniques while providing an innovative strategy for reducing drug-related systemic side effects and improving therapeutic efficacy.

PMID:40184721 | DOI:10.1016/j.colsurfb.2025.114652

Pediatr Allergy Immunol Pulmonol. 2025 Apr 4. doi: 10.1089/ped.2024.0142. Online ahead of print.

ABSTRACT

Background: Although the forced oscillation technique has been used for many years in children, there is still inconclusive results about its efficiency in cystic fibrosis (CF). Moreover, no studies have been conducted on impulse oscillometry (IOS) in children with primary ciliary dyskinesia (PCD). Methods: Age, sex, weight, height, body mass index, and oscillometric parameters were compared in 3-6-year-old children with CF and PCD and healthy children. Results: This prospective study included 27 children with CF, 21 with PCD, and 27 healthy children, with mean ages of 4.11 ± 1.08, 4.33 ± 1.23, and 4.41 ± 0.79 years, respectively. No significant differences were revealed in the comparison of the z-scores of the parameters of the CF group with those of the healthy group. However, in the PCD group, z-scores of R5 and Z5 were significantly higher than those in the healthy group (P = 0.018 and P = 0.008, respectively). In addition, z-scores of X10, X15, and X20 were significantly lower in children with PCD compared with the healthy group (P = 0.013, P = 0.033, and P = 0.029, respectively). Conclusions: This first study simultaneously reporting IOS results in preschool children with CF or PCD showed a significant difference of resistance and reactance of airways between PCD and healthy children. This study is also very significant in showing that IOS can be performed in young children who are unable to cooperate with spirometry. In contrast, no such differences were noted between CF and healthy controls, possibly due to thick mucus affecting sound wave transmission through the airways in CF. In addition, IOS may be less effective in detecting early pulmonary disease, as in some studies it failed to identify abnormalities in young children with CF even when spirometry is abnormal.

PMID:40184233 | DOI:10.1089/ped.2024.0142

Invest Radiol. 2025 Apr 7. doi: 10.1097/RLI.0000000000001186. Online ahead of print.

ABSTRACT

OBJECTIVE: The aim of the study was to compare the diagnostic quality of deep learning (DL) reconstructed balanced steady-state free precession (bSSFP) single-shot (SSH) cine images with standard, multishot (also: segmented) bSSFP cine (standard cine) in cardiac MRI.

METHODS AND MATERIALS: This prospective study was performed in a cohort of participants with clinical indication for cardiac MRI. SSH compressed-sensing bSSFP cine and standard multishot cine were acquired with breath-holding and electrocardiogram-gating in short-axis view at 1.5 Tesla. SSH cine images were reconstructed using an industry-developed DL super-resolution algorithm (DL-SSH cine). Two readers evaluated diagnostic quality (endocardial edge definition, blood pool to myocardium contrast and artifact burden) from 1 (nondiagnostic) to 5 (excellent). Functional left ventricular (LV) parameters were assessed in both sequences. Edge rise distance, apparent signal-to-noise ratio (aSNR) and contrast-to-noise ratio were calculated. Statistical analysis for the comparison of DL-SSH cine and standard cine included the Student's t-test, Wilcoxon signed-rank test, Bland-Altman analysis, and Pearson correlation.

RESULTS: Forty-five participants (mean age: 50 years ±18; 30 men) were included. Mean total scan time was 65% lower for DL-SSH cine compared to standard cine (92 ± 8 s vs 265 ± 33 s; P < 0.0001). DL-SSH cine showed high ratings for subjective image quality (eg, contrast: 5 [interquartile range {IQR}, 5-5] vs 5 [IQR, 5-5], P = 0.01; artifacts: 4.5 [IQR, 4-5] vs 5 [IQR, 4-5], P = 0.26), with superior values for sharpness parameters (endocardial edge definition: 5 [IQR, 5-5] vs 5 [IQR, 4-5], P < 0.0001; edge rise distance: 1.9 [IQR, 1.8-2.3] vs 2.5 [IQR, 2.3-2.6], P < 0.0001) compared to standard cine. No significant differences were found in the comparison of objective metrics between DL-SSH and standard cine (eg, aSNR: 49 [IQR, 38.5-70] vs 52 [IQR, 38-66.5], P = 0.74). Strong correlation was found between DL-SSH cine and standard cine for the assessment of functional LV parameters (eg, ejection fraction: r = 0.95). Subgroup analysis of participants with arrhythmia or unreliable breath-holding (n = 14/45, 31%) showed better image quality ratings for DL-SSH cine compared to standard cine (eg, artifacts: 4 [IQR, 4-5] vs 4 [IQR, 3-5], P = 0.04).

CONCLUSIONS: DL reconstruction of SSH cine sequence in cardiac MRI enabled accelerated acquisition times and noninferior diagnostic quality compared to standard cine imaging, with even superior diagnostic quality in participants with arrhythmia or unreliable breath-holding.

PMID:40184545 | DOI:10.1097/RLI.0000000000001186

J Psychol. 2025 Apr 4:1-24. doi: 10.1080/00223980.2025.2456801. Online ahead of print.

ABSTRACT

This study investigated the relationships between familial factors in terms of parental autonomy support and parental support and Vocational Education and Training (VET) students' learning motivation, learning approaches, and cognitive flexibility. In this cross-sectional study, a convenient sample of 557 VET students (males = 56.7% and females = 43.35; mean age = 18.41 and SD = 0.85) from ten vocational schools in Bangkok areas, Thailand, responded to a questionnaire of adapted scales on familial factors (i.e., parental autonomy support and parental support), learning motivation (i.e., intrinsic motivation, extrinsic motivation, and utility value), learning approaches (i.e., deep learning approaches and surface learning approaches), and cognitive flexibility (i.e., alternatives). Structural equation analyses revealed that parental autonomy support had indirect relationship with alternatives via learning motivation and deep learning approaches, whereas parental support had both direct and indirect association with alternatives through learning motivation and deep learning approaches. Surface learning approaches were not found to significantly predict alternatives. These findings suggest that a familial context that stresses autonomy support and helpful support from parents can motivate VET students to learn and adopt deep approaches to learning, which in turn encourages the development of their cognitive flexibility.

PMID:40184534 | DOI:10.1080/00223980.2025.2456801

Sci Adv. 2025 Apr 4;11(14):eadr7134. doi: 10.1126/sciadv.adr7134. Epub 2025 Apr 4.

ABSTRACT

Joint analysis of transcriptomic and T cell receptor (TCR) features at single-cell resolution provides a powerful approach for in-depth T cell immune function research. Here, we introduce a deep learning framework for single-T cell transcriptome and receptor analysis, MIST (Multi-insight for T cell). MIST features three latent spaces: gene expression, TCR, and a joint latent space. Through analyses of antigen-specific T cells, and T cell datasets related to lung cancer immunotherapy and COVID19, we demonstrate MIST's interpretability and flexibility. MIST easily and accurately resolves cell function and antigen specificity by vectorizing and integrating transcriptome and TCR data of T cells. In addition, using MIST, we identified the heterogeneity of CXCL13+ subsets in lung cancer infiltrating CD8+ T cells and their association with immunotherapy, providing additional insights into the functional transition of CXCL13+ T cells related to anti-PD-1 therapy that were not reported in the original study.

PMID:40184452 | DOI:10.1126/sciadv.adr7134

Oncotarget. 2025 Apr 4;16:249-255. doi: 10.18632/oncotarget.28709.

ABSTRACT

Recent advances in deep learning models have transformed medical imaging analysis, particularly in radiology. This editorial outlines how uncertainty quantification through embedding-based approaches enhances diagnostic accuracy and reliability in hepatobiliary imaging, with a specific focus on oncological conditions and early detection of precancerous lesions. We explore modern architectures like the Anisotropic Hybrid Network (AHUNet), which leverages both 2D imaging and 3D volumetric data through innovative convolutional approaches. We consider the implications for quality assurance in radiological practice and discuss recent clinical applications.

PMID:40184325 | DOI:10.18632/oncotarget.28709

IEEE Trans Pattern Anal Mach Intell. 2025 Mar 7;PP. doi: 10.1109/TPAMI.2025.3548810. Online ahead of print.

ABSTRACT

Zeroth-order optimization algorithms recently emerge as a popular research theme in optimization and machine learning, playing important roles in many deep-learning related tasks such as black-box adversarial attack, deep reinforcement learning, as well as hyper-parameter tuning. Mainstream zeroth-order optimization algorithms, however, concentrate on exploiting zeroth-order-estimated first-order gradient information of the objective landscape. In this paper, we propose a novel meta-algorithm called Hessian-Aware Zeroth-Order (ZOHA) optimization algorithm, which utilizes several canonical variants of zeroth-order-estimated second-order Hessian information of the objective: power-method-based, and Gaussian-smoothing-based. We conclude theoretically that ZOHA enjoys an improved convergence rate compared with existing work without incorporating in zeroth-order optimization second-order Hessian information. Empirical studies on logistic regression as well as the black-box adversarial attack are provided to validate the effectiveness and improved success rates with reduced query complexity of the zeroth-order oracle.

PMID:40184293 | DOI:10.1109/TPAMI.2025.3548810

IEEE Trans Neural Netw Learn Syst. 2025 Apr 4;PP. doi: 10.1109/TNNLS.2025.3551778. Online ahead of print.

ABSTRACT

Enhancing the prediction of volatile and intermittent electric loads is one of the pivotal elements that contributes to the smooth functioning of modern power grids. However, conventional deep learning-based forecasting techniques fall short in simultaneously taking into account both the temporal dependencies of historical loads and the spatial structure between residential units, resulting in a subpar prediction performance. Furthermore, the representation of the spatial graph structure is frequently inadequate and constrained, along with the complexities inherent in Spatial-Temporal data, impeding the effective learning among different households. To alleviate those shortcomings, this article proposes a novel framework: Spatial-Temporal fusion adaptive gated graph convolution networks (STFAG-GCNs), tailored for residential short-term load forecasting (STLF). Spatial-Temporal fusion graph construction is introduced to compensate for several existing correlations where additional information are not known or unreflected in advance. Through an innovative gated adaptive fusion graph convolution (AFG-Conv) mechanism, Spatial-Temporal fusion graph convolution network (STFGCN) dynamically model the Spatial-Temporal correlations implicitly. Meanwhile, by integrating a gated temporal convolutional network (Gated TCN) and multiple STFGCNs into a unified Spatial-Temporal fusion layer, STFAG-GCN handles long sequences by stacking layers. Experimental results on real-world datasets validate the accuracy and robustness of STFAG-GCN in forecasting short-term residential loads, highlighting its advancements over state-of-the-art methods. Ablation experiments further reveal its effectiveness and superiority.

PMID:40184286 | DOI:10.1109/TNNLS.2025.3551778

IEEE Trans Pattern Anal Mach Intell. 2025 Apr 4;PP. doi: 10.1109/TPAMI.2025.3558267. Online ahead of print.

ABSTRACT

By abusing access to a well-trained classifier, model inversion (MI) attacks pose a significant threat as they can recover the original training data, leading to privacy leakage. Previous studies mitigated MI attacks by imposing regularization to reduce the dependency between input features and outputs during classifier training, a strategy known as unilateral dependency optimization. However, this strategy contradicts the objective of minimizing the supervised classification loss, which inherently seeks to maximize the dependency between input features and outputs. Consequently, there is a trade-off between improving the model's robustness against MI attacks and maintaining its classification performance. To address this issue, we propose the bilateral dependency optimization strategy (BiDO), a dual-objective approach that minimizes the dependency between input features and latent representations, while simultaneously maximizing the dependency between latent representations and labels. BiDO is remarkable for its privacy-preserving capabilities. However, models trained with BiDO exhibit diminished capabilities in out-of-distribution (OOD) detection compared to models trained with standard classification supervision. Given the open-world nature of deep learning systems, this limitation could lead to significant security risks, as encountering OOD inputs-whose label spaces do not overlap with the in-distribution (ID) data used during training-is inevitable. To address this, we leverage readily available auxiliary OOD data to enhance the OOD detection performance of models trained with BiDO. This leads to the introduction of an upgraded framework, unknown-aware BiDO (BiDO+), which mitigates both privacy and security concerns. As a highlight, with comparable model utility, BiDO-HSIC+ reduces the FPR95 by $55.02\%$ and enhances the AUCROC by $9.52\%$ compared to BiDO-HSIC, while also providing superior MI robustness.

PMID:40184277 | DOI:10.1109/TPAMI.2025.3558267