Dynamic Routing and Knowledge Re-Learning for Data-Free Black-Box Attack.

Deep learning models have emerged as strong and efficient tools that can be applied to a broad spectrum of complex learning problems and many real-world applications. However, more and more works show that deep models are vulnerable to adversarial examples. Compared to vanilla attack settings, this paper advocates a more practical setting of data-free black-box attack, for which the attackers can completely not access the structures and parameters of the target model, as well as the intermediate features and any training data associated with the model.

Building an Open-Vocabulary Video CLIP Model With Better Architectures, Optimization and Data.

Despite significant results achieved by Contrastive Language-Image Pretraining (CLIP) in zero-shot image recognition, limited effort has been made exploring its potential for zero-shot video recognition. This paper presents Open-VCLIP++, a simple yet effective framework that adapts CLIP to a strong zero-shot video classifier, capable of identifying novel actions and events during testing. Open-VCLIP++ minimally modifies CLIP to capture spatial-temporal relationships in videos, thereby creating a specialized video classifier while striving for generalization.

Adaptive Cross-Modal Transferable Adversarial Attacks From Images to Videos.

The cross-model transferability of adversarial examples makes black-box attacks to be practical. However, it typically requires access to the input of the same modality as black-box models to attain reliable transferability. Unfortunately, the collection of datasets may be difficult in security-critical scenarios.

Towards Transferable Adversarial Attacks on Image and Video Transformers.

The transferability of adversarial examples across different convolutional neural networks (CNNs) makes it feasible to perform black-box attacks, resulting in security threats for CNNs. However, fewer endeavors have been made to investigate transferable attacks for vision transformers (ViTs), which achieve superior performance on various computer vision tasks. Unlike CNNs, ViTs establish relationships between patches extracted from inputs by the self-attention module.

Blueberry extracts antagonize Aβ neurotoxicity and exert a neuroprotective effect through MEK-ERK-BDNF/UCH-L1 signaling pathway in rat and mouse hippocampus.

Background: The neuroprotective potential of blueberry (BB) extracts against Alzheimer's disease (AD) has been previously hinted at, while its exact mechanism has remained largely enigmatic.

Objective: Our study endeavored to unravel the impacts and mechanisms by which BB extracts ameliorated the learning and memory prowess of AD-afflicted mice, with a specific focus on the MEK-ERK pathway.

Methods: We employed 3-month-old APP/PS1 transgenic mice and stratified them into three distinct groups: AD+BB, AD, and control (CT).

Circular RNA CircPDS5B impairs angiogenesis following ischemic stroke through its interaction with hnRNPL to inactivate VEGF-A.

Background: Ischemic stroke (IS) is the primary cause of mortality and disability worldwide. Circular RNAs (circRNAs) have been proposed as crucial regulators in IS. This study focused on the role of circPDS5B in IS and its underlying mechanism.

The Prognosis of Pineal Parenchymal Tumors: Development and Validation of a Nomogram Based on Surveillance, Epidemiology and End Results.

Background: Pineal parenchymal tumors are exceedingly rare, and optimal disease management has yet to be defined. In this study, we aimed to identify prognostic factors and establish a predictive model for the prognosis of patients with pineal parenchymal tumors.

Methods: All patients with pineal parenchymal tumors in the Surveillance, Epidemiology and End Results database between 1975 and 2019 were reviewed.

Blueberry-Mulberry Extract Alleviates Cognitive Impairment, Regulates Gut Metabolites, and Inhibits Inflammation in Aged Mice.

Cognitive impairment is associated with aging; however, the underlying mechanism remains unclear. Our previous study found that polyphenol-rich blueberry-mulberry extract (BME) had an antioxidant capability and effectively alleviated cognitive impairment in a mouse model of Alzheimer's disease. Thus, we hypothesized that BME would improve cognitive performance in naturally aging mice and assessed its effects on related signaling pathways.

Association between circular stapler size and anastomotic leakage after laparoscopic low anterior resection for rectal cancer.

Background: Anastomotic leakage (AL) is one of the most severe and frequent complications occurring after laparoscopic low anterior resection (LAR) for rectal cancer. This study aimed to examine the association between circular stapler size and AL after laparoscopic LAR.

Methods: This retrospective single-institution study involved 181 patients with rectal cancer who underwent laparoscopic LAR performed by a single surgical team between July 2016 and June 2021.

LncRNA PVT1 Promotes Neuronal Cell Apoptosis and Neuroinflammation by Regulating miR-488-3p/FOXD3/SCN2A Axis in Epilepsy.

It is vital to understand the mechanism of epilepsy onset and development. Dysregulated lncRNAs are closely associated with epilepsy. Our work probed the role of lncRNA PVT1/miR-488-3p/FOXD3/SCN2A axis in epilepsy.

Generalized Meta-FDMixup: Cross-Domain Few-Shot Learning Guided by Labeled Target Data.

The vanilla Few-shot Learning (FSL) learns to build a classifier for a new concept from one or very few target examples, with the general assumption that source and target classes are sampled from the same domain. Recently, the task of Cross-Domain Few-Shot Learning (CD-FSL) aims at tackling the FSL where there is a huge domain shift between the source and target datasets. Extensive efforts on CD-FSL have been made via either directly extending the meta-learning paradigm of vanilla FSL methods, or employing massive unlabeled target data to help learn models.

5 kHz, 4.2mJ, 900 ps end-pumped Nd:YVO MOPA laser system.

A 5 kHz sub-nanosecond master oscillator power amplifier (MOPA) laser system was reported in this paper. The master oscillator was an electro-optically Q-switched Nd:YVO laser directly pumped at 879 nm, yielding a pulse energy of 520 µJ and a pulse width of 900 ps at 5 kHz. With two Nd:YVO amplifiers directly pumped at 914 nm, the pulse energy was further scaled up.

Molecular diagnosis and treatment of meningiomas: an expert consensus (2022).

Abstract: Meningiomas are the most common primary intracranial neoplasm with diverse pathological types and complicated clinical manifestations. The fifth edition of the WHO Classification of Tumors of the Central Nervous System (WHO CNS5), published in 2021, introduces major changes that advance the role of molecular diagnostics in meningiomas. To follow the revision of WHO CNS5, this expert consensus statement was formed jointly by the Group of Neuro-Oncology, Society of Neurosurgery, Chinese Medical Association together with neuropathologists and evidence-based experts.

Incision pressing, a simple and effective intervention to reduce colorectal surgical site infection: A propensity score-matched study.

Background: Colorectal surgery is associated with a high risk of surgical site infection (SSI). In March 2017, we developed an intervention, called "PRESS", with the aim of reducing colorectal superficial SSI. This study assessed the effect of the new intervention in reducing the rates of superficial SSI in colorectal surgery.

Acute exposure to simulated high-altitude hypoxia alters gut microbiota in mice.

Gut microbiota bears adaptive potential to different environments, but little is known regarding its responses to acute high-altitude exposure. This study aimed to evaluate the microbial changes after acute exposure to simulated high-altitude hypoxia. C57BL/6 J mice were divided into hypoxia and normoxia groups.

Building roof extraction from ASTIL echo images applying OSA-YOLOv5s.

Light detection and ranging (LiDAR) is a type of essential tool for urban planning and geoinformation extraction. Airborne streak tube imaging LiDAR (ASTIL) is a new system with great advantages in the rapid collection of remote sensing data. To the best of our knowledge, a new method to extract a building roof from the echo images of ASTIL is proposed.

The mediating role of self-efficacy of managing chronic disease between the dual-mode of self-control and the fatigue in breast cancer patients undergoing postoperative chemotherapy.

Background: Fatigue is prevalent in breast cancer patients undergoing postoperative chemotherapy, which seriously affects physical and mental health. The present study aimed to investigate the relevance of fatigue, the self-efficacy of managing chronic disease (SEMCD), and the dual-mode of self-control (DMSC) in patients.

Methods: Three hundred and seventy six breast cancer patients undergoing postoperative chemotherapy participated in this cross-sectional study.

Risk factors impacting intra- and postoperative cerebrospinal fluid rhinorrhea on the endoscopic treatment of pituitary adenomas: A retrospective study of 250 patients.

We aimed to identify the risk factors associated with intra- and postoperative cerebrospinal fluid (CSF) leakage in pituitary adenomas treated with endoscopic transsphenoidal surgery.This study is a retrospective analysis of 250 pituitary adenoma cases from January 2017 to December 2019 at our hospital. All patients underwent endoscopic endonasal transsphenoidal surgeries.

Molecular and Clinical Characterization of a Novel Prognostic and Immunologic Biomarker GPSM3 in Low-Grade Gliomas.

as the most common malignancy of the central nervous system, low-grade glioma (LGG) patients suffered a poor prognosis. Tumor microenvironment, especially immune components, plays an important role in the progression of tumors. Thus, it is critical to explore the key immune-related genes, a comprehensive understanding of the TME in LGG helps us find novel cancer biomarkers and therapeutic targets.

LncRNA RBPMS-AS1 promotes NRGN transcription to enhance the radiosensitivity of glioblastoma through the microRNA-301a-3p/CAMTA1 axis.

Objective: Glioblastoma (GBM) is the most frequent brain malignancy with high incidence, and long noncoding RNAs (lncRNAs) exerts functions in GBM. In this research, we focused on the capabilities of lncRNA RBPMS-AS1 in radiosensitivity of GBM.

Methods: RBPMS-AS1 and CAMTA1 expression levels were determined in GBM tissues and cells.

CD73-positive extracellular vesicles promote glioblastoma immunosuppression by inhibiting T-cell clonal expansion.

Extracellular vesicles are involved in the occurrence, progression and metastasis of glioblastoma (GBM). GBM can secrete a variety of tumour-derived extracellular vesicles (TDEVs) with high immunosuppressive activity that remotely suppress the systemic immune system, and therapy targeting TDEVs has potential efficacy. In this study, we detected a higher concentration of CD73+ TDEVs enriched in exosomes in central and peripheral body fluids of GBM patients than in those of patients with other brain tumours (low-grade glioma or brain metastases from melanoma or non-small-cell lung cancer).

Anthocyanin-rich blueberry extracts and anthocyanin metabolite protocatechuic acid promote autophagy-lysosomal pathway and alleviate neurons damage in in vivo and in vitro models of Alzheimer's disease.

Objectives: As the global aging phenomenon intensifies, the incidence of Alzheimer's disease (AD) is gradually increasing. Diet appears to be an effective way to prevent and delay the progression of AD. Previous studies have found that cognitive impairment and neuronal damage were effectively alleviated by blueberry extract (BBE) in AD mice, but its mechanism is still unclear.