La Roche University is a private university in McCandless, Pennsylvania. It was founded in 1963 by the Sisters of Divine Providence as a Roman Catholic college and now sits on an 80-acre (320,000 m2) campus in McCandless within the Diocese of Pittsburgh.
As AI-generated art becomes more prevalent, questions arise about how such works are perceived and valued across cultures. This study examined Chinese participants’ (N = 471) ability to distinguish AI-generated from human-created artworks and compared their performance to previously collected U.S. data (N = 406). Participants evaluated ten artworks (five AI-generated, five human-created), judged authorship, indicated willingness to purchase, and Chinese participants entered price offers. Results show that only four artworks were identified above chance, three below chance, and three at chance, indicating the difficulty of distinguishing AI art from human works. Chinese participants were more accurate in recognizing AI-generated art, while U.S. participants better identified human-created works. Across both groups, willingness to purchase was consistently higher when an artwork was believed to be human-created, though price differences appeared for only two pieces among Chinese participants. Despite prior research suggesting relative openness toward AI in some Chinese contexts, these findings suggest persistent preference for human authorship. Limitations include the use of a small stimulus set and the omission of qualitative appraisal data, which will be addressed in future research. This cross-cultural design highlights how cultural context shapes perceptions of authorship and authenticity in the age of AI-generated visual production. Rather than addressing broader debate about whether AI should be considered creative, this exploratory study examines how viewers in two cultural contexts interpret authorship, authenticity, and related evaluative judgments surrounding AI-generated art.
Distributed Denial of Service (DDoS) is found to be the most disruptive threat for the current network infrastructures, especially with the advent of cloud computing, Internet of Things (IoT), and Software-Defined Networking (SDN). The detection methods traditionally used fail to effectively counter the sophistication of DDoS attacks. The recent developments in machine learning (ML) and deep learning (DL) algorithms have greatly impacted the detection of DDoS attacks, as these algorithms can well identify the complex, nonlinear behavior of DDoS attacks. Class imbalance has been reported to affect the detection of DDoS attacks by contemporary detection systems. Then we propose an innovative hybrid model that uses Artificial Neural Networks (ANN) and Decision Trees (DT). The proposed hybrid model has achieved an impressive accuracy of 98.77% on the CICDDoS2019 data set, with precision, recall, and F1-score values of 98%, 99%, and 98.5%, respectively, outperforming the traditional detection methods. The presented hybrid model has effectively addressed the problem of class imbalance, providing an effective solution for the detection of DDoS attacks.
Asparagus racemosus (Shatavari) is one of the most valuable medicinal plants in traditional medicine and has been widely used for centuries to manage reproductive, gastrointestinal, neurological, and inflammatory disorders. Growing scientific interest has prompted extensive investigations into its phytochemical composition and pharmacological activities. This review provides a comprehensive overview of the botanical characteristics, phytochemical constituents, traditional applications, and therapeutic potential of A. racemosus. The plant is rich in steroidal saponins, particularly shatavarins, along with flavonoids, alkaloids, polyphenols, essential fatty acids, vitamins, and trace minerals that collectively contribute to its diverse biological activities. Experimental and clinical studies demonstrate that A. racemosus exhibits antioxidants, anti-inflammatory, immunomodulatory, adaptogenic, antimicrobial, antiulcer, antidiabetic, hepatoprotective, neuroprotective, galactagogue, aphrodisiac, and anticancer properties. In addition, emerging evidence supports its beneficial effects on female reproductive health, cognitive function, stress reduction, and metabolic regulation. Despite substantial preclinical evidence, further well-designed clinical trials, standardized phytochemical profiling, pharmacokinetic investigations, and long-term safety evaluations are required to establish its therapeutic efficacy and facilitate its integration into evidence-based medicine. Although substantial preclinical evidence is available, further well-designed in vivo studies and clinical trials, together with standardized phytochemical profiling, pharmacokinetic investigations, and long-term safety evaluations, are needed to establish therapeutic efficacy and safety and support its integration into evidence-based medicine.
Mental health conditions have appeared to be a significant societal health issue in modern society, impacting members of various demographic and socio-economic groups. Even though recent developments in machine learning and deep learning have demonstrated that these methods have been promising in the field of mental health assessment, current techniques lack scalability in generalization, sensitivity to noise, and lack of combination of both time and latent features representations. To overcome these shortcomings, we introduce the Mental Health Feature Fusion Network (MH-fuse-net), a new hybrid framework that incorporates Autoencoder-based representation learning, Bidirectional Long Short-Term Memory (LSTM)-based temporal dependency modelling, and Light Gradient Boosting machine (LightGBM)-based vigorous ensemble classification. An autoencoder terminates redundancy and noise from the features, and a Bidirectional Long Short-Term Memory (BiLSTM) apprehends bidirectional contextual patterns in the mental health data. LightGBM then improves decision-making robustness and generalization by leveraging gradient-boosted trees. The model is compared to the popular baseline techniques such as Logistic Regression, single-user Autoencoders, BiLSTM, and XGBoost. Experimental that MH-FuseNet performs significantly better than the baseline models, with better accuracy, precision, recall, F1-score, and ROC-AUC.
WRN is an enzyme critical for DNA repair and genome stability and is a promising synthetic lethal target for MSI cancers. 40-70 % of patients (pts) with MSI/dMMR solid tumors do not respond to immune checkpoint inhibitors (ICI) or develop resistance, representing an unmet need in MSI cancers. RO7589831 is a novel first-in-class covalent, irreversible WRN inhibitor that induces dose-dependent DNA damage and tumor growth inhibition in MSI preclinical models, supporting its clinical evaluation in MSI pts. This open-label, multi-center, FIH study assesses the safety, pharmacokinetics (PK), pharmacodynamics (PD) and preliminary anti-tumor activity of RO7589831 in pts with MSI and/or dMMR advanced solid tumors (NCT06004245). The primary objective is to determine the maximum tolerated dose (MTD) and/or recommended Phase 2 dose (RP2D). 14 paired tumor biopsies were collected for translational analysis. As of Nov 25, 2024, 44 pts (24 F, 20 M; median age 62y (32-77y)) were enrolled in six dose cohorts at RO7589831 daily doses from 150 mg to 2000 mg. Based on clinical first dose PK data (rapid absorption, mean Tmax 2.6 h [± 1.3h] and half-life 4.4 h [±1.8h], n=41), RO7589831 was dosed orally once (QD) or twice (BID) daily. Among the 44 pts (22 colorectal (CRC) and 22 non-CRC), median prior treatment lines were 3 (range 1-12); 89% had received ICIs. 35 pts had MSI tumors (18 CRC and 17 non-CRC). No dose-limiting toxicities (DLTs) have been reported and the MTD has not been reached. The most common treatment-emergent AEs across all doses were nausea (52.3%), diarrhea (34.1%), and vomiting (31.8%). Grade (G)3 treatment related AEs (TRAE) were nausea and increased AST/ALT in 2 pts each (4.3%) as well as fatigue and anemia in 1 pt each (2.1%). No G4 or worse TRAEs were observed. Of the efficacy evaluable MSI pts (n= 32), four RECISTv1.1 partial responses (PRs) (2 confirmed and 2 ongoing unconfirmed PRs) were observed in post-ICI-treated MSI CRC (n=1), ovarian (n=1), and endometrial cancer (n=2) for up to 9.5+ mths. Disease control rate (DCR) was 68.8% (95%CI 51.13, 86.37) and 20 (62.5%) pts achieved RECISTv1.1 stable disease (SD). 48.4% (15/31) of FDG-PET scan evaluable pts across all dose levels showed metabolic responses to RO7589831. All pts with RECIST PR showed pronounced metabolic responses (-50% to -90%). ctDNA molecular responses were observed in 2/2 pts with RECISTv1.1 PR, 8/9 pts with SD and 0/5 pts with disease progression. RO7589831 is generally safe and well tolerated, with promising DCR, durable RECISTv1.1, FDG-PET metabolic and ctDNA molecular responses in pts with advanced MSI cancers, including ICI-treated pts, providing the first early clinical proof-of-concept for effectively drugging WRN. Dose optimization cohorts are ongoing to establish the RP2D. Timothy A. Yap, Natalie Cook, Elisa Fontana, Erica S. Tsang, Oliver Bechter, Marwan Fakih, Kristoffer S. Rohrberg, Emiliano Calvo, Elena M. Elez, Meredith Pelster, John H. Strickler, Ignacio Matos, Adam Sharp, Sophia Blake, David Dejardin, Stephen Fowler, Nina Henkel, Michael Hettich, Angelika Lahr, Ting Liu, Christophe Meille, Piergiorgio Pettazzoni, Denise Reeves, Barbara Romagnoli, Sotirios Sotiriou, Daniel Waterkamp, Yali Fu, Sophie Postel-Vinay. First-in-human (FIH) phase 1 trial of the oral first-in-class covalent Werner helicase (WRN) inhibitor RO7589831 in patients with microsatellite instable (MSI) and/or mismatch repair deficient (dMMR) advanced solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_2):Abstract nr CT016.