Shanghai Jinyuan Senior High School

Using immunotherapy to treat breast cancer, one of the most common cancers that affects women and the second most common cancer among women in the United States, has achieved many successes. However, how to solve the problem of cancer resistance to therapies caused by myeloid-derived suppressor cells and immunosuppressive tumor microenvironment remains a major challenge. To improve the performance of immunotherapy, we chose to explore the three putative therapeutic approaches including anti IL 8 antibody injection, PDH inhibition, and the use of tasquinimod, then combining them to observe anti-tumor effects in a murine model. We anticipate the greatest decrease of tumor volume in triple combination, concluding the feasibility of our method on treating breast cancer.

With recent breakthroughs in computer vision and deep learning, people's lives have become more convenient. However, urban barrier-free facilities for the equal mobility and independence of the elderly and disabled are increasingly ignored. Traditional recognition methods—such as manual annotation, laser radar (LiDAR), infrared imaging, and other optical sensing technologies—face limitations in scalability, accuracy, and cost. This paper proposes an advanced deep learningbased recognition system that integrates optical sensing data and visual imagery, leveraging the YOLOv8 object detection architecture to automatically identify barrier-free elements (e.g., ramps, handrails, steps) in complex urban environments. The multi-layered system architecture encompasses heterogeneous data acquisition (including RGB cameras, LiDAR point clouds, and infrared sensors), model training, deployment, and GIS-based map integration. A composite dataset was constructed by fusing public imagery, optical sensor data, and synthetic augmentations to enhance robustness under diverse urban conditions. Our model achieves a mean Average Precision (mAP) of 87.6% across scenarios with variable lighting and weather, outperforming previous methods such as YOLOv5, SSD, and Faster RCNN in both precision and inference speed. The feasibility of deployment on mobile and embedded platforms is further discussed, with an emphasis on real-time processing and sensor fusion. This study makes both technical and social contributions by introducing a scalable, real-time, and inclusive urban accessibility detection framework that harnesses the synergy of optical and visual recognition technologies for next-generation smart cities.

NBA (national basketball association), It was officially founded in 1949, consisting with 30 basketball teams that compete together for the championship. This event has become one of the most influential sports competitions globally and has garnered a massive fan base and audience. This study mainly focuses on the impacts exerted on the NBA by the following three dimensions: Young fan groups and middle-aged/elderly fan groups; Different geographical regions; Fans with different psychological motivations like emotionally driven fans and investment-driven fans. this study mainly conduct our research by collecting and analyzing data. Different ages spectators --Young fans drive digital engagement & emerging consumption. Middle-aged/elderly support traditional revenue. NBA responds with streaming partnerships (ESPN+) & nostalgic content for seniors. Domestic fans in the United States form the foundation of the NBA, while China serves as the cornerstone of the NBA's digital and broadcasting economy. Countries like those in Europe are key directions for the NBA's major expansion efforts. Emotionally driven fans require emotional connections with teams, such as the display of retro jerseys or commentary from retired star players. Investment-driven fans need relevant guidance and data support provided by teams.

This paper explores the role of artificial intelligence (AI) platforms in supporting web developers, with a focus on their practical application and impact on the development process. The study compares a website coded by students during specialized Information Technology (IT) classes with automatically generated code produced by AI tools. The research also examines how AI-driven platforms can assist students in refining their original projects.

With economic development and the transformation of social structures, the social roles of contemporary Chinese women have undergone a significant shift. This study focuses on the complex social roles of modern Chinese women, particularly in the contexts of choice, conflict, and the integration of work and family duties. In order to fully explore the multiple choices, conflicts, and balancing of social roles among contemporary Chinese women, this study conducted qualitative research by using literature review methods as well as semi-structured interviews with 40 adult women from Beijing, Shanghai, and Chengdu. The findings reveal that contemporary Chinese women have three predominant types of social roles: career woman, full-time housewife, and a combination of both. The study identifies conflict over time, responsibilities, and the worth of those responsibilities in shifting roles, shaped by the changes brought by the era of economic growth, policies, and social attitudes. Regarding the difficult balancing and choice women face between work and family responsibilities, various stakeholders can take measures to provide support. For instance, governments may introduce childcare subsidies and anti-discrimination legislation, businesses may establish female-friendly workplace policies, family members may help share the burden, and individual women may enhance their capacity for self-adaptation.

Translanguaging, a pedagogical strategy that alternates between the use of L1 and L2 in EFL/ESL classrooms, has been shown to enhance comprehension, facilitate scaffolding, and promote inclusive and equitable classroom participation in language education. Although translanguaging has attracted growing academic attention in China, little is known about how university English teachers employ translanguaging strategies in multilingual classrooms in Xinjiang. This area would benefit from such research, as it responds to Xinjiang’s multilingual reality and supports high-quality English education in minority areas. This research investigated teachers’ pedagogical translanguaging practices and students’ attitudes in college English classrooms at Yili Normal University in Xinjiang. Adopting a mixed-methods design consisting of an 8-week classroom observation, a questionnaire survey with 170 students (89 minority and 81 Han), and 18 semi-structured interviews, this study aims to categorise the types of translanguaging practices employed by teachers, examine both Han and minority students’ attitudes toward these practices, and identify the factors shaping such attitudes. The findings show that teachers’ translanguaging practices fell into three categories—explanatory, managerial, and interpersonal—comprising 12 specific subcategories. Quantitative results indicate that a majority of students expressed positive attitudes (scoring above 4) towards these practices, with minority and Han students showing overall optimism but diverging in certain classroom contexts, with minority students less likely to view translanguaging as signalling deficiency or hindering language acquisition (t=2.712, p=0.007). Thematic analysis further revealed four key factors influencing students’ attitudes: English proficiency, teachers’ language choice, the perceived psychological distance between Chinese and English, and awareness of linguistic diversity. As one of the early systematic attempts to explore translanguaging in higher education settings in China’s ethnic minority regions, this research proposes that teachers’ adoption of translanguaging practices, such as using bilingual explanations, the use of activities explicitly encouraging students to draw on their full linguistic repertoires, as well as teacher training programmes, should include practical workshops on translanguaging pedagogy.

Runaway stars depart their birthplaces with high peculiar velocities. Two mechanisms are commonly invoked to explain their origin, the binary supernova scenario and the dynamical ejection scenario. Investigating the kinematic properties of runaway stars is key to understanding their origins. We investigated the origins of B-type runaway stars from LAMOST using orbital traceback analysis. From the LAMOST catalog, we selected B-type runaway stars and determined their spectral subtypes from key absorption lines. We then derived atmospheric parameters for each star using the Stellar Label Machine, which is trained on TLUSTY synthetic spectra computed under the nonlocal thermodynamic equilibrium assumption. Using the derived atmospheric parameters as input, we estimated stellar masses and ages with a machine learning model trained on PARSEC evolutionary tracks. We finally performed orbital traceback with GALPY to analyze their origins. Through orbital traceback, we find that stars have trajectories entirely within the Galactic disk, whereas are disk-passing yet still trace back to the disk. Two stars have trajectories that intersect those of known clusters. Their orbits show similar morphologies in both the X-Y and R-Z planes, and their M/H values are comparable, suggesting possible cluster origins. However, definitive confirmation will require additional evidence. In addition, the V_ ̊m Sp - v i plane shows that runaway stars with low peculiar space velocities but high v remain on the Galactic disk, whereas those with high peculiar space velocities but low v pass through the disk, possibly reflecting two distinct origins.