Search for: All records

Low-income students remain underrepresented in STEM (Science, Technology, Engineering, and Mathematics) fields, with high attrition rates posing challenges to educational equity and workforce development. This study examined how participation in the S-STEM program at Tennessee State University (TSU) influences students’ sense of belonging (SOB). A survey of 101 undergraduate STEM students (61 S-STEM scholars, 40 non-scholars) was analyzed using descriptive statistics, t-tests, ANOVA, and linear regression. Results showed no statistically significant differences in overall SOB between S-STEM scholars and non-scholars; however, regression analysis revealed that campus environment and students’ willingness to choose the department again were significant positive predictors of SOB, whereas S-STEM participation, faculty support, faculty inclusion of diverse perspectives, and administrative support were not significant. These findings suggest that while S-STEM provides important financial and community resources, perceptions of a welcoming campus environment and overall departmental satisfaction are stronger determinants of belonging, highlighting the importance of broader institutional integration and supportive environments for low-income STEM students. 

We study operads with trivial A-actions and prove an equivalence between the category of A-trivial operads and that of pseudo-graded-Perm associative algebras. As a consequence, we show that finitely generated A-trivial operads are right noetherian of integral Gelfand-Kirillov dimension and that every element in a prime A-trivial operad is central. 

This study employed an empirical research design to examine factors influencing engaged learning among undergraduate engineering students. The participants consisted of 43 undergraduates enrolled in the college of engineering at Tennessee State University. Data were collected using Schreiner and Louis’s engaged learning survey, which assesses students’ demographic characteristics, academic performance, learning satisfaction, and levels of engaged learning. Following data cleaning and imputation, a feature selection procedure was conducted. Hierarchical linear regression analyses were first performed to examine the incremental contributions of demographic variables, academic performance indicators, and satisfaction-related factors to engaged learning. Based on the full model, key predictors were subsequently entered into simplified hierarchical regression models to identify the most influential variables. The selected predictors were then entered into a final linear regression model to evaluate their overall effects. The results indicated that overall satisfaction was the strongest and most consistent positive predictor of engaged learning among engineering undergraduates. Learning satisfaction and critical thinking gain showed weaker, though positive, correlations with engaged learning, whereas ethnicity exhibited a negative relationship. These findings emphasize the importance of students’ overall educational experience in fostering engaged learning in engineering education and suggest that efforts to enhance student satisfaction may contribute to improved engagement and learning outcomes in engineering programs. 

AI-generated images have become pervasive, raising critical concerns around content authenticity, intellectual property, and the spread of misinformation. Invisible watermarks offer a promising solution for identifying AI-generated images, preserving content provenance without degrading visual quality. However, their real-world robustness remains uncertain due to the lack of standardized evaluation protocols and large-scale stress testing. To bridge this gap, we organized “Erasing the Invisible,” a NeurIPS 2024 competition and newly established benchmark designed to systematically stress testing the resilience of watermarking techniques. The competition introduced two attack tracks—Black-box and Beige-box—that simulate practical scenarios with varying levels of attacker knowledge on watermarks, providing a comprehensive assessment of watermark robustness. The competition attracted significant global participation, with 2,722 submissions from 298 teams. Through a rigorous evaluation pipeline featuring real-time feedback and human-verified final rankings, participants developed and demonstrated new attack strategies that revealed critical vulnerabilities in state-of-the-art watermarking methods. On average, the top-5 teams in both tracks could remove watermarks from $$\geq$$ 89% of the images while preserving high visual quality, setting strong baselines for future research on watermark attacks and defenses. To support continued progress in this field, we summarize the insights and lessons learned from this competition in this paper, and release the benchmark dataset, evaluation toolkit, and competition results. “Erasing the Invisible” establishes a valuable open resource for advancing more robust watermarking techniques and strengthening content provenance in the era of generative AI. 

Polycrystalline graphite and the MAX phase TiSiC are layered crystalline solids with similar deformation mechanisms, including basal slip, ripplocation boundaries (RBs), kink boundaries (KBs), and cracking. The interplay of these mechanisms, notably in energy dissipation, has been much discussed in the past twenty-five years. This study builds upon previous work, investigating deformation with a renewed emphasis on the bulk-scale and given recent findings concerning RBs. Our investigation compares the evolution of energy dissipation, nonlinear recoverable and irrecoverable strain, and damage upon increasing stress for graphite and TiSiC. Benitez et al.’s (2016) methodology of compressive cyclic loading and post-mortem electron backscatter diffraction (EBSD) to assess the prevalence of kinking based on low-angle grain boundaries (LAGBs) was used. Strains were measured with digital image correlation and EBSD was conducted on TiSiC leveraging dictionary indexing, which was necessary herein to identify LAGBs accurately. The stress–strain stages of TiSiC agree with literature on TiAlC. Damage and energy dissipation were more accelerated in graphite. No significant difference was observed in the fraction of LAGBs between pristine and unloaded TiSiC. Trends observed and EBSD evidence that KBs were not dominant suggest that RBs are the primary dissipator of energy in both materials. 

Finetuned large language models (LLMs) have shown remarkable performance in financial tasks, such as sentiment analysis and information retrieval. Due to privacy concerns, finetuning and deploying financial LLMs (FinLLMs) locally are crucial for institutions and individuals. In this paper, we employ quantized low-rank adaptation (QLoRA) to finetune FinLLMs, which leverage low-rank structure and quantization technique to significantly reduce computational requirements while maintaining model performance. We also employ data and pipeline parallelism to enable local finetuning on commodity GPUs. Experiments on financial datasets validate the efficacy of our approach in yielding notable improvements over the base models.