Influencers exert significant promotional effects in tourism marketing; however, few empirical studies have examined how different types of tourism activities can strategically leverage these effects through influencer narrative style. To address this gap, this study draws on Heuristic-Analytic Theory to investigate how personal and professional narrative styles align with relaxing and challenging tourism activities to maximize influencers’ persuasiveness. The study analyzes secondary data from 1187 social media posts and conducts three scenario-based experiments with 1438 participants. Results show that personal narrative style is more persuasive for relaxing activities, whereas professional narrative style works better for challenging activities. Moreover, experience resonance and knowledge acquisition are identified as key underlying mechanisms. Notably, the study introduces temporal distance in travel decision as a moderator, highlighting the importance of dynamic temporal cues in influencer recommendations. The findings provide valuable theoretical contributions and practical implications for optimizing influencer marketing strategies in the tourism context.
As intelligent malware detection systems are deployed in evolving mobile ecosystems, both benign and malicious applications change in structure and behavior, inducing non-stationary data distributions that challenge long-term model reliability. In parallel, adversarial perturbations expose additional vulnerabilities in machine learning–based detectors. Despite extensive work on concept drift and adversarial robustness independently, their interaction in adaptive intelligent systems remains insufficiently characterized.We present a longitudinal, drift-aware evaluation of adversarial robustness across more than a decade of Android applications using static and dynamic feature representations extracted from emulator and real-device executions. The dataset is organized into yearly slices and evaluated under three deployment protocols that emulate realistic learning scenarios: (1) same-year training and testing, (2) cross-year deployment without model updates, and (3) expanding-window retraining with cumulative historical data. Across multiple classifier families, adversarial examples are generated using FGSM and SPSA under feasibility constraints. We measure clean performance, Adversarial Accuracy (AA), Attack Success Rate (ASR), and introduce temporal linkage metrics—RobustDrop, ΔASR, and Adversarial Amplification Factor (AAF)—to quantify the relationship between distribution shift and robustness degradation. Results show that temporal separation is associated with reductions in both clean accuracy and adversarial accuracy under the evaluated transfer-based feature-space setting. A controlled analysis shows that the residual drift–robustness relationship is configuration-dependent, with a stronger residual temporal association for static features than for dynamic features. Expanding-window retraining mitigates, but does not eliminate, robustness loss under continued distributional evolution. These findings highlight the need for drift-aware robustness assessment frameworks in long-lived adversarial environments.
Chapter 1. Introduction. Section One. Foundations of Teachers' Beliefs Research. Chapter 2. The Promises, Problems, and Prospects of Research on Teachers' Beliefs. Chapter 3. Historical Overview and Theoretical Perspectives of Research on Teachers' Beliefs. Chapter 4. The Development of Teachers' Beliefs. Chapter 5. The Relationship between Teachers' Beliefs and Teachers' Practices. Section Two. Studying Teachers' Beliefs. Chapter 6. Assessing Teachers' Beliefs: Challenges and Solutions. Chapter 7. Measuring Teachers' Beliefs: For What Purpose? Chapter 8. Qualitative Approaches to Studying Teachers' Beliefs. Chapter 9. Methods for Studying Beliefs: Teacher Writing, Scenarios, and Metaphor Analysis. Section Three. Teachers' Identity, Motivation, and Affect. Chapter 10. The Intersection of Identity, Beliefs, and Politics in Conceptualizing `Teacher Identity'. Chapter 11. A Motivational Analysis of Teacher Beliefs. Chapter 12. The Career Development of Preservice and Inservice Teachers: Why Teachers' Self-Efficacy Beliefs Matter. Chapter 13. A Hot Mess: Unpacking the Relation between Teachers' Beliefs and Emotions. Section Four. Contexts and Teachers' Beliefs. Chapter 14. Teachers' Beliefs about Teaching (and Learning). Chapter 15. Teachers' Instructional Beliefs and the Classroom Climate: Connections and Conundrums. Chapter 16. Teachers' Beliefs about Assessment. Chapter 17. Context Matters: The Influence of Collective Beliefs and Shared Norms. Section 5. Teachers' Beliefs about Knowing and Teaching within Academic Domains. Chapter 18. Personal Epistemologies and Teaching. Chapter 19. The Individual, the Context and Practice: A Review of the Research on Teachers' Beliefs Related to Mathematics. Chapter 20. Beliefs about Reading, Text, and Learning from Text. Chapter 21. Science Teachers' Beliefs: Perceptions of Efficacy and the Nature of Scientific Knowledge and Knowing. Chapter 22. Teachers' Beliefs about Social Studies. Chapter 23. Teacher Beliefs and Uses of Technology to Support 21st Century Teaching and Learning. Section Six. Teachers' Beliefs about Learners. Chapter 24. Preschool Teachers' Ideas about How Children Learn Best: An Examination of Beliefs about the Principles of Developmentally Appropriate Practice. Chapter 25. Teachers' Beliefs about Cultural Diversity: Problems and Possibilities. Chapter 26. Teachers' Beliefs about English Language Learners. Chapter 27. Teachers' Beliefs about Students with Special Needs and Inclusion.
Large Language Models (LLMs) have transformed the natural language processing landscape and brought to life diverse applications. Pretraining on vast web-scale data has laid the foundation for these models, yet the research community is now increasingly shifting focus toward post-training techniques to achieve further breakthroughs. While pretraining provides a broad linguistic foundation, post-training methods enable LLMs to refine their knowledge, improve reasoning, enhance factual accuracy, and align more effectively with user intents and ethical considerations. Fine-tuning, reinforcement learning, and test-time scaling have emerged as critical strategies for optimizing LLMs performance, ensuring robustness, and improving adaptability across various real-world tasks. This survey provides a systematic exploration of post-training methodologies, analyzing their role in refining LLMs beyond pretraining, addressing key challenges such as catastrophic forgetting, reward hacking, and inference-time trade-offs. We highlight emerging directions in model alignment, scalable adaptation, and inference-time reasoning, and outline future research directions. We also provide a public repository to continually track developments in this fast-evolving field: https://github.com/mbzuai-oryx/Awesome-LLM-Post-training.
Characterizing the genetic structure and connectivity between populations of endangered species can be used to inform management actions. In vagile species with high gene flow or recently established populations, such characterizations can be difficult to undertake using traditional genetic markers, and genetic stock identification (GSI) may be confounded by allele-sharing between populations. Loggerhead sea turtles (Caretta caretta) in the southeastern United States comprise seven management units (MUs) based on female philopatry inferred via mitochondrial DNA sequences, yet nuclear microsatellite data do not reflect divergence. Further, loci for accurate GSI are not currently known. To address this, we generated genome-wide single nucleotide polymorphism (SNP) data from 146 females nesting at individual sites representative of each southeastern United States MU. We found weak (FST=0.001–0.003) but significant divergence among all MUs, with more notable divergence between the Gulf Coast and Atlantic Ocean MUs, and amongst the Atlantic Ocean MUs. We then used an iterative leave-one-out approach to identify candidate loci for GSI. This approach identified loci that could assign individuals to natal ocean basins (i.e., to the Gulf Coast or to the Atlantic Ocean), and to individual MUs within the Atlantic Ocean, with high (≥90
