ai research

Current AI research often treats models as static artifacts, overlooking the fundamental influence of training dynamics on critical properties such as capability, bias, robustness, and safety. This work proposes shifting the focus toward the training process itself to establish a science of AI centered on training dynamics. By analyzing the interactions among data, objectives, architectures, and optimizers, the paper develops a theoretical framework that is predictive, intervenable, and design-oriented. Integrating approaches from mechanistic interpretability, fairness, memory mechanisms, and simplicity biases, it uncovers causal links between early-training signals and final model behavior. The study systematically outlines key challenges and open problems, offering both theoretical pathways and practical foundations for extending scaling laws beyond performance to encompass multidimensional model attributes.

This paper addresses the extended Turing Test question—“Can machines grow?”—by proposing GROW-AI, the first systematic framework for assessing AI developmental maturity. It establishes four evaluation arenas—psychological, ethical, computational, and robotic—and introduces six standardized, game-based benchmark tasks (C1–C6). Leveraging behavioral logging, expert-weighted scoring, and a quantified Grow Up Index, GROW-AI enables reproducible, cross-modal (LLMs, robots, software agents) assessment of AI evolution from immaturity to maturity. Its core innovation lies in adapting human developmental psychology constructs to AI evaluation, thereby establishing the first traceable, multidimensional, process-oriented growth assessment system. Empirical validation demonstrates its capacity to identify maturity bottlenecks and evolutionary trajectories across diverse AI systems, confirming both universality and effectiveness. (149 words)

This study addresses the prevailing gap in AI education, which emphasizes model development while neglecting system engineering practices, leaving students ill-equipped to handle real-world challenges such as architectural design, deployment, and monitoring. To bridge this gap, the authors implemented a master’s-level course in which students built a movie recommendation system under realistic constraints, with a focus on integrating AI components into robust software systems, adopting data-driven machine learning practices, and cultivating systems-level thinking. Using a mixed-methods approach—combining analysis of student project artifacts with survey data—the research evaluates learners’ performance in architectural decision-making, integration of heterogeneous models, and adaptation to evolving requirements. Findings reveal common difficulties students encounter in AI system engineering and demonstrate the course’s effectiveness in addressing critical deficiencies in AI engineering education and enhancing systems-aware competencies.

This study addresses the lack of systematic synthesis at the intersection of artificial intelligence (AI) and modeling and simulation (M&S) by proposing, for the first time, a structured framework based on the full M&S lifecycle—encompassing model construction, input modeling, execution, experimentation, validation, and output analysis. It elucidates the bidirectional integration mechanisms between AI and simulation: how AI enhances or substitutes traditional simulation components, and how simulation supports AI training and evaluation. Incorporating generative AI technologies such as large language models, the paper identifies representative application paradigms and integration approaches across each phase, synthesizes key achievements, and presents a conceptual roadmap tailored to the rapidly evolving ecosystem, while highlighting current limitations and open research challenges.

This study addresses the challenges organizations face—such as skill gaps, ethical concerns, and data governance issues—when integrating artificial intelligence (AI) into experimental innovation approaches like growth hacking, lean startup, design thinking, and agile methodologies. Drawing on a systematic literature review of 37 studies published between 2018 and 2024 and guided by the PRISMA 2020 framework, this work offers the first integrative synthesis of AI with these four dominant experimental innovation paradigms. It reveals cross-method synergies through which AI enhances iterative cycles, boosts creativity, and improves decision-making via data analytics, real-time feedback, automation, and process optimization. The findings not only demonstrate AI’s significant contribution to amplifying experimental innovation efficacy but also identify critical success factors and implementation barriers, thereby proposing a strategic adoption pathway encompassing talent development, data governance, and ethical guidelines.

This study investigates the mechanisms underlying the emergence of social bias in artificial intelligence systems, practitioners’ understandings of these issues, and potential mitigation strategies. Employing a qualitative multiple-case design grounded in an interpretivist paradigm, the research integrates intersectionality theory and cognitive science, drawing on semi-structured interviews, document analysis, and triangulation to examine AI practitioners’ experiences across design, development, and governance. Findings reveal that algorithmic bias is deeply rooted in historical inequities, exclusionary assumptions, and organizational pressures for efficiency, underscoring the insufficiency of purely technical fixes. The study proposes an innovative approach that embeds ethical considerations early in the development lifecycle, strengthens structural accountability, fosters diverse stakeholder participation and cognitive awareness, and actively reshapes organizational culture to cultivate AI systems that are transparent, accountable, and aligned with community values.

This work addresses the critical gap between the widespread deployment of AI models and the limited understanding of their internal mechanisms, as conventional benchmarks often fail to uncover root causes of failures such as hallucination and shortcut learning. It proposes the first systematic “model science” framework, integrating paradigms from cognitive science, neuroscience, and related disciplines to enable in-depth analysis of individual model instances through four complementary lenses: Verify, Explore, Steer, and Refine. By establishing a shared knowledge repository and collaborative research infrastructure, the framework transcends the limitations of population-level performance evaluation, offering both theoretical foundations and practical pathways to enhance AI interpretability, reliability, and continuous improvement. This paradigm shift moves AI research beyond performance-centric metrics toward a deeper, understanding-driven approach.

Current AI systems struggle to autonomously execute end-to-end scientific discovery pipelines in neuroscience, particularly due to a lack of capability in assessing scientific plausibility. This work presents the first systematic evaluation of general-purpose code-generating agents on a real-world, large-scale optogenetics task in Drosophila, whose scale and complexity substantially exceed existing benchmarks. By integrating iterative code generation, visualization of intermediate outputs, and rigorous domain-expert-driven assessment, the study reveals that while agents can successfully complete individual pipeline stages, they fail to reliably orchestrate the full workflow. Performance degrades markedly in the absence of predefined evaluation criteria. These findings highlight critical limitations in AI’s capacity for scientific reasoning and self-evaluation, and propose principles for constructing and evaluating agent-based approaches to open-ended scientific problems.