2051 papers
The range from Cs to Cy captures a vibrant spectrum of cutting-edge research, spanning everything from complex systems and cybernetics to cosmology and celestial mechanics. These fields explore how information flows through networks, how the universe began, and the fundamental laws governing matter and energy. While the original research is often dense and technical, the core ideas drive our understanding of both digital and physical realities.
At Gist.Science, we make this knowledge accessible by processing every new preprint in this category directly from arXiv. For each paper, our team provides a clear, plain-language summary alongside a detailed technical breakdown, ensuring that both experts and curious readers can grasp the latest breakthroughs without getting lost in the jargon. Below are the most recent additions to this collection, featuring fresh insights from the global scientific community.
This paper proposes a comprehensive ethical framework to guide the responsible development and deployment of foundational AI models in medical imaging, addressing critical challenges such as patient data privacy, algorithmic bias, and the need for transparency and human oversight to ensure patient-centered care.
This paper demonstrates that Large Language Model agents, when placed in a simulated survival environment, naturally evolve from a chaotic "state of nature" into an ordered society with an absolute sovereign, thereby empirically validating Thomas Hobbes's Social Contract Theory and highlighting the potential of LLM-driven multi-agent simulations for studying complex social dynamics.
This paper argues that claims of large language models passing the Turing Test are premature, as a rigorous, unconstrained implementation of Turing's original imitation game revealed that human judges could easily distinguish the AI from a human, unlike previous studies that may have used shorter time limits.
This paper proposes a new paradigm for Artificial General Intelligence called "True Intelligence" (TI), arguing that current performance-based definitions are inadequate and instead defining TI through six components—five assessable architectural pillars and one emergent quality of interconnectedness—while critiquing the field's neglect of four decades of cognitive architecture research.
This study proposes and evaluates machine learning models, including logistic regression and stacked generalization, to effectively detect fraudulent banking transactions in the post-pandemic digital landscape, achieving a maximum AUC of 0.954 through advanced data preprocessing and feature engineering techniques.
This paper proposes "Flow-by-Flow," a governance paradigm for high-loss AI domains that bypasses content evaluation to overcome human cognitive limits by imposing nonlinear costs on high-volume production and enforcing institutional capacity caps, thereby ensuring supervisory load remains manageable without relying on error-prone content judgment.
This study evaluates the socio-communicative competencies of three large language models (GPT-4o, Llama 3, and Command R+) in healthcare dialogues using the IC-MD instrument, finding that while they exhibit non-hostility, they lack the consistent sensitivity, non-intrusiveness, and structuring skills necessary for safe and effective use as healthcare advisors.
This paper argues that AI ethics must extend beyond software to address the semiconductor industry's intertwined social, environmental, and geopolitical challenges by proposing a participatory, cross-disciplinary framework that resolves tensions between sovereignty and sustainability, supply chain opacity, and knowledge divides through strategies like standardized emissions labeling and epistemic redistribution.
This paper demonstrates that crowd-sourced Google Maps Points of Interest can serve as a high-frequency, low-cost proxy for estimating household income at the sub-municipal level in São Paulo, achieving a strong predictive performance (R² of 0.65) that offers a viable alternative to costly, infrequent census data for social policy in Brazil.
This paper critiques current AI child safety evaluations for lacking real-world grounding and proposes a new framework based on interviews with 19 youth practitioners to identify harmful chatbot behaviors and define more effective, supportive responses for vulnerable youth.