Sci

Cement production in Africa remains carbon-intensive, primarily due to the use of coal-based thermal energy. This study conducts a comparative cradle-to-gate life cycle assessment (LCA) of cement production using 100% coal (Scenario A) against partial substitution with refuse-derived fuel (RDF) at a 20% thermal input rate (Scenario B), with case studies in South Africa and Ethiopia. The LCA, modeled in SimaPro 9.2.0.1 with Ecoinvent v3.7.1 and regional data, evaluates midpoint environmental impacts across the following five stages: raw materials, clinker production, electricity, fuel use, and transportation. The results show that Scenario B reduces the global warming potential (GWP) by 3.3–4.2% per kg of cement, with minimal increases in other impact categories. When avoided landfill methane is accounted for, GWP reduction improves to 6.7%. Fossil resource depletion drops by 10%, and toxicity and particulate emissions show marginal improvements. Economic analysis under South Africa’s 2025 carbon policy reveals a modest net cost increase of $2–3 per ton of cement and an abatement cost of $64–87 per ton of CO₂. The study provides new insights by harmonizing LCA models across national contexts, linking emissions reductions to economic instruments, and quantifying the co-benefits of RDF for waste management. The results support RDF co-processing as a scalable mitigation strategy for the African cement sector, recommending substitution rates of 15–30%, policy alignment, and enhancement of the RDF supply chain to maximize impact.

Large language models (LLMs) are rapidly being explored as tools to support learning and assessment in health science education, yet their performance across discipline-specific evaluations remains underexamined. This study evaluated the accuracy of two prominent LLMs on university-level pharmacognosy examinations and compared their performance to that of pharmacy students. Authentic exam papers comprising a range of question formats and content categories were administered to ChatGPT and DeepSeek using a structured prompting approach. Student data were anonymized and LLM responses were graded using the same marking criteria applied to student cohorts, and a Monte Carlo simulation was conducted to determine whether observed performance differences were statistically meaningful. Facility Index (FI) values were calculated to contextualize item difficulty and identify where LLM performance aligned or diverged from student outcomes. The models demonstrated variable accuracy across question types, with a stronger performance in recall-based and definition-style items and comparatively weaker outputs for applied or interpretive questions. Simulated comparisons showed that LLM performance did not uniformly exceed or fall below that of students, indicating dimension-specific strengths and constraints. These findings suggest that while LLM-resistant examination design is contingent on question structure and content, further research should refine their integration into pharmacy education.

Abductive reasoning—the search for plausible explanations—has long been central to human inquiry, from forensics to medicine and scientific discovery. Yet formal approaches in AI have largely reduced abduction to eliminative search: hypotheses are treated as mutually exclusive, evaluated against consistency constraints or probability updates, and pruned until a single “best” explanation remains. This reductionist framing fails on two critical fronts. First, it overlooks how human reasoners naturally sustain multiple explanatory lines in suspension, navigate contradictions, and generate novel syntheses. Second, when applied to complex investigations in legal or scientific domains, it forces destructive competition between hypotheses that later prove compatible or even synergistic, as demonstrated by historical cases in physics, astronomy, and geology. This paper introduces quantum abduction, a non-classical paradigm that models hypotheses in superposition, allowing them to interfere constructively or destructively, and collapses only when coherence with evidence is reached. Grounded in quantum cognition and implemented with modern NLP embeddings and generative AI, the framework supports dynamic synthesis rather than premature elimination. For immediate decisions, it models expert cognitive processes; for extended investigations, it transforms competition into “co-opetition” where competing hypotheses strengthen each other. Case studies span historical mysteries (Ludwig II of Bavaria, the “Monster of Florence”), literary demonstrations (Murder on the Orient Express), medical diagnosis, and scientific theory change. Across these domains, quantum abduction proves more faithful to the constructive and multifaceted nature of human reasoning, while offering a pathway toward expressive and transparent AI reasoning systems.