Search

This study investigates the strategic decision-making abilities of large language models (LLMs) via the game of Tic-Tac-Toe, renowned for its straightforward rules and definitive outcomes. We developed a mobile application coupled with web services, facilitating gameplay among leading LLMs, including Jurassic-2 Ultra by AI21, Claude 2.1 by Anthropic, Gemini-Pro by Google, GPT-3.5-Turbo and GPT-4 by OpenAI, Llama2-70B by Meta, and Mistral Large by Mistral, to assess their rule comprehension and strategic thinking. Using a consistent prompt structure in 10 sessions for each LLM pair, we systematically collected data on wins, draws, and invalid moves across 980 games, employing two distinct prompt types to vary the presentation of the game's status. Our findings reveal significant performance variations among the LLMs. Notably, GPT-4, GPT-3.5-Turbo, and Llama2 secured the most wins with the list prompt, while GPT-4, Gemini-Pro, and Mistral Large excelled using the illustration prompt. GPT-4 emerged as the top performer, achieving victory with the minimum number of moves and the fewest errors for both prompt types. This research introduces a novel methodology for assessing LLM capabilities using a game that can illuminate their strategic thinking abilities. Beyond enhancing our comprehension of LLM performance, this study lays the groundwork for future exploration into their utility in complex decision-making scenarios, offering directions for further inquiry and the exploration of LLM limits within game-based frameworks. [ABSTRACT FROM AUTHOR]

Photocatalytic degradation, as a very significant advanced oxidation technology in the field of environmental purification, has attracted extensive attention in recent years. The design and synthesis of catalysts with high-intensity photocatalytic properties have been the focus of many researchers in recent years. In this contribution, two new supramolecular materials {[(L1)·(Ag4I7)]CH3CN} (1), {[(L2)·(Ag4I7)]CH3CN} (2) were synthesized by solution volatilization reaction of two cationic templates 1,3,5-Tris(4-aminopyridinylmethyl)-2,4,6-Trimethylphenyl bromide (L1) and 1,3,5-Tris(4-methyl pyridinyl methyl)-2,4,6-trimethylphenyl bromide (L2) with metal salt AgI at room temperature, respectively. The degradation effect of 1 and 2 as catalyst on tetracycline (TC) under visible light irradiation was studied. The results showed that the degradation of TC by 1 was better than that by 2 and both of them had good stability and cyclability. The effects of pH value, catalyst dosage, and anion in water on the photocatalytic performance were also investigated. The adsorption kinetics fit the quasi-first-order model best. After 180 min of irradiation with 1, the degradation rate of TC can reach 97.91%. In addition, the trapping experiments showed that ·OH was the main active substance in the photocatalytic degradation of TC compared with ·O2− and h+. Because of its simple synthesis and high removal efficiency, catalyst 1 has potential value for the treatment of wastewater containing organic matter. [ABSTRACT FROM AUTHOR]

Most Artificial Intelligence (AI) implementations so far are based on the exploration of how the human brain is designed. Nevertheless, while significant progress is shown on specialized tasks, creating an Artificial General Intelligence (AGI) remains elusive. This manuscript proposes that instead of asking how the brain is constructed, the main question should be how it was evolved. Since neurons can be understood as intelligent agents, intelligence can be thought of as a construct of multiple agents working and evolving together as a society, within a long-term memory and evolution context. More concretely, we suggest placing Multiple Neighborhood Cellular Automata (MNCA) on a blockchain with an interaction protocol and incentives to create an AGI. Given that such a model could become a "strong" AI, we present the conjecture that this infrastructure is possible to simulate the properties of cognition as an emergent phenomenon. [ABSTRACT FROM AUTHOR]

Endgame studies have long served as a tool for testing human creativity and intelligence. We find that they can serve as a tool for testing machine ability as well. Two of the leading chess engines, Stockfish and Leela Chess Zero (LCZero), employ significantly different methods during play. We use Plaskett's Puzzle, a famous endgame study from the late 1970s, to compare the two engines. Our experiments show that Stockfish outperforms LCZero on the puzzle. We examine the algorithmic differences between the engines and use our observations as a basis for carefully interpreting the test results. Drawing inspiration from how humans solve chess problems, we ask whether machines can possess a form of imagination. On the theoretical side, we describe how Bellman's equation may be applied to optimize the probability of winning. To conclude, we discuss the implications of our work on artificial intelligence (AI) and artificial general intelligence (AGI), suggesting possible avenues for future research. [ABSTRACT FROM AUTHOR]

Artificial general intelligence (AGI) progression metrics indicate AGI will occur within decades. No proof exists that AGI will benefit humans and not harm or eliminate humans. A set of logically distinct conceptual components is proposed that are necessary and sufficient to (1) ensure various AGI scenarios will not harm humanity, and (2) robustly align AGI and human values and goals. By systematically addressing pathways to malevolent AI we can induce the methods/axioms required to redress them. Distributed ledger technology (DLT, "blockchain") is integral to this proposal, e.g., "smart contracts" are necessary to address the evolution of AI that will be too fast for human monitoring and intervention. The proposed axioms: (1) Access to technology by market license. (2) Transparent ethics embodied in DLT. (3) Morality encrypted via DLT. (4) Behavior control structure with values at roots. (5) Individual bar-code identification of critical components. (6) Configuration Item (from business continuity/disaster recovery planning). (7) Identity verification secured via DLT. (8) "Smart" automated contracts based on DLT. (9) Decentralized applications--AI software modules encrypted via DLT. (10) Audit trail of component usage stored via DLT. (11) Social ostracism (denial of resources) augmented by DLT petitions. (12) Game theory and mechanism design. [ABSTRACT FROM AUTHOR]

The formation phenomena of silver carbonate (Ag2CO3)–silver iodide (AgI) solid solutions were investigated by X-ray diffraction, thermogravimetry-differential thermal analysis, and electrical conductivity measurement. Results revealed that AgI and Ag2CO3 reacted with each other when mixed at room temperature. The reaction products were classified into three types: (1) AgI-based solid solutions in the AgI-rich region for x = 10% or less in x Ag2CO3–(1 − x) AgI; (2) Ag2CO3-based solid solutions in the Ag2CO3-rich region for x = 60% or more; and (3) silver carbonate iodides in the intermediate range for x between 10% and 60%. For the AgI-based solid solutions, the incorporation of Ag2CO3 into the AgI lattice expanded the unit cell and enhanced electrical conductivity. The solubility limit of Ag2CO3 into the AgI lattice estimated from the differential thermal analysis was x ≈ 5%. [ABSTRACT FROM AUTHOR]

Diabetes mellitus, a group of metabolic disorders characterized by hyperglycemia, is one of the most serious and common diseases around the world and is associated with major complications such as diabetic neuropathy, retinopathy, and cardiovascular diseases. A widely used treatment for non-insulin-dependent diabetes is α-glucosidase inhibitors (AGIs) such as acarbose, which hinders hydrolytic cleavage of disaccharides and retard glucose absorption. The ability to inhibit α-glucosidase activity has been reported in leaf and fruit of pepper (Capsicum annuum L.). In this study, we aimed to identify quantitative trait loci (QTLs) controlling α-glucosidase inhibitory activity (AGI activity) in pepper leaf and fruit using enzyme assay and genotyping-by-sequencing (GBS) analysis. The AGI activity at three stages of leaf and one stage of fruit development was analyzed by 96 F2 individuals. GBS analysis identified 17,427 SNPs that were subjected to pepper genetic linkage map construction. The map, consisting of 763 SNPs, contained 12 linkage groups with a total genetic distance of 2379 cM. QTL analysis revealed seven QTLs (qAGI1.1, qAGI11.1, qAGI5.1, qAGI9.1, qAGI12.1, qAGI5.2, and qAGI12.2) controlling AGI activity in pepper leaf and fruit. The QTLs for AGI activity varied by plant age and organ. This QTL information is expected to provide a significant contribution to developing pepper varieties with high AGI activity. [ABSTRACT FROM AUTHOR]

An exemplary paradigm of how an AI can be a disruptive technological paragon via the utilization of blockchain comes straight from the world of deep learning. Data scientists have long struggled to maintain the quality of a dataset for machine learning by an AI entity. Datasets can be very expensive to purchase, as, depending on both the proper selection of the elements and the homogeneity of the data contained within, constructing and maintaining the integrity of a dataset is difficult. Blockchain as a highly secure storage medium presents a technological quantum leap in maintaining data integrity. Furthermore, blockchain's immutability constructs a fruitful environment for creating high quality, permanent and growing datasets for deep learning. The combination of AI and blockchain could impact fields like Internet of things (IoT), identity, financial markets, civil governance, smart cities, small communities, supply chains, personalized medicine and other fields, and thereby deliver benefits to many people. [ABSTRACT FROM AUTHOR]