Subjects: Computer Science >> Computer Application Technology submitted time 2024-04-10
Exploration of the Integration and Application of Large Model and Standard Literature Knowledge Base
Abstract： In the context of artificial intelligence and big data technology, the use of large models and the construction of standard literature knowledge bases are of great value for scientific research innovation, knowledge mining, and information retrieval. The standard literature knowledge base provides solid support for the standardization and standardization of various industries. This study first explores the current status of standard literature, then builds a framework for integrating large models and standard literature knowledge bases based on retrieval enhancement, and proposes exploration of enhancement optimization in each stage. Finally, it looks forward to future research directions and application prospects.
Peer Review Status:
Awaiting Review
Subjects: Other Disciplines >> Synthetic discipline submitted time 2023-03-28 Cooperative journals: 《中国科学院院刊》
Abstract：以 ChatGPT及其大语言模型为代表的人工智能将对经济学研究范式产生深远影响。目前，经济学实证研究一般使用参数维数较少、经济可解释性较强的小模型。然而，经济金融系统是一个超高维动态复杂系统，受多种因素的影响，且这些因素之间的关系呈现非线性与时变性特征，小模型无法刻画其本质规律。大模型可有效减少系统偏差，更好刻画复杂经济系统的特征与运行规律；而海量数据的使用可避免模型过度拟合，使大模型具有较好的泛化能力即样本外预测能力。为支持经济学及社会科学其他领域的大模型的估计、推断与预测，需要利用人工智能技术整合各种异构、异源、异频数据，构建大规模计量经济学数据库，并加强大算力等信息技术基础设施建设。目前，ChatGPT 及大模型等前沿人工智能技术仍存在局限性，如：无法像人类一样进行批判性思考或想象，只有预测能力；基于大数据的人工智能因果推断本质上是一种统计关系推断，需要引入经济理论或实验方法帮助识别真正的因果关系；人工智能技术不能改变经济学实证研究从样本推断总体性质的本质；同时，由于互联网大数据存在大量虚假信息或样本选择偏差等问题，基于人工智能所获得的结论的可靠性需要验证。
Subjects: Library Science，Information Science >> Information Science submitted time 2024-05-08
Abstract： Purpose/Significance This paper starts from the field of humanities and social sciences, and compares the model performance of humanities and social sciences from the aspects of basic knowledge and academic texts of humanities and social sciences. It aims to provide a systematic large language model evaluation benchmark for the field of humanities and social sciences for the reference of researchers in humanities and social sciences related fields. Methods/Processes Seven evaluation tasks related to the field of humanities and social sciences were designed and corresponding indicators were selected. On this basis, the current open-source and high-performance general-purpose domain Chinese large language models were selected to complete the domain-specific tasks in the form of questions and answers by invoking the local models, and their performance in the field of humanities and social sciences was quantitatively evaluated by selecting relevant indicators. Results/Conclusions The evaluation results show that among the open-source models selected in this paper, Qwen has the best performance, followed by Baichuan2, InternLM, and Atom is the worst performer in both the base model and the dialog model; moreover, in most cases, the dialog model shows more superior performance compared to the base model.
Subjects: Agriculture, Forestry,Livestock & Aquatic Products Science >> Other Disciplines of Agriculture, Forestry,Livestock & Aquatic Products Science submitted time 2024-06-26 Cooperative journals: 《农业图书情报学报》
Abstract： Purpose/Significance The new quality productivity advancing AI technology, especially exemplified by large language models (LLMs), is rapidly updating and attracting wide attention. In order to accelerate the implementation of AI technologies, it is urgent for advanced AI technologies to acquire support from knowledge resources in scientific and technological (S & T) information and libraries. Meanwhile, S & T information provides significant potential service scenarios for the application of AI technologies such as LLMs. This study aims to explore and design the method and path for constructing AI-ready data resources in the field of S & T information, and proposes a comprehensive and operable construction model that adapts to the new technical environment of AI, thereby facilitating comprehensive readiness in the field of intelligence. Method/Process This study first focuses on the concept and development status of AI-ready construction, and examines the development of AI-ready construction at home and abroad from three aspects: governments, enterprises and research institutions. The survey shows that the application of artificial intelligence has been highly valued by various fields of scientific research and production. However, the groundwork and preparation for AI applications are still relatively lagging behind, and AI tools cannot be fully implemented in key application scenarios due to the lack of high-quality and refined data resources. Based on the research results, the study made a preliminary definition of AI-ready construction, that is, we defined AI-ready construction as: the various development and improvement actions to adapt the object to the AI technical environment and promote the long-term benefits. The research then focuses on the field of S & T information, and systematically discusses and designs the AI-ready construction mode in the field of S & T information from six aspects: connotation category, construction angle, construction object, construction principle, control dimension and types of construction mode. Results/Conclusions The construction of AI-ready S & T information resources is a comprehensive and multi-angle transformation and upgrading process, which is located between the knowledge resource end and the intelligence application end. It is carried out in four aspects, including standards, methods, tools and platforms. The main content of the construction includes channels of AI technology, data transformation, data resources, and data management. At the same time, the construction is comprehensively controlled by six principles and four control dimensions. Besides, this study proposes the way of the practi cal construction of AI-ready S & T data resources, including the construction of intelligent data systems, and the construction of integrated platforms for the whole life cycle of S&T information data. The path reflects the process of the variation of knowledge resources from diversification to organization and then to integration, which not only serves the scientific information field itself, but also provides more intelligent, convenient, rich and powerful S&T information support for various fields. In the future, it is hoped that further research can delve into more micro and practical aspects, review the specific characteristics of different AI technologies, and provide more detailed suggestions for specific application scenarios at the operational level, providing a solid guarantee for scientific research institutions to achieve the leading strategic position in research and development.
Subjects: Library Science，Information Science >> Information Science submitted time 2024-03-01
Abstract： Objective This paper aims to improve the accuracy of automatic extraction of key technical words and corresponding function words from patent. Methods ChatGPT was used as the Teacher-model, and ChatGLM3 was used as the Student-model. Through knowledge distillation method, the training data generated by ChatGPT was used to fine-tune ChatGLM3, and multiple technical word extraction models and a function word extraction model were obtained. The technical words are extracted from the abstract, the first claim and the technical function paragraph, respectively, by using multiple technical word extraction models, and the function words are extracted from the technical function paragraph by using the function words extraction model. Results Compared with ChatGPT, the fine-tuned multiple technical word extraction models and function word extraction model show higher accuracy and lower recall rate, when extracting technical words and function words. The ChatGLM3 fine-tuning model of the first claim has the highest accuracy and F1 values of 0.734 and 0.724 respectively. Moreover, The accuracy of the function words extracted by the function word extraction model is 0.649, which is higher than the accuracy of the function words labeled by the commercial tool, which is is 0.53. Limitations The technical field and patent language of this research are single, the amount of patent verification data is small, and the data cleaning rules expect to be further optimized. Conclusions This research scheme improves the efficiency accuracy of automatic extraction of large language model through knowledge distillation operation. At the same time, this study can support the mining of cutting-edge innovative and hot technologies from patent texts, and support higher quality intelligent patent analysis.
Subjects: Other Disciplines >> Synthetic discipline submitted time 2023-10-09 Cooperative journals: 《心理学报》
Abstract： Self-injury usually emerges in early adolescence and has a high incidence among adolescents worldwide. Self-injury not only damages body tissue but is also associated with depression, anxiety, personality disorders, substance abuse, and a higher-than-average risk of suicide. Given the high incidence of self-injury and the severity of its consequences, it is important to explore its predictors and specific mechanisms. Interpersonal theories of developmental psychopathology maintain that interpersonal stress is a critical risk factor for adolescent self-injury behavior. However, the ways the source and duration of exposure to that stress affect adolescent self-injury behavior are unclear. Adolescents also differ in their sensitivity to interpersonal stress. Stress-related genetic factors may play an important moderating role. The current study selected child abuse and recent peer victimization as distal and proximal interpersonal stress, respectively, and FKBP5 gene rs3800373 polymorphism as the genetic factor. The purpose of this study was to build upon the results of previous studies by exploring the relative and interactive effects of distal and proximal interpersonal stress on adolescent self-injury behavior.The participants were 436 adolescents (12.84 ± 0.89 years, 49.8% males) recruited from four junior high schools in Guizhou Province. All were tracked from grade 7 to grade 9. At Time 1, adolescents reported child abuse via the Parent-Child Conflict Tactics Scale, reported peer victimization via the Multidimensional Peer Victimization Scale, and reported self-injury behavior via the Short Version of Self-Injury Behavior Scale. At Time 2, adolescents reported peer victimization and self-injury behavior, and saliva samples were collected. Genotyping with respect to the FKBP5 gene was performed with Agena MassArray software, and the corresponding typing results were analyzed using MassARRAY Typer software version 4.0.Results showed that both distal and proximal interpersonal stress significantly predicted adolescent self-injury behavior, but the relative effect sizes differed in early and middle adolescence. Across the entire sample, distal and proximal interpersonal stress had an interactive effect on adolescent self-injury, and the interaction pattern was consistent with the stress amplification model. However, when the FKBP5 gene was considered, the interaction pattern was found to differ between adolescents in different genotype groups. Specifically, compared with AA homozygous adolescents who experienced less childhood abuse, those who experienced more child abuse were easily to be impacted by recent peer victimization and engage in NSSI. This was consistent with the stress amplification model. However, in adolescents with the AC/CC genotype who experienced more childhood abuse, mild recent peer victimization triggered adolescent self-injury. These participants showed lower self-injury thresholds and higher scores for self-injury than those who experienced less childhood abuse, which was consistent with a stress sensitization model. These relationships were stable in both early and middle adolescence.These findings showed different patterns of interaction between interpersonal and intrapersonal factors on self-injury behavior in adolescents of different genotypes. Using an integrative, dynamic, and developmental framework, this study provides important insights into the relevant interpersonal theories. It is also valuable for the accurate identification of adolescents at high risk of self-injury and for both prevention and intervention.
Subjects: Statistics >> Social Statistics submitted time 2024-03-27 Cooperative journals: 《中国科学院院刊》
Abstract： In recent years, artificial intelligence (AI) has achieved numerous disruptive breakthroughs in frontier scientific and technological fields, such as AlphaFold2 for protein structure prediction, intelligent control of nuclear fusion, and drug design for COVID-19. These achievements indicate that AI for Science is becoming a new paradigm in research. To achieve fundamental scientific innovation and major technological breakthroughs in the era of intelligence, two core issues should be addressed: 1) how to harness the generality and creativity of the new-generation of AI, especially generative AI and large language models (LLMs), to promote the formation of new paradigms; 2) how to empower and transform traditional scientific facilities using AI. To tackle these challenges, this study proposes a concept of AI-enabled scientific facility (AISF) that caters to the requirements of both establishment of totally new intelligent scientific facility and AI empowerment of existing scientific facilities. It aims to construct an infrastructure system for AI for Science, enabling innovative functionalities such as scientific large language models (LLMs), generative simulation and inversion, autonomous intelligent unmanned experiments, and large-scale trustworthy scientific collaboration. These advancements will accelerate scientific discoveries, synthesis of transformative materials, and application of related engineering technologies.
Subjects: Agriculture, Forestry,Livestock & Aquatic Products Science >> Other Disciplines of Agriculture, Forestry,Livestock & Aquatic Products Science submitted time 2024-06-17 Cooperative journals: 《智慧农业（中英文）》
Abstract： Significance Big Models, or Foundation Models, have offered a new paradigm in smart agriculture. These models, built on the Transformer architecture, incorporate numerous parameters and have undergone extensive training, often showing excellent performance and adaptability, making them effective in addressing agricultural issues where data is limited. Integrating big models in agriculture promises to pave the way for a more comprehensive form of agricultural intelligence, capable of processing diverse inputs, making informed decisions, and potentially overseeing entire farming systems autonomously. Progress The fundamental concepts and core technologies of big models are initially elaborated from five aspects: the generation and core principles of the Transformer architecture, scaling laws of extending big models, large-scale self-supervised learning, the general capabilities and adaptions of big models, and the emerging capabilities of big models. Subsequently, the possible application scenarios of the big model in the agricultural field are analyzed in detail, the development status of big models is described based on three types of the models: Large language models (LLMs), large vision models (LVMs), and large multi-modal models (LMMs). The progress of applying big models in agriculture is discussed, and the achievements are presented. Conclusions and Prospects The challenges and key tasks of applying big models technology in agriculture are analyzed. Firstly, the current datasets used for agricultural big models are somewhat limited, and the process of constructing these datasets can be both expensive and potentially problematic in terms of copyright issues. There is a call for creating more extensive, more openly accessible datasets to facilitate future advancements. Secondly, the complexity of big models, due to their extensive parameter counts, poses significant challenges in terms of training and deployment. However, there is optimism that future methodological improvements will streamline these processes by optimizing memory and computational efficiency, thereby enhancing the performance of big models in agriculture. Thirdly, these advanced models demonstrate strong proficiency in analyzing image and text data, suggesting potential future applications in integrating real-time data from IoT devices and the Internet to make informed decisions, manage multi-modal data, and potentially operate machinery within autonomous agricultural systems. Finally, the dissemination and implementation of these big models in the public agricultural sphere are deemed crucial. The public availability of these models is expected to refine their capabilities through user feedback and alleviate the workload on humans by providing sophisticated and accurate agricultural advice, which could revolutionize agricultural practices.
Subjects: Agriculture, Forestry,Livestock & Aquatic Products Science >> Other Disciplines of Agriculture, Forestry,Livestock & Aquatic Products Science submitted time 2024-06-17 Cooperative journals: 《智慧农业（中英文）》
Abstract： Significance The crop phenotype is the visible result of the complex interplay between crop genes and the environment. It reflects the physiological, ecological, and dynamic aspects of crop growth and development, serving as a critical component in the realm of advanced breeding techniques. By systematically analyzing crop phenotypes, researchers can gain valuable insights into gene function and identify genetic factors that influence important crop traits. This information can then be leveraged to effectively harness germplasm resources and develop breakthrough varieties. Utilizing data-driven, intelligent, dynamic, and non-invasive methods for measuring crop phenotypes allows researchers to accurately capture key growth traits and parameters, providing essential data for breeding and selecting superior crop varieties throughout the entire growth cycle. This article provides an overview of intelligent identification technologies for crop agronomic traits and morphological structural phenotypes. Progress Crop phenotype acquisition equipment serves as the essential foundation for acquiring, analyzing, measuring, and identifying crop phenotypes. This equipment enables detailed monitoring of crop growth status. The article presents an overview of the functions, performance, and applications of the leading high-throughput crop phenotyping platforms, as well as an analysis of the characteristics of various sensing and imaging devices used to obtain crop phenotypic information. The rapid advancement of high-throughput crop phenotyping platforms and sensory imaging equipment has facilitated the integration of cutting-edge imaging technology, spectroscopy technology, and deep learning algorithms. These technologies enable the automatic and high-throughput acquisition of yield, resistance, quality, and other relevant traits of large-scale crops, leading to the generation of extensive multi-dimensional, multiscale, and multi-modal crop phenotypic data. This advancement supports the rapid progression of crop phenomics. The article also discusses the research progress of intelligent recognition technologies for agronomic traits such as crop plant height acquisition, crop organ detection, and counting, as well as crop ideotype recognition, crop morphological information measurement, and crop three-dimensional reconstruction for morphological structure intelligent recognition. Furthermore, this article outlines the main challenges faced in this field, including: difficulties in data collection in complex environments, high requirements for data scale, diversity, and preprocessing, the need to improve the lightweight nature and generalization ability of models, as well as the high cost of data collection equipment and the need to enhance practicality. Conclusions and Prospects Finally, this article puts forward the development directions of crop phenotype intelligent recognition technology, including: developing new and low cost intelligent field equipment for acquiring and analyzing crop phenotypes, enhancing the standardization and consistency of field crop phenotype acquisition, strengthening the generality of intelligent crop phenotype recognition models, researching crop phenotype recognition methods that involve multi-perspective, multimodal, multi-point continuous analysis, and spatiotemporal feature fusion, as well as improving model interpretability.
Subjects: Computer Science >> Natural Language Understanding and Machine Translation submitted time 2023-07-29
Abstract： At present, LLM（Large Language Model）adopts the technical path of "attention mechanism+deep learning"+"Reinforcement learning". In the field of AIGC (Artificial Intelligence Generated Content)，significant progress has been made, sparking hope for realization of AGI (Artificial General Intelligence). However, in those fields that need to interact with the actual environment, such as elderly care, family nanny, agricultural production, vehicle driving and other fields, the cost of trial and error is very high, and the reinforcement learning process that requires a lot of trial and error is difficult to achieve. So, in order to achieve universal artificial intelligence that can be applied to any field, we need to not only utilize existing technology, but also solve the shortcomings of existing technology, thereby promoting the further development of the technological wave of artificial intelligence. In this paper, we analyze the limitations of the Technology roadmap of the LLM, and propose solutions to these limitations, thus solving the inherent defects of Large Model. In this article, we will reveal how to implement True AGI step by step.
Peer Review Status:Awaiting Review
Hits
Hits
Downloads
Comment
YES