设计前沿技术预测总体思路是开展前沿技术预测的重要基础和前提条件。文章分析了前沿技术预测主要作用和特点，研究了前沿技术预测概念和目标，提出了“识(Identity)—研(Research)—决(Decision)”前沿技术预测总体思路，详细描述了前沿技术识别、研判和决策的方法和步骤，分析了由构建技术树、确定重点和精准跟踪等组成的前沿技术识别初步思路，构建了基于“技术概率+技术影响”的前沿技术研判矩阵，提出了前沿技术决策的初步思路。研究结论可为科学实施前沿技术预测工作提供理论参考和借鉴。

依据前沿技术的前瞻性、先导性、探索性及颠覆性特征, 构建基于数据挖掘的前沿技术识别方法, 可为高效制定技术发展策略、合理调整技术规划布局提供支撑。首先, 以工程索引(engineering index, EI)和德温特创新(Derwent innovation, DI) 平台数据库为数据源, 制定检索表达式采集数据, 并对数据进行预处理。其次, 建立表征前沿技术特征的指标体系, 并对据此筛选出的技术文本数据进行挖掘, 识别前沿技术主题。最后, 以量子计算领域为例开展实证研究, 识别出量子纠错技术、光量子芯片技术等前沿技术点, 经技术专家等验证, 表明识别结果科学有效。

Based on the forward-looking, pioneering, exploratory, and disruptive characteristics of frontier technologies, a method of frontier technology identification based on data mining is constructed to provide support for efficient formulation of technology development strategies and rational adjustment of technology planning and layout. Firstly, data is collected and processed from the engineering index (EI) and Derwent innovation (DI) platform databases as a data source by formulating retrieval expressions. Secondly, the index system representing the characteristics of frontier technology is established, and the selected technical text data is mined to identify the frontier technology topics. Finally, empirical research is conducted in the field of quantum computing as an example, identifying frontier technology topics such as quantum error correction and optical quantum chips. This verification is confirmed by expert review and other methods, making identifying issues scientifically and practically applicable.

前沿技术对作战效能的影响评估在武器装备发展论证、技术体系决策建设等方面具有重要意义。从前沿技术发展决策的实际需求出发,开展前沿技术对作战效能评估方法研究。通过研究国内外技术评估流程及方法,建立前沿技术对军事应用价值的评估流程,设计技术关联模型与二次效应模型,提出基于指标体系与技术关联模型的前沿技术评估方法。在典型作战场景下,通过构建前沿技术评估指标体系、前沿技术对作战能力影响评估指标体系、大数据统计分析、专家经验校对打分,进行10项典型前沿技术对作战效能影响评估。评估结果表明,新设计的基于指标体系与技术关联模型的评估方法可以有效地实现对前沿技术作战效能影响的评估。

The evaluation of the influence of frontier technology on combat effectiveness is of great significance to the demonstration of weapon and equipment development and the decision-making in construction of technology system. The evaluation method for evaluating the effect of frontier technology in combat effectiveness is studied based on the actual demand of frontier technology development decision. By studying the technology evaluation process and methods at home and abroad, a military value evaluation process of frontier technology is established, a technology correlation model and a secondary effect model are designed, and a frontier technology evaluation method based on the index system and technology correlation model is proposed. The influences of ten typical frontier technologies on combat effectiveness in typical combat scenarios are evaluated by constructing an evaluation index system of frontier technologies, an evaluation index system for evaluating the influence of frontier technologies on combat capability, the big data statistical analysis, and the expert experience checking score. The evaluated results show that the proposed evaluation method based on index system and technology correlation model can effectively evaluate the influence of frontier technology on combat effectiveness.

人工智能技术在医疗领域的应用渗透正在加速，应用场景愈发丰富。基于智能医疗专利数据，利用LDA主题模型对智能医疗产业应用场景进行识别，以53581件智能医疗专利为源数据，挖掘得到智能医疗产业的12个应用场景主题，识别出智能诊疗、医疗机器人、药物研发、智能医学影像、智能健康管理五大主要应用场景，并从“技术—市场—政策”3个维度开展场景驱动视角下我国智能医疗产业演化过程分析。研究发现：①在技术方面，智能医疗技术处于不断创新和发展阶段，我国在各应用场景下均有研发布局；②在市场方面，国内外不同类型企业对智能医疗五大应用场景进行积极探索，我国智能医疗市场规模逐年扩大，前景广阔，但处于早期发展阶段；③在政策方面，我国尝试利用多种政策工具全方位支持智能医疗产业发展，但不同应用场景政策倾斜力度不同。研究结论可为我国智能医疗产业规划和政策制定提供数据参考与决策支撑。

【目的】分析基于大规模语言模型的提示学习方法在学术论文实体识别任务上的可用性。【方法】以ChatGPT这一大规模语言模型为例，将ChatGPT视为实体识别工具、伪标签生成工具以及训练数据生成工具，从性能、价格和时间等维度出发分析以上三个视角下ChatGPT的可用性。【结果】三个视角下基于ChatGPT的方法的F1值高于少量样本训练得到的神经网络基线模型，比如实体识别工具视角的F1宏平均值超过10个学术论文人工标注摘要训练得到的模型21.4个百分点。基于ChatGPT的方法在不同学科领域的学术论文数据集上性能较稳定。【局限】仅在英文学术论文摘要数据集上展开实验，但中文与英文学术论文、学术论文摘要与全文存在逻辑结构和表述上的差异。【结论】当缺少人工标注数据时，将ChatGPT视为实体识别工具可从学术论文摘要中识别出部分实体，但识别结果需进一步过滤以应用到下游任务中。

[Objective] This paper aims to use a large language model for entity recognition tasks of academic papers. [Methods] We utilized ChatGPT, a large language model, as an entity recognition tool, a pseudo-label generation tool, and a training set generation tool. Then, we analyzed ChatGPT’s performance, price, and time for the tasks. [Results] The F1 of the ChatGPT-based method in all three perspectives is higher than that of the neural network baseline model trained with a small dataset. For example, the F1 from the perspective of entity recognition was 21.4% higher than the model trained by manually annotating 10 abstracts. The ChatGPT-based methods had stable performance on academic paper datasets in different disciplines. [Limitations] We only examined the new method with English academic paper abstract datasets. More research is needed to examine it with the Chinese datasets. [Conclusions] ChatGPT can identify entities from academic paper abstracts with little manually annotated data. The recognition results need to be further filtered to be applied to downstream tasks.

立场检测旨在分析文本对给定目标的态度。当前社交媒体的文本通常简短且话题演变迅速，传统立场检测方法面临着语义信息稀少和立场特征表示不充分等挑战，且许多现有方法往往忽略了情感信息对立场检测的影响。为了应对上述两方面挑战，提出了一种基于大语言模型的社交媒体文本立场检测方法。通过设计包含明确任务指令的立场检测提示模板，调用模型在预训练阶段获得的与立场检测相关的知识，解决语义信息稀少的问题；通过在任务指令中加入情感分析任务，引导模型关注情感信息，为立场检测提供更多的语义线索，解决情感信息利用不足的问题。在此基础上，在模型内部添加针对立场检测任务的适配器，专注于提取和表示立场特征，增强模型对立场特征的表示能力，实现了更好的立场检测效果；将不同架构的大语言模型的结果进行集成投票提高预测结果的稳定性。为验证该方法，构建多组对比实验，实验结果表明该方法在SemEval-2016 Task 6A数据集上的有效性显著优于现有基准方法。

[目的/意义] 新兴技术识别有助于及时跟踪技术发展动态，摸清技术发展脉络，更早地捕捉领域未来的发展契机。但一方面新兴技术尚未形成统一的定义与特征属性；另一方面随着研究的深入与新技术的应用，新兴技术识别方法不断多样化与复杂化，因此，亟待对新兴技术定义属性与识别方法进行梳理与归纳。[方法/过程] 通过对国内外相关文献的调研，首先明晰了新兴技术的概念和属性特征，并对相近概念进行辨析。接下来对新兴技术识别方法作系统性梳理，重点介绍定量分析方法，并将其划分为科学计量分析法、文本挖掘分析法、机器学习方法3类，还揭示了新兴技术识别方法的新趋势，以及结果验证的常用方法。[结果/结论] 在对各类新兴技术识别方法的优劣势进行揭示的基础上，提出了现有各类方法的共性问题，即概念不明确，标准不统一；数据源类型选择不均衡，偏向性明显；“渐进式”回顾研究成主流，“跃进式”预测研究仍待发展，并针对每种问题提出了相应建议。

跨学科是一个复杂、多维的概念，单一的测度方法无法全面描述其本质特征。先前的研究大多针对跨学科某一维度进行分析，缺乏整体视角和系统建构。本研究对跨学科的外部知识融合、内在知识会聚与科学合作模式三个不同维度进行有效分解与整合，分别从参考文献学科多样性、目标文献学科多样性和合作机构学科多样性三个维度进行测度和综合比较分析，以普赖斯奖获得者的研究论文为例，探索不同测度方法的共性与差异、适应性与结合点，以及跨学科研究不同维度之间的内在关联。与单一维度的方法相比，三维测度可以更加全面地剖析研究对象的跨学科特征，同时也为跨学科的测度研究提供了新思路。

Interdisciplinarity is a complex and multidimensional concept, so it is difficult to fully describe its’ essential characteristics by a single indicator. Most previous studies have analyzed interdisciplinarity from single dimension, lacking an overall and systematic perspective. In this paper, we decompose, and then integrate the external knowledge integration, internal knowledge convergence and research activity patterns of interdisciplinarity from three different dimensions. Based on publication data of the winners of Derek de Solla Price, a comparative analysis of category integration, knowledge convergence and scientific cooperation are conducted, to explore the differences, adaptability, relations and possible combination of the three different measurement methods. Compared with the single indicator, our measurement from three dimensions can better profile the interdisciplinary characteristics of research object. Such a comprehensive dimension also provides a new perspective of evaluating interdisciplinary research.

颠覆性技术作为技术创新的重要组成部分，已成为推动新一轮技术变革浪潮的强力引擎。如何在激烈的市场竞争中，更早地发现和识别颠覆性技术，对企业和国家把握技术发展机遇具有重要意义。为此，本文在总结颠覆性技术特征的基础上，基于专利视角，从技术融合性、新颖性、扩张性和影响力四个维度，运用熵权法和模糊一致性矩阵方法，构建了一套系统的颠覆性技术识别模型。此外，为证实该模型的可行性和有效性，本文选取5个技术领域开展实证研究，分别从同领域传统技术与颠覆性技术对比、不同领域颠覆性技术对比、以及颠覆性技术在不同时间段的对比三个维度开展分析，一方面验证了本文模型的可行性，同时探究了各技术颠覆性潜力指数的有效性和适用性。

Disruptive technology, as an important part of technological innovation, has become a powerful engine driving of technological change. How to discover and identify disruptive technologies earlier, is of great significance, which is helpful for enterprises and countries to grasp the opportunities of technological development. For this reason, this paper proposes a technological disruptive potential identification model based on patent bibliometric. Firstly, we summarized the properties and characteristics of disruptive technology. Secondly, based on patent database, we built the quantitative index to characterize technical disruptive properties. The disruptive indexes include technical fusion index, technical novelty index, technical expansion index, and technical influence index. Thirdly, we used entropy method and fuzzy consistent matrix method to quantitative measure disruptive potential of technology. Finally, we choose five technologies to carry out empirical research from three aspects: 1) cooperation of traditional technology with disruptive technology in same field, 2) cooperation of disruptive technologies in different period, 3) cooperation of disruptive technologies in different field. In this way, the feasibility and effectiveness of the proposed method is confirmed. Meanwhile, the empirical research can further demonstrate the applicability of the disruptive potential index.

核心技术的识别是技术创新的重要环节，发现面向老年人的AAL监测领域核心技术，对预测该领域的技术发展趋势，提升应对老龄化社会技术需求能力有重要意义。已有研究聚集于分析技术交叉影响关系，基于技术交叉影响的核心技术识别尚未得到充分认识和应用。鉴于此，本文提出了基于技术交叉影响的核心技术识别模型，该模型以核心技术特征与技术交叉影响理论为基础。首先，阐述“核心技术”的概念，分析核心技术应具备的特性；其次，针对传统PCIA算法在核心技术识别中，难以直接套用条件概率计算技术交叉影响的不足，提出NPCIA算法测度技术交叉影响，并通过对核心技术已知的成熟技术分析，验证NPCIA算法的有效性，进而给出一种基于NPCIA算法的核心技术识别模型；最后，将提出的基于NPCIA的核心技术识别模型应用于面向老年人的AAL监测技术领域，验证所提核心技术识别模型的有效性

The identification of core technologies is an important part of technological innovation. Identifying the core technologies of AAL monitoring for the elderly plays a significant role in predicting the development trend of technology and improving the ability to cope with the technological needs of the aging society. The existing research focuses on analyzing the cross-impact relationship of techniques, and the core technology identification based on the cross-impact of technology has not been fully recognized and applied. In view of this, this paper puts forward a core technology recognition research framework based on cross-impact of technology, which is based on core technologies characters and technical cross-impact theory. Firstly, this paper describes the concept of "core technology", and analyzes the characteristics of core technologies; Secondly, in order to overcome the shortcomings of traditional PCIA algorithm in identifying core technology, which is difficult to directly apply conditional probability estimating the technological impact, we propose NPCIA algorithm to measure the cross-impact of technologies, and verify the effectiveness of NPCIA algorithm by analyzing the mature technology that the core technology already knows, and then put forward a core technology identification model based on NPCIA algorithm; Finally, the proposed model is applied to the AAL monitoring technology for the elderly and the effectiveness of the proposed core technology identification model is verified.

论文提出了一种基于集合论的技术应用领域转移研究方法。根据学科-领域对应关系，将文献归类到不同学科大类，然后将学科大类视为技术应用领域；再后定义技术应用领域文献的最大、最小集合及交叉集合，通过计算各集合内文献量随时间的变化，实现技术应用领域转移研究的目标。研究结论对于企业把握研发方向，有的放矢地配置研发资源起到决策支持的作用。文章以无线电力传输技术为例，对所提出的研究方法进行了说明。

We proposed a new research method about transfer of technology application domain based on Set Theory. According to the corresponding relation between disciplines and field, we categorized the literatures of certain technology into different application domains. First, We define the maximum set, minimum set and cross set of literatures in different technology domains, by calculating the amount of literatures changed over time in these sets, We realize the goal that researching transfer of technology application domain. Research results could help enterprises identify development directions, and allocate resources into research and development perfectly. We showed an example about technology of wireless power transfer and illustrated the research method we proposed.

从颠覆性技术成长周期入手，根据萌芽期颠覆性技术特点，采用基于创新性、独创性与功能分析的识别方法，在技术创新没有造成市场显著变化时实现预警决策。首先进行技术生命周期分析，随后利用创新性和独创性特点衡量技术颠覆性，排除渐进性技术干扰，最后运用功能分析方法研究技术新功能对未来市场的影响。以工业机器人专利技术领域进行实证，验证了方法的可行性和有效性。

最近，智能体代理能在复杂任务中提供高效的解决方案，在工业界备受关注。作为智能体代理的常见范式之一，检索增强生成（retrieval-augmented generation，RAG）旨在结合信息检索和内容生成技术增强生成响应质量，已逐步成为研究的重点。在对国内外检索增强生成方法研究的基础上，阐述了RAG的基本概念及工作流程，归纳了技术现状，分析了现有RAG技术的优缺点，梳理了现有评估指标、数据集和基准。最后探讨了RAG技术在未来应用场景下所面临的挑战，并展望了其未来发展方向。

Artificial intelligence agents provide efficient solutions in complex tasks, which have recently gained attention in industry. As one of the paradigms of artificial intelligence agents, retrieval-augmented generation (RAG), which aims to enhance the quality of generated responses by combining information retrieval and content generation techniques, has gradually become the focus of research. According to the studies on retrieval enhancement generation methods at home and abroad, the basic concept and workflow of RAG are elaborated, the current state of the technology is summarized, the advantages and disadvantages of the existing RAG technology are analyzed, and the existing evaluation indexes, datasets and benchmarks are sorted out. Finally, challenges faced by RAG technology in future application scenarios are discussed and the future development direction of RAG technology is envisioned.