基于RoBERTa和T5的两阶段医学术语标准化

医学术语标准化作为消除实体歧义性的重要手段,被广泛应用于知识图谱的构建过程之中.针对医学领域涉及大量的专业术语和复杂的表述方式,传统匹配模型往往难以达到较高的准确率的问题,提出语义召回加精准排序的两阶段模型来提升医学术语标准化效果.首先在语义召回阶段基于改进的有监督对比学习和RoBERTa-wwm提出语义表征模型CL-BERT,通过CL-BERT生成实体的语义表征向量,根据向量之间的余弦相似度进行召回并得到标准词候选集,其次在精准排序阶段使用T5结合prompt tuning构建语义精准匹配模型,并将FGM对抗训练应用到模型训练中,然后使用精准匹配模型对原词和标准词候选集分别进行精准排序得到最终标准词.采用ccks2019公开数据集进行实验,F1值达到了0.9206,实验结果表明所提出的两阶段模型具有较高的性能,为实现医学术语标准化提供了新思路.

基于Prompt的两阶段澄清问题生成方法

在自然语言相关系统中,当用户输入存在歧义时,生成澄清问题询问用户有助于系统理解用户需求;基于Prompt的方法可以更好地挖掘预训练语言模型的潜在知识,但往往需要手动设计模板,限制其生成澄清问题的多样性。为解决这一问题,提出了TSCQG(two-stage clarification question generation)方法。首先,在动态Prompt模板生成阶段,利用歧义上下文和预训练语言模型生成动态的Prompt模板;然后在缺失信息生成阶段,将Prompt模板与外部知识相结合,充分利用预训练语言模型的生成能力生成相应的缺失信息。实验结果表明,在CLAQUA数据集的多轮对话情况中,BLEU值和ROUGE-L值分别达到了58.31和84.33,在ClariQ-FKw数据集上,BLEU值和ROUGE-L值分别达到了31.18和58.86。实验结果证明了TSCQG方法在澄清问题生成任务上的有效性。

基于完形填空的方面级情感四元组预测

方面情感四元组预测(ASQP)任务旨在从给定的评论语句中提取所有方面词以及相应的方面类别、观点表达和情感极性,有助于全面了解用户对产品或服务不同方面的评价情况.现有情感四元组预测方法主要存在以下局限:(1)判别式模型没有利用prompt捕获情感元素之间的语义关系;(2)生成式模型要么简单地将情感元素类型标签组合形成prompt,缺乏理解标签类型涵义的语境;要么将离散模板作为解码器的输入,而编码器则无法捕获到模板中情感元素之间的语义关系.为了缓解这些问题,本文首先基于完形填空思想研制离散和连续2类prompt,提供理解4个情感元素类型涵义的语境,帮助模型更好地捕获情感元素之间的语义关系;然后,基于设计的prompt,提出C-ASQP框架,包含判别式模型DC-ASQP和生成式模型GC-ASQP.在DC-ASQP中,采用2阶段策略,先预测4个情感元素中2个较为容易的情感元素,再将预测结果嵌入到设计的prompt中,帮助模型理解情感元素类型的涵义,从而有效预测另外2个情感元素.在GC-ASQP中,将设计的prompt作为编码器的输入,借助预训练模型的学习模式,充分利用预训练模型蕴含的知识提升四元组的生成效果.实验结果显示,DC-ASQP模型在4个常用数据集上的F1值相比同类判别式最优模型分别提高4.70%、6.48%、6.97%和2.60%,GC-ASQP模型的F1值比最优基准模型分别提高0.86%、1.67%、0.15%和1.02%,验证了将ASQP建模为完形填空任务的有效性,所设计的2类prompt以及C-ASQP框架是有效的.

Social Radars: Finding Targets in Cyberspace for Cybersecurity

INSPIRED by the insight from American political scientist Lasswell, who summarized the environmental role in societal surveillance [1], Schramm coined the term “social radar” [2] as it resembles the activities of radar in collecting and processing information, playing a crucial role in helping humans perceive changes in the internal and external environment and promptly adjusting adaptive behaviors.

Considerations for exerkine research focusing on the response to exercise training

García-Hermoso and colleagues1 recently published a systematic literature review and meta-analysis on exercise training-induced changes in exerkine concentrations in type 2diabetes mellitus patients,providing a contemporary view on how exerkines respond to exercise training.That review prompted us to highlight 2 additional considerations that should be taken into account when studying the response of exerkines to exercise training.Firstly,whether exerkines can exhibit discordant responses to acute exercise compared to exercise training,and secondly,the need to consider the residual effects of the most recent exercise bout.

Comprehension-driven design of advanced multi-block single-ion conducting polymer electrolytes for high-performance lithium-metal batteries

The continuously growing importance of batteries for powering(hybrid)electric vehicles and storing renewable energy has prompted a renewed focus on lithium-metal batteries(LMBs)in recent years,as its high theoretical specific capacity of about 3860 mA h g^(-1) and very low redox potential(-3.04 V vs.the standard hydrogen electrode)promise substantially higher energy densities compared to current lithium-ion batteries(LIBs)[1].However,lithium metal electrodes face severe challenges associated with the risk of dendritic lithium deposition and the high reactivity with traditional organic liquid electrolytes,resulting in a continuous loss of electrochemically active lithium and a relatively low Coulombic efficiency[2].To address these challenges,solid inorganic and polymer electrolytes have emerged as a potentially saferalternative.

Human radiological safety assessment for petawatt laser-driven ion acceleration experiments in CLAPA-T

The newly built Compact Laser Plasma Accelerator-Therapy facility at Peking University will deliver 60 J/1 Hz laser pulses with 30 fs duration.Driven by this petawatt laser facility,proton beams with energy up to 200 MeV are expected to be generated for tumor therapy.During high-repetition operation,both prompt radiation and residual radiation may cause safety problems.Therefore,human radiological safety assessment before commissioning is essential.In this paper,we simulate both prompt and residual radiation using the Geant4 and FLUKA Monte Carlo codes with reasonable proton and as-produced electron beam parameters.We find that the prompt radiation can be shielded well by the concrete wall of the experimental hall,but the risk from residual radiation is nonnegligible and necessitates adequate radiation cooling.On the basis of the simulation results,we discuss the constraints imposed by radiation safety considerations on the annual working time,and we propose radiation cooling strategies for different shooting modes.

Transactions of Nanjing University of Aeronautics and Astronautics Information for Contributors

Transactions of Nanjing University of Aeronautics&Astronautics (TNUAA) is a bimonthly journal facing international academic community.Emphasizing prompt and effective dissemination of key data and new scientific insights,TNUAA publishes experimental and theoretical papers bearing on applications to all branches of aeronautics, astronautics and civil aviation.

Perspectives on enhancing vaccination coverage and confidence in the Philippines

Vaccination plays a vital role in protecting against communicable infections like measles and their associated complications[1,2].While vaccination doesn't guarantee complete immunity,the authors also highlighted that vaccinated individuals tend to experience milder infections compared to those who are unvaccinated[1].Despite the availability of free vaccines for measles,polio,and pertussis through the national vaccination program[3],the Philippines faced outbreaks of measles and pertussis across its islands in 2024[4].The Bangsamoro Autonomous Region in Muslim Mindanao reported 500 measles cases,while pertussis outbreaks were recorded in various parts of Luzon and Visayas,prompting several cities to declare states of calamity[4,5].Additionally,despite successfully eradicating poliomyelitis in 2000,approximately 70 percent of provinces and cities remained at high risk of polio outbreaks[3,4].

Transactions of Nanjing University of Aeronautics and Astronautics Information for Contributors

Transactions of Nanjing University of Aeronautics&Astronautics (TNUAA) is a bimonthly journal facing international academic community.Emphasizing prompt and effective dissemination of key data and new scientific insights,TNUAA publishes experimental and theoretical papers bearing on applications to all branches of aeronautics, astronautics and civil aviation.

Transactions of Nanjing University of Aeronautics and Astronautics Information for Contributors

Transactions of Nanjing University of Aeronautics&Astronautics (TNUAA) is a bimonthly journal facing international academic community.Emphasizing prompt and effective dissemination of key data and new scientific insights,TNUAA publishes experimental and theoretical papers bearing on applications to all branches of aeronautics, astronautics and civil aviation.

Testicular dislocation:A rare complication of scrotal trauma

Scrotal trauma,though uncommon,can have severe consequences,and one such complication is testicular dislocation.This occurs when the testicle is displaced from its normal anatomical position within the scrotum due to a traumatic event.Understanding the mechanisms,diagnosis,and management of testicular dislocation is paramount in providing optimal care for affected individuals[1].Traumatic incidents,such as sports injuries,motor vehicle accidents,or direct blows to the scrotum,can result in testicular trauma[2].The condition demands immediate medical attention to prevent further complications,including vascular compromise and testicular ischemia.Diagnosing testicular dislocation involves a thorough physical examination and often necessitates imaging studies,such as ultrasound,to precisely determine the extent of the displacement and assess potential vascular compromise.Once diagnosed,prompt intervention is crucial to reposition the dislocated testicle and restore blood flow.Treatment options may include manual reduction or surgical intervention,depending on the severity of the dislocation and associated injuries[3].Close monitoring in the post-treatment period is essential to identify and manage any complications promptly.

基于ChatGLM的煤矿安全预警知识抽取及应用

煤炭安全预警可以预测和确定煤矿生产过程中的潜在风险,从而预防和减少煤矿事故的发生,减少事故造成的损失。构建煤矿安全灾害预警的本体模型,设计提示工程方法,利用ChatGLM大语言模型解决煤矿安全预警知识的自动提取问题,构建一个煤矿安全知识图知识库管理原型,验证了应用煤矿安全预警知识辅助决策的可行性。

China-Latin America Green Cooperation under the Global Development Initiative

The international development environment is entering a chaotic and volatile period.The world economy remains sluggish,while geopolitical conflicts,climate change,food crisis and a multitude of other challenges become more closely intertwined and mutually reinforcing.In this context,countries across the world have been prompted to look for healthy,feasible and sustainable solutions to keep their socioeconomic development momentum on track.

Security Vulnerability Analyses of Large Language Models (LLMs) through Extension of the Common Vulnerability Scoring System (CVSS) Framework

Large Language Models (LLMs) have revolutionized Generative Artificial Intelligence (GenAI) tasks, becoming an integral part of various applications in society, including text generation, translation, summarization, and more. However, their widespread usage emphasizes the critical need to enhance their security posture to ensure the integrity and reliability of their outputs and minimize harmful effects. Prompt injections and training data poisoning attacks are two of the most prominent vulnerabilities in LLMs, which could potentially lead to unpredictable and undesirable behaviors, such as biased outputs, misinformation propagation, and even malicious content generation. The Common Vulnerability Scoring System (CVSS) framework provides a standardized approach to capturing the principal characteristics of vulnerabilities, facilitating a deeper understanding of their severity within the security and AI communities. By extending the current CVSS framework, we generate scores for these vulnerabilities such that organizations can prioritize mitigation efforts, allocate resources effectively, and implement targeted security measures to defend against potential risks.

基于Word2Vec模型与RAG框架的医疗检索增强生成算法

当今通用人工智能(AGI)发展火热,各大语言模型(LLMs)层出不穷。大语言模型的广泛应用大大提高了人们的工作水平和效率,但大语言模型也并非完美的,同样伴随着诸多缺点。如:敏感数据安全性、幻觉性、时效性等。同时对于通用大语言模型来讲,对于一些专业领域问题的回答并不是很准确,这就需要检索增强生成(RAG)技术的支持。尤其是在智慧医疗领域方面,由于相关数据的缺乏,不能发挥出大语言模型优秀的对话和解决问题的能力。本算法通过使用Jieba分词,Word2Vec模型对文本数据进行词嵌入,计算句子间的向量相似度并做重排序,帮助大语言模型快速筛选出最可靠可信的模型外部的医疗知识数据,再根据编写相关的提示词(Prompt),可以使大语言模型针对医生或患者的问题提供令人满意的答案。Nowadays, general artificial intelligence is developing rapidly, and major language models are emerging one after another. The widespread application of large language models has greatly improved people’s work level and efficiency, but large language models are not perfect and are also accompanied by many shortcomings. Such as: data security, illusion, timeliness, etc. At the same time, for general large language models, the answers to questions in some professional fields are not very accurate, which requires the support of RAG technology. Especially in the field of smart medical care, due to the lack of relevant data, the excellent conversation and problem-solving capabilities of the large language model cannot be brought into play. This algorithm uses Jieba word segmentation and the Word2Vec model to embed text data, calculate the vector similarity between sentences and reorder them, helping the large language model to quickly screen out the most reliable and trustworthy medical knowledge data outside the model, and then write relevant prompts to enable the large language model to provide satisfactory answers to doctors or patients’ questions.

AIGC重建影视工业化构架

本文介绍了笔者所在单位通过AIGC技术进行影视工业流程的改造工作。从建立企业管理平台、确定研究方向、落地应用场景等几个主要角度阐述如何将AIGC技术成功应用到数字动画、数字展示和文旅、数字影视等多个业务形态中,并不断努力塑造基于AIGC的全新生产流程,展示了AIGC技术对于构建符合业务端的工业化架构和制作流程以及在商业实践中的应用价值,总结了现阶段AIGC技术的特性与企业未来探索方向。

Preventing bad behavior in academia

We gave young scientists this prompt:Describe one change to scientific policy or culture that would substantially decrease incidents of scientific misconduct or other unethical behavior.

通过多文档精排与融合的开放域问答任务增强实现

开放域问答(OpenQA)是自然语言处理中的一项具有挑战性的任务,传统的机器学习和深度学习技术通常用于从原始语料库中检索与问题相关的候选文档片段以进行答案提取。然而,当前方法检索的候选文档片段往往包含大量的噪声以及与问题无关的信息,并且主流的OpenQA模型在准确响应需要多个文档片段作为相关证据的问题方面存在不足。鉴于此,提出通过多文档精排与融合增强开放域问答的方法(RFMD),该方法在检索阶段设计了基于Transformer的文档精排模块,以减少候选文档中的噪声信息;在阅读理解阶段,RFMD采用以文本生成为中心的问答模块,通过构建跨文档片段的全局注意力机制,整合多个相关文档片段的信息,准确回答需要多个文档片段作为支持证据的问题。RFMD在NaturalQuestions和TriviaQA数据集上的EM得分分别达到45.8%和63.4%,验证了该模型在OpenQA任务中的有效性和优越性。

GUARDIAN: A Multi-Tiered Defense Architecture for Thwarting Prompt Injection Attacks on LLMs

This paper introduces a novel multi-tiered defense architecture to protect language models from adversarial prompt attacks. We construct adversarial prompts using strategies like role emulation and manipulative assistance to simulate real threats. We introduce a comprehensive, multi-tiered defense framework named GUARDIAN (Guardrails for Upholding Ethics in Language Models) comprising a system prompt filter, pre-processing filter leveraging a toxic classifier and ethical prompt generator, and pre-display filter using the model itself for output screening. Extensive testing on Meta’s Llama-2 model demonstrates the capability to block 100% of attack prompts. The approach also auto-suggests safer prompt alternatives, thereby bolstering language model security. Quantitatively evaluated defense layers and an ethical substitution mechanism represent key innovations to counter sophisticated attacks. The integrated methodology not only fortifies smaller LLMs against emerging cyber threats but also guides the broader application of LLMs in a secure and ethical manner.