arXiv - CS - Computers and Society最新文献_第5页

ChatGPT vs Social Surveys: Probing the Objective and Subjective Human Society ChatGPT 与社会调查：探究人类社会的客观与主观

arXiv - CS - Computers and Society Pub Date : 2024-09-04 DOI: arxiv-2409.02601

Muzhi Zhou, Lu Yu, Xiaomin Geng, Lan Luo

{"title":"ChatGPT vs Social Surveys: Probing the Objective and Subjective Human Society","authors":"Muzhi Zhou, Lu Yu, Xiaomin Geng, Lan Luo","doi":"arxiv-2409.02601","DOIUrl":"https://doi.org/arxiv-2409.02601","url":null,"abstract":"The extent to which Large Language Models (LLMs) can simulate the\u0000data-generating process for social surveys remains unclear. Current research\u0000has not thoroughly assessed potential biases in the sociodemographic population\u0000represented within the language model's framework. Additionally, the subjective\u0000worlds of LLMs often show inconsistencies in how closely their responses match\u0000those of groups of human respondents. In this paper, we used ChatGPT-3.5 to\u0000simulate the sampling process and generated six socioeconomic characteristics\u0000from the 2020 US population. We also analyzed responses to questions about\u0000income inequality and gender roles to explore GPT's subjective attitudes. By\u0000using repeated random sampling, we created a sampling distribution to identify\u0000the parameters of the GPT-generated population and compared these with Census\u0000data. Our findings show some alignment in gender and age means with the actual\u00002020 US population, but we also found mismatches in the distributions of racial\u0000and educational groups. Furthermore, there were significant differences between\u0000the distribution of GPT's responses and human self-reported attitudes. While\u0000the overall point estimates of GPT's income attitudinal responses seem to align\u0000with the mean of the population occasionally, their response distributions\u0000follow a normal distribution that diverges from human responses. In terms of\u0000gender relations, GPT's answers tend to cluster in the most frequently answered\u0000category, demonstrating a deterministic pattern. We conclude by emphasizing the\u0000distinct design philosophies of LLMs and social surveys: LLMs aim to predict\u0000the most suitable answers, while social surveys seek to reveal the\u0000heterogeneity among social groups.","PeriodicalId":501112,"journal":{"name":"arXiv - CS - Computers and Society","volume":"35 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142227768","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Governing dual-use technologies: Case studies of international security agreements and lessons for AI governance 管理双重用途技术：国际安全协议案例研究及人工智能治理的经验教训

arXiv - CS - Computers and Society Pub Date : 2024-09-04 DOI: arxiv-2409.02779

Akash R. Wasil, Peter Barnett, Michael Gerovitch, Roman Hauksson, Tom Reed, Jack William Miller

引用次数: 0

METcross: A framework for short-term forecasting of cross-city metro passenger flow METcross：跨城市地铁客流短期预测框架

arXiv - CS - Computers and Society Pub Date : 2024-09-03 DOI: arxiv-2409.01515

Wenbo Lu, Jinhua Xu, Peikun Li, Ting Wang, Yong Zhang

{"title":"METcross: A framework for short-term forecasting of cross-city metro passenger flow","authors":"Wenbo Lu, Jinhua Xu, Peikun Li, Ting Wang, Yong Zhang","doi":"arxiv-2409.01515","DOIUrl":"https://doi.org/arxiv-2409.01515","url":null,"abstract":"Metro operation management relies on accurate predictions of passenger flow\u0000in the future. This study begins by integrating cross-city (including source\u0000and target city) knowledge and developing a short-term passenger flow\u0000prediction framework (METcross) for the metro. Firstly, we propose a basic\u0000framework for modeling cross-city metro passenger flow prediction from the\u0000perspectives of data fusion and transfer learning. Secondly, METcross framework\u0000is designed to use both static and dynamic covariates as inputs, including\u0000economy and weather, that help characterize station passenger flow features.\u0000This framework consists of two steps: pre-training on the source city and\u0000fine-tuning on the target city. During pre-training, data from the source city\u0000trains the feature extraction and passenger flow prediction models. Fine-tuning\u0000on the target city involves using the source city's trained model as the\u0000initial parameter and fusing the feature embeddings of both cities to obtain\u0000the passenger flow prediction results. Finally, we tested the basic prediction\u0000framework and METcross framework on the metro networks of Wuxi and Chongqing to\u0000experimentally analyze their efficacy. Results indicate that the METcross\u0000framework performs better than the basic framework and can reduce the Mean\u0000Absolute Error and Root Mean Squared Error by 22.35% and 26.18%, respectively,\u0000compared to single-city prediction models.","PeriodicalId":501112,"journal":{"name":"arXiv - CS - Computers and Society","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183897","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A+AI: Threats to Society, Remedies, and Governance A+AI：对社会的威胁、补救措施和治理

arXiv - CS - Computers and Society Pub Date : 2024-09-03 DOI: arxiv-2409.02219

Don Byrd

{"title":"A+AI: Threats to Society, Remedies, and Governance","authors":"Don Byrd","doi":"arxiv-2409.02219","DOIUrl":"https://doi.org/arxiv-2409.02219","url":null,"abstract":"This document focuses on the threats, especially near-term threats, that\u0000Artificial Intelligence (AI) brings to society. Most of the threats discussed\u0000here can result from any algorithmic process, not just AI; in addition,\u0000defining AI is notoriously difficult. For both reasons, it is important to\u0000think of \"A+AI\": Algorithms and Artificial Intelligence. In addition to the threats, this paper discusses countermeasures to them, and\u0000it includes a table showing which countermeasures are likely to mitigate which\u0000threats. Thoughtful governance could manage the risks without seriously\u0000impeding progress; in fact, chances are it would accelerate progress by\u0000reducing the social chaos that would otherwise be likely. The paper lists\u0000specific actions government should take as soon as possible, namely: * Require all social media platforms accessible in the U.S. to offer users\u0000verification that their accounts are owned by citizens, and to display every\u0000account's verification status * Establish regulations to require that all products created or significantly\u0000modified with A+AI be clearly labeled as such; to restrict use of generative AI\u0000to create likenesses of persons; and to require creators of generative AI\u0000software to disclose materials used to train their software and to compensate\u0000the creators of any copyrighted material used * Fund a crash project of research on mitigating the threats * Fund educational campaigns to raise awareness of the threats","PeriodicalId":501112,"journal":{"name":"arXiv - CS - Computers and Society","volume":"23 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183894","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Observing Context Improves Disparity Estimation when Race is Unobserved 当种族无法观测时，观测背景可改进差异估计

arXiv - CS - Computers and Society Pub Date : 2024-09-03 DOI: arxiv-2409.01984

Kweku Kwegyir-Aggrey, Naveen Durvasula, Jennifer Wang, Suresh Venkatasubramanian

引用次数: 0

Comprehensive up-to-date impact of the IoMT in healthcare and patients 全面介绍 IoMT 对医疗保健和患者的最新影响

arXiv - CS - Computers and Society Pub Date : 2024-09-02 DOI: arxiv-2409.01287

Guy. Mouanda

引用次数: 0

Lecture Notes from the NaijaCoder Summer Camp 来自 NaijaCoder 夏令营的讲座笔记

arXiv - CS - Computers and Society Pub Date : 2024-09-02 DOI: arxiv-2409.01499

Daniel Alabi, Joseph Ekpenyong, Alida Monaco

引用次数: 0

Agentic Society: Merging skeleton from real world and texture from Large Language Model 代理社会：融合现实世界的骨架和大型语言模型的纹理

arXiv - CS - Computers and Society Pub Date : 2024-09-02 DOI: arxiv-2409.10550

Yuqi Bai, Kun Sun, Huishi Yin

{"title":"Agentic Society: Merging skeleton from real world and texture from Large Language Model","authors":"Yuqi Bai, Kun Sun, Huishi Yin","doi":"arxiv-2409.10550","DOIUrl":"https://doi.org/arxiv-2409.10550","url":null,"abstract":"Recent advancements in large language models (LLMs) and agent technologies\u0000offer promising solutions to the simulation of social science experiments, but\u0000the availability of data of real-world population required by many of them\u0000still poses as a major challenge. This paper explores a novel framework that\u0000leverages census data and LLMs to generate virtual populations, significantly\u0000reducing resource requirements and bypassing privacy compliance issues\u0000associated with real-world data, while keeping a statistical truthfulness.\u0000Drawing on real-world census data, our approach first generates a persona that\u0000reflects demographic characteristics of the population. We then employ LLMs to\u0000enrich these personas with intricate details, using techniques akin to those in\u0000image generative models but applied to textual data. Additionally, we propose a\u0000framework for the evaluation of the feasibility of our method with respect to\u0000capability of LLMs based on personality trait tests, specifically the Big Five\u0000model, which also enhances the depth and realism of the generated personas.\u0000Through preliminary experiments and analysis, we demonstrate that our method\u0000produces personas with variability essential for simulating diverse human\u0000behaviors in social science experiments. But the evaluation result shows that\u0000only weak sign of statistical truthfulness can be produced due to limited\u0000capability of current LLMs. Insights from our study also highlight the tension\u0000within LLMs between aligning with human values and reflecting real-world\u0000complexities. Thorough and rigorous test call for further research. Our codes\u0000are released at https://github.com/baiyuqi/agentic-society.git","PeriodicalId":501112,"journal":{"name":"arXiv - CS - Computers and Society","volume":"8 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142263705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Human-Centered AI Applications for Canada's Immigration Settlement Sector 加拿大移民安置部门以人为本的人工智能应用

arXiv - CS - Computers and Society Pub Date : 2024-09-02 DOI: arxiv-2409.01461

Isar Nejadgholi, Maryam Molamohammadi, Kimiya Missaghi, Samir Bakhtawar

引用次数: 0

AI Literacy for All: Adjustable Interdisciplinary Socio-technical Curriculum 全民人工智能扫盲：可调整的跨学科社会技术课程

arXiv - CS - Computers and Society Pub Date : 2024-09-02 DOI: arxiv-2409.10552

Sri Yash Tadimalla, Mary Lou Maher

{"title":"AI Literacy for All: Adjustable Interdisciplinary Socio-technical Curriculum","authors":"Sri Yash Tadimalla, Mary Lou Maher","doi":"arxiv-2409.10552","DOIUrl":"https://doi.org/arxiv-2409.10552","url":null,"abstract":"This paper presents a curriculum, \"AI Literacy for All,\" to promote an\u0000interdisciplinary understanding of AI, its socio-technical implications, and\u0000its practical applications for all levels of education. With the rapid\u0000evolution of artificial intelligence (AI), there is a need for AI literacy that\u0000goes beyond the traditional AI education curriculum. AI literacy has been\u0000conceptualized in various ways, including public literacy, competency building\u0000for designers, conceptual understanding of AI concepts, and domain-specific\u0000upskilling. Most of these conceptualizations were established before the public\u0000release of Generative AI (Gen-AI) tools like ChatGPT. AI education has focused\u0000on the principles and applications of AI through a technical lens that\u0000emphasizes the mastery of AI principles, the mathematical foundations\u0000underlying these technologies, and the programming and mathematical skills\u0000necessary to implement AI solutions. In AI Literacy for All, we emphasize a\u0000balanced curriculum that includes technical and non-technical learning outcomes\u0000to enable a conceptual understanding and critical evaluation of AI technologies\u0000in an interdisciplinary socio-technical context. The paper presents four\u0000pillars of AI literacy: understanding the scope and technical dimensions of AI,\u0000learning how to interact with Gen-AI in an informed and responsible way, the\u0000socio-technical issues of ethical and responsible AI, and the social and future\u0000implications of AI. While it is important to include all learning outcomes for\u0000AI education in a Computer Science major, the learning outcomes can be adjusted\u0000for other learning contexts, including, non-CS majors, high school summer\u0000camps, the adult workforce, and the public. This paper advocates for a shift in\u0000AI literacy education to offer a more interdisciplinary socio-technical\u0000approach as a pathway to broaden participation in AI.","PeriodicalId":501112,"journal":{"name":"arXiv - CS - Computers and Society","volume":"12 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142263698","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0