Aprajit Mahajan, Shekhar Mittal, Ofir Reich, Taha Barwahwala
We investigate the use of a machine learning (ML) algorithm to identify fraudulent non-existent firms that are used for tax evasion. Using a rich dataset of tax returns in an Indian state over several years, we train an ML-based model to predict fraudulent firms. We then use the model predictions to carry out field inspections of firms identified as suspicious by the ML tool. We find that the ML model is accurate in both simulated and field settings in identifying non-existent firms. Withholding a randomly selected group of firms from inspection, we estimate the causal impact of ML driven inspections. Despite the strong predictive performance, our model driven inspections do not yield a significant increase in enforcement as evidenced by the cancellation of fraudulent firm registrations and tax recovery. We provide two explanations for this discrepancy based on a close analysis of the tax department’s operating protocols: overfitting to proxy-labels, and institutional friction in integrating the model into existing administrative systems. Our study serves as a cautionary tale for the application of machine learning in public policy contexts and of relying solely on test set performance as an effectiveness indicator. Field evaluations are critical in assessing the real-world impact of predictive models.
我们研究了如何利用机器学习（ML）算法来识别用于逃税的不存在的欺诈性公司。利用印度某邦数年来丰富的纳税申报数据集，我们训练了一个基于 ML 的模型来预测欺诈性公司。然后，我们利用模型预测结果，对 ML 工具识别出的可疑公司进行实地检查。我们发现，无论是在模拟环境中还是在实地环境中，ML 模型都能准确识别不存在的公司。在不对随机抽取的一组企业进行检查的情况下，我们估算了 ML 驱动检查的因果影响。尽管具有很强的预测性能，但我们的模型驱动检查并没有显著提高执法力度，虚假企业注册的注销和税款的追缴都证明了这一点。基于对税务部门操作规程的仔细分析，我们对这一差异提供了两种解释：对代理标签的过度拟合，以及将模型整合到现有管理系统中的制度摩擦。我们的研究为机器学习在公共政策环境中的应用以及单纯依赖测试集性能作为有效性指标提供了警示。实地评估对于评估预测模型在现实世界中的影响至关重要。
{"title":"Using Machine Learning to Catch Bogus Firms","authors":"Aprajit Mahajan, Shekhar Mittal, Ofir Reich, Taha Barwahwala","doi":"10.1145/3676188","DOIUrl":"https://doi.org/10.1145/3676188","url":null,"abstract":"We investigate the use of a machine learning (ML) algorithm to identify fraudulent non-existent firms that are used for tax evasion. Using a rich dataset of tax returns in an Indian state over several years, we train an ML-based model to predict fraudulent firms. We then use the model predictions to carry out field inspections of firms identified as suspicious by the ML tool. We find that the ML model is accurate in both simulated and field settings in identifying non-existent firms. Withholding a randomly selected group of firms from inspection, we estimate the causal impact of ML driven inspections. Despite the strong predictive performance, our model driven inspections do not yield a significant increase in enforcement as evidenced by the cancellation of fraudulent firm registrations and tax recovery. We provide two explanations for this discrepancy based on a close analysis of the tax department’s operating protocols: overfitting to proxy-labels, and institutional friction in integrating the model into existing administrative systems. Our study serves as a cautionary tale for the application of machine learning in public policy contexts and of relying solely on test set performance as an effectiveness indicator. Field evaluations are critical in assessing the real-world impact of predictive models.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141811225","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}
Although internet access and affordability are increasingly at the center of policy decisions around issues of the “digital divide” in the US, the complex nature of usage as it relates to structural inequality is not well-understood. We partnered with Project Waves, a community internet provider, to set up connectivity across the urban landscape of a city in the Eastern United States to study factors that impact the rollout of affordable broadband internet connectivity to low-income communities during the COVID-19 pandemic. The organization endeavored to meet structural challenges, provide community support for adoption, and stave off attendant privacy concerns. We present three dimensions of equitable use prioritized by the community internet provider: safety from COVID-19 through social distancing enabled by remote access, trusted connectivity, and private internet access. We use employee interviews and a phone survey of internet recipients to investigate how the provider prioritized these dimensions and who uses their service.
{"title":"Connecting in Crisis: Investigating Equitable Community Internet Access in the US During the COVID-19 Pandemic","authors":"Nora Mcdonald, Lydia Stamato, Foad Hamidi","doi":"10.1145/3677326","DOIUrl":"https://doi.org/10.1145/3677326","url":null,"abstract":"Although internet access and affordability are increasingly at the center of policy decisions around issues of the “digital divide” in the US, the complex nature of usage as it relates to structural inequality is not well-understood. We partnered with Project Waves, a community internet provider, to set up connectivity across the urban landscape of a city in the Eastern United States to study factors that impact the rollout of affordable broadband internet connectivity to low-income communities during the COVID-19 pandemic. The organization endeavored to meet structural challenges, provide community support for adoption, and stave off attendant privacy concerns. We present three dimensions of equitable use prioritized by the community internet provider: safety from COVID-19 through social distancing enabled by remote access, trusted connectivity, and private internet access. We use employee interviews and a phone survey of internet recipients to investigate how the provider prioritized these dimensions and who uses their service.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141655821","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}
The rising ubiquity of smartphones for navigation, driver mode, etc., has increased their use significantly among drivers; however, there are growing numbers of road fatalities being reported due to distractions from the phone while driving. In contrast to the existing solutions that use a camera or other communication media on the car or need external setups, this paper proposes a solution called ZeCA , where the smartphone itself can identify in real-time with zero pre-configurations whether its user is driving while engaging in a high-distraction interaction with the phone. ZeCA runs as a smartphone background service and generates audio-visual alerts when the phone can distract the driver. A thorough evaluation and usability study of ZeCA with 50 different models of vehicles driven by 70 drivers over 5 countries indicates that the proposed solution can infer distracting smartphone interactions with (gt 80% ) accuracy and a (70% ) reduction in smartphone usage during driving.
{"title":"Zero-configuration Alarms: Towards Reducing Distracting Smartphone Interactions while Driving","authors":"Sugandh Pargal, Neha Dalmia, Harshal R. Borse, Bivas Mitra, Sandip Chakraborty","doi":"10.1145/3675159","DOIUrl":"https://doi.org/10.1145/3675159","url":null,"abstract":"\u0000 The rising ubiquity of smartphones for navigation, driver mode, etc., has increased their use significantly among drivers; however, there are growing numbers of road fatalities being reported due to distractions from the phone while driving. In contrast to the existing solutions that use a camera or other communication media on the car or need external setups, this paper proposes a solution called\u0000 ZeCA\u0000 , where the smartphone itself can identify in real-time with zero pre-configurations whether its user is driving while engaging in a high-distraction interaction with the phone.\u0000 ZeCA\u0000 runs as a smartphone background service and generates audio-visual alerts when the phone can distract the driver. A thorough evaluation and usability study of\u0000 ZeCA\u0000 with 50 different models of vehicles driven by 70 drivers over 5 countries indicates that the proposed solution can infer distracting smartphone interactions with\u0000 \u0000 (gt 80% )\u0000 \u0000 accuracy and a\u0000 \u0000 (70% )\u0000 \u0000 reduction in smartphone usage during driving.\u0000","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141656587","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}
Earning a living often leaves low-income individuals with little time for learning new skills, perpetuating a cycle where the need for immediate income restricts access to learning. In this study, we investigate if digital work, specifically speech data generation, can facilitate domain-specific knowledge acquisition. For the purposes of this study we focus on finance and banking. We conducted a two-week financial literacy program with low-income individuals (n=55) in Wagholi, a semi-urban area in Pune, India. Participants read aloud and recorded a nine-lesson financial curriculum, earning ₹2000 (≈ $24) for ≈ 90 minutes of voice-recording. By conducting pre- and post-tests, we found a significant increase in participants’ financial knowledge with a high effect size (cohen’s d = 1.32) and medium normalised score gain (hake’s g = 0.58). Fourteen follow-up interviews indicated the work was accessible and conveniently integrated into participants’ daily lives. Additionally, the program triggered attitude change among participants and community dialogue about critical financial concepts. Our results suggest that digital work can become an effective method for knowledge acquisition and should be tested at a larger scale.
为了生计，低收入者往往没有时间学习新技能，这就形成了一个恶性循环，即急需收入限制了学习机会。在本研究中，我们探讨了数字工作（特别是语音数据生成）能否促进特定领域知识的获取。在本研究中，我们重点关注金融和银行业。我们在印度浦那的一个半城市地区--瓦格霍利（Wagholi）为低收入者（人数=55）开展了一项为期两周的金融扫盲计划。参与者朗读并录制了九节金融课程，90 分钟的录音可赚取 2000 英镑（约合 24 美元）。通过进行前测和后测，我们发现参与者的金融知识有了显著提高，效果大小较高（cohen's d = 1.32），归一化得分收益中等（hake's g = 0.58）。14 次后续访谈表明，这项工作可以方便地融入参与者的日常生活。此外，该计划还引发了参与者的态度转变和有关重要财务概念的社区对话。我们的研究结果表明，数字作品可以成为获取知识的有效方法，并应在更大范围内进行测试。
{"title":"Speaking in Terms of Money: Financial Knowledge Acquisition via Speech Data Generation","authors":"Advait Bhat, Nidhi Kulkarni, Safiya Husain, Aditya Yadavalli, Jivat Kaur, Anurag Shukla, Monali Shelar, Vivek Seshadri","doi":"10.1145/3663775","DOIUrl":"https://doi.org/10.1145/3663775","url":null,"abstract":"Earning a living often leaves low-income individuals with little time for learning new skills, perpetuating a cycle where the need for immediate income restricts access to learning. In this study, we investigate if digital work, specifically speech data generation, can facilitate domain-specific knowledge acquisition. For the purposes of this study we focus on finance and banking. We conducted a two-week financial literacy program with low-income individuals (n=55) in Wagholi, a semi-urban area in Pune, India. Participants read aloud and recorded a nine-lesson financial curriculum, earning ₹2000 (≈ $24) for ≈ 90 minutes of voice-recording. By conducting pre- and post-tests, we found a significant increase in participants’ financial knowledge with a high effect size (cohen’s d = 1.32) and medium normalised score gain (hake’s g = 0.58). Fourteen follow-up interviews indicated the work was accessible and conveniently integrated into participants’ daily lives. Additionally, the program triggered attitude change among participants and community dialogue about critical financial concepts. Our results suggest that digital work can become an effective method for knowledge acquisition and should be tested at a larger scale.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141676129","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}
The economic costs of Internet shutdowns are far-reaching and widespread, and span beyond the simple disruption to communication networks that are reliant on access to the Internet. Existing work on the impacts of the Internet shutdowns does not extensively exploit the fact that they can have adverse effects on the local economy in terms of output, employment, and investments. There is a lack of rigorous economic analysis of the impacts of shutdowns that can be more broadly applied to specific regions that account for variations in the intensity (or type) of shutdowns, as well as go beyond providing broad GDP cost estimates which may be misleading. This paper aims to bridge this gap by providing an econometric approach to estimate the impact of Internet shutdowns on GDP, employment, and foreign direct investment using panel data on 92 countries. We show that a point increase in the likelihood of an Internet shutdown was statistically significantly associated with a 15.6 percentage point reduction in the GDP per capita on average and every additional day of an Internet shutdown costs $86.58 per person on average.
互联网关闭造成的经济损失是深远而广泛的，不仅仅是对依赖于互联网接入的通信网络的简单破坏。现有关于互联网关闭影响的研究并没有广泛利用互联网关闭会在产出、就业和投资方面对当地经济产生不利影响这一事实。目前缺乏对互联网关闭影响的严谨经济分析，这些分析可以更广泛地应用于特定地区，并考虑到关闭强度（或类型）的变化，以及提供可能具有误导性的广义 GDP 成本估算。本文旨在利用 92 个国家的面板数据，提供一种计量经济学方法来估算互联网关闭对国内生产总值、就业和外国直接投资的影响，从而弥补这一差距。我们的研究表明，在统计上，互联网关闭的可能性每增加一个点，人均 GDP 就会平均减少 15.6 个百分点，而互联网每多关闭一天，人均成本就会增加 86.58 美元。
{"title":"Net Loss: An econometric method to measure the impact of Internet shutdowns","authors":"A. Tagat, Amreesh Phokeer, Hanna Kreitem","doi":"10.1145/3659466","DOIUrl":"https://doi.org/10.1145/3659466","url":null,"abstract":"The economic costs of Internet shutdowns are far-reaching and widespread, and span beyond the simple disruption to communication networks that are reliant on access to the Internet. Existing work on the impacts of the Internet shutdowns does not extensively exploit the fact that they can have adverse effects on the local economy in terms of output, employment, and investments. There is a lack of rigorous economic analysis of the impacts of shutdowns that can be more broadly applied to specific regions that account for variations in the intensity (or type) of shutdowns, as well as go beyond providing broad GDP cost estimates which may be misleading. This paper aims to bridge this gap by providing an econometric approach to estimate the impact of Internet shutdowns on GDP, employment, and foreign direct investment using panel data on 92 countries. We show that a point increase in the likelihood of an Internet shutdown was statistically significantly associated with a 15.6 percentage point reduction in the GDP per capita on average and every additional day of an Internet shutdown costs $86.58 per person on average.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140691762","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}
Early detection and treatment of Social Anxiety Disorder (SAD) is crucial. However, current diagnostic methods have several drawbacks, including being time-consuming for clinical interviews, susceptible to emotional bias for self-reports, and inconclusive for physiological measures. Our research focuses on a digital approach using acoustic and linguistic features extracted from participants’ “speech” for diagnosing SAD. Our methodology involves identifying correlations between extracted features and SAD severity, selecting the effective features, and comparing classical machine learning and deep learning methods for predicting SAD. Our results demonstrate that both acoustic and linguistic features outperform deep learning approaches when considered individually. Logistic Regression proves effective for acoustic features, while Random Forest excels with linguistic features, achieving the highest accuracy of 85.71%. Our findings pave the way for non-intrusive SAD diagnosing that can be used conveniently anywhere, facilitating early detection.
及早发现和治疗社交焦虑症（SAD）至关重要。然而，目前的诊断方法有几个缺点，包括临床访谈耗时长，自我报告易受情绪偏差影响，生理测量不确定。我们的研究重点是利用从参与者 "讲话 "中提取的声学和语言特征来诊断 SAD 的数字化方法。我们的方法包括识别所提取特征与 SAD 严重程度之间的相关性，选择有效的特征，并比较经典的机器学习和深度学习方法来预测 SAD。我们的研究结果表明，如果单独考虑深度学习方法，声学特征和语言特征都优于深度学习方法。逻辑回归证明了声学特征的有效性，而随机森林则在语言特征方面表现出色，达到了 85.71% 的最高准确率。我们的研究结果为非侵入式 SAD 诊断铺平了道路，它可以方便地在任何地方使用，从而促进早期检测。
{"title":"Unveiling Social Anxiety: Analyzing Acoustic and Linguistic Traits in Impromptu Speech within a Controlled Study","authors":"N. K. Sahu, Manjeet Yadav, H. Lone","doi":"10.1145/3657245","DOIUrl":"https://doi.org/10.1145/3657245","url":null,"abstract":"Early detection and treatment of Social Anxiety Disorder (SAD) is crucial. However, current diagnostic methods have several drawbacks, including being time-consuming for clinical interviews, susceptible to emotional bias for self-reports, and inconclusive for physiological measures. Our research focuses on a digital approach using acoustic and linguistic features extracted from participants’ “speech” for diagnosing SAD. Our methodology involves identifying correlations between extracted features and SAD severity, selecting the effective features, and comparing classical machine learning and deep learning methods for predicting SAD. Our results demonstrate that both acoustic and linguistic features outperform deep learning approaches when considered individually. Logistic Regression proves effective for acoustic features, while Random Forest excels with linguistic features, achieving the highest accuracy of 85.71%. Our findings pave the way for non-intrusive SAD diagnosing that can be used conveniently anywhere, facilitating early detection.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-04-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140711278","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}
Anik Sinha, Nova Ahmed, Sabbir Ahmed, Ifti Azad Abeer, Rahat Jahangir Rony, Anik Saha, Syeda Shabnam Khan, Shajnush Amir, Shabana Khan
The COVID-19 pandemic required handling a clear communication of risk and community engagement. A gap is noted in scholarly studies portraying strong community engagement for risk handling, particularly in resource constrained regions in HCI community. This study covers community engagement and its use of technology during COVID-19 through a qualitative study of Bangladesh. The study looks at marginalized communities who have struggled through the pandemic yet handled the difficult time through their effective problem solving, working together as a community when there was not enough support from authorities. It is a qualitative study during the pandemic consisting of 9 communities, presenting 58 participants (N=58, Female= 33, Male=23, Transgender =2) across four divisions of Bangladesh covering urban, semi urban, and rural regions. The study uncovers the challenges and close community structures. It also shows the enhanced and increased positive role of technology during the pandemic while referring to a few communities being digitally disconnected communities that could benefit from digital connectivity in the future through increased awareness and support.
{"title":"Roles of Technology for Risk Communication and Community Engagement in Bangladesh during COVID-19 Pandemic","authors":"Anik Sinha, Nova Ahmed, Sabbir Ahmed, Ifti Azad Abeer, Rahat Jahangir Rony, Anik Saha, Syeda Shabnam Khan, Shajnush Amir, Shabana Khan","doi":"10.1145/3648433","DOIUrl":"https://doi.org/10.1145/3648433","url":null,"abstract":"The COVID-19 pandemic required handling a clear communication of risk and community engagement. A gap is noted in scholarly studies portraying strong community engagement for risk handling, particularly in resource constrained regions in HCI community. This study covers community engagement and its use of technology during COVID-19 through a qualitative study of Bangladesh. The study looks at marginalized communities who have struggled through the pandemic yet handled the difficult time through their effective problem solving, working together as a community when there was not enough support from authorities. It is a qualitative study during the pandemic consisting of 9 communities, presenting 58 participants (N=58, Female= 33, Male=23, Transgender =2) across four divisions of Bangladesh covering urban, semi urban, and rural regions. The study uncovers the challenges and close community structures. It also shows the enhanced and increased positive role of technology during the pandemic while referring to a few communities being digitally disconnected communities that could benefit from digital connectivity in the future through increased awareness and support.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140447442","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}
A key question in e-participation is what roles public officials can play to harness the benefits of emerging technologies and practices, mitigate potential harms and, ultimately, ensure more inclusive and effective public involvement in decision-making. This paper presents results from a desktop analysis of e-participation projects from the African continent to highlight the diversity of public official roles and associated skills and perspectives that would be relevant to e-participation implementation. The identified roles and activities range from legal specialists developing guidelines to comply with personal data protection legislation, and stakeholder managers designing models of collaboration with commons-based platforms; to communications officials learning how to moderate social media conversations, and technology developers exploring new ways of verifying online identity.
{"title":"Implementing e-participation in Africa: What Roles can Public Officials Play?","authors":"P. Plantinga, N. Dlamini, Tanja Gordon","doi":"10.1145/3648438","DOIUrl":"https://doi.org/10.1145/3648438","url":null,"abstract":"A key question in e-participation is what roles public officials can play to harness the benefits of emerging technologies and practices, mitigate potential harms and, ultimately, ensure more inclusive and effective public involvement in decision-making. This paper presents results from a desktop analysis of e-participation projects from the African continent to highlight the diversity of public official roles and associated skills and perspectives that would be relevant to e-participation implementation. The identified roles and activities range from legal specialists developing guidelines to comply with personal data protection legislation, and stakeholder managers designing models of collaboration with commons-based platforms; to communications officials learning how to moderate social media conversations, and technology developers exploring new ways of verifying online identity.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139958468","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}
Developing countries need to better manage fast increasing traffic flows, owing to rapid urbanization. Else, increasing traffic congestion would increase fatalities due to reckless driving, as well as keep vehicular emissions and air pollution critically high in cities like New Delhi. State-of-the-art traffic signal control methods in developed countries, however, use expensive sensing, computation and communication resources. How far can control algorithms go, under resource constraints, is explored through the design and evaluation of FrugalLight (FL) in this paper. We also captured and processed a real traffic dataset at a busy intersection in New Delhi, India, using efficient techniques on low cost embedded devices. This dataset ( https://delhi-trafficdensity-dataset.github.io ) contains traffic density information at fine time granularity of one measurement every second, from all approaches of the intersection for 40 days. FrugalLight ( https://github.com/sachin-iitd/FrugalLight ) is evaluated on the collected traffic dataset from New Delhi and another open source traffic dataset from New York. FrugalLight matches the performance of state-of-the-art Convolutional Neural Network (CNN) based sensing and Deep Reinforcement Learning (DRL) based control algorithms, while utilizing resources less by an order of magnitude. We further explore improvements using a careful combination of knowledge distillation and domain knowledge based DRL model compression, with employing Model-Agnostic Meta-Learning to quickly adapt to traffic at new intersections. The collected real dataset and FrugalLight therefore opens up opportunities for resource efficient RL based intersection control design for the ML research community, where the controller should have limited carbon footprint. Such intelligent, green, intersection controllers can help reduce traffic congestion and associated vehicular emissions, even if compute and communication infrastructure is limited in low resource regions. This is a critical step towards achieving two of the United Nations Sustainable Development Goals (SDG), namely sustainable cities and communities and climate action.
{"title":"FrugalLight\u0000 : Symmetry-Aware Cyclic Heterogeneous Intersection Control using Deep Reinforcement Learning with Model Compression, Distillation and Domain Knowledge","authors":"Sachin Kumar Chauhan, Rijurekha Sen","doi":"10.1145/3648599","DOIUrl":"https://doi.org/10.1145/3648599","url":null,"abstract":"\u0000 Developing countries need to better manage fast increasing traffic flows, owing to rapid urbanization. Else, increasing traffic congestion would increase fatalities due to reckless driving, as well as keep vehicular emissions and air pollution critically high in cities like New Delhi. State-of-the-art traffic signal control methods in developed countries, however, use expensive sensing, computation and communication resources. How far can control algorithms go, under resource constraints, is explored through the design and evaluation of\u0000 FrugalLight\u0000 (FL) in this paper. We also captured and processed a real traffic dataset at a busy intersection in New Delhi, India, using efficient techniques on low cost embedded devices. This dataset (\u0000 https://delhi-trafficdensity-dataset.github.io\u0000 ) contains traffic density information at fine time granularity of one measurement every second, from all approaches of the intersection for 40 days.\u0000 FrugalLight\u0000 (\u0000 https://github.com/sachin-iitd/FrugalLight\u0000 ) is evaluated on the collected traffic dataset from New Delhi and another open source traffic dataset from New York.\u0000 FrugalLight\u0000 matches the performance of state-of-the-art Convolutional Neural Network (CNN) based sensing and Deep Reinforcement Learning (DRL) based control algorithms, while utilizing resources less by an order of magnitude. We further explore improvements using a careful combination of knowledge distillation and domain knowledge based DRL model compression, with employing Model-Agnostic Meta-Learning to quickly adapt to traffic at new intersections. The collected real dataset and\u0000 FrugalLight\u0000 therefore opens up opportunities for resource efficient RL based intersection control design for the ML research community, where the controller should have limited carbon footprint. Such intelligent, green, intersection controllers can help reduce traffic congestion and associated vehicular emissions, even if compute and communication infrastructure is limited in low resource regions. This is a critical step towards achieving two of the United Nations Sustainable Development Goals (SDG), namely sustainable cities and communities and climate action.\u0000","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139958856","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}
T. Chatty, Bryton L. Moeller, Ioana A. Pantelimon, Catherine D. Parnell, Tahsin M. Khan, Lise Laurin, Jeremy Faludi, Elizabeth L. Murnane
Sustainability has long been a topic of substantial interest the design and human-centered computing communities. With industries increasingly prioritizing climate targets, there is a growing demand for sustainable product design. This paper addresses this need through EcoSketch, a digital tool designed to democratize environmental impact assessments for product designers. Shifting typically retrospective evaluations to the early stages of product development, EcoSketch enables proactive consideration and adoption of sustainable alternatives. Unlike software tailored to environmental scientists, it minimizes the need for specialized training or extensive data inputs. We delve into the development and evaluation of EcoSketch, highlighting its unique features and usability strengths. The paper concludes by discussing design implications and proposing future research avenues to strengthen the intersection of human-computer interaction and sustainable product design, advancing progress on environmental challenges at the systems level.
{"title":"EcoSketch: promoting sustainable design through iterative environmental assessment during early-stage product development","authors":"T. Chatty, Bryton L. Moeller, Ioana A. Pantelimon, Catherine D. Parnell, Tahsin M. Khan, Lise Laurin, Jeremy Faludi, Elizabeth L. Murnane","doi":"10.1145/3648436","DOIUrl":"https://doi.org/10.1145/3648436","url":null,"abstract":"Sustainability has long been a topic of substantial interest the design and human-centered computing communities. With industries increasingly prioritizing climate targets, there is a growing demand for sustainable product design. This paper addresses this need through EcoSketch, a digital tool designed to democratize environmental impact assessments for product designers. Shifting typically retrospective evaluations to the early stages of product development, EcoSketch enables proactive consideration and adoption of sustainable alternatives. Unlike software tailored to environmental scientists, it minimizes the need for specialized training or extensive data inputs. We delve into the development and evaluation of EcoSketch, highlighting its unique features and usability strengths. The paper concludes by discussing design implications and proposing future research avenues to strengthen the intersection of human-computer interaction and sustainable product design, advancing progress on environmental challenges at the systems level.","PeriodicalId":505364,"journal":{"name":"ACM Journal on Computing and Sustainable Societies","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139959744","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：