Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8460059
Simon Fauvel, Han Yu, C. Miao, Li-zhen Cui, Hengjie Song, L. Zhang, Xiaoming Li, Cyril Leung
Massive Open Online Courses (MOOCs) have gained tremendous popularity in the last few years. Thanks to MOOCs, millions of learners from all over the world have taken thousands of high-quality courses for free. Artificial intelligence (AI) has played an important role in making MOOCs what they are today. By exploiting the vast amount of data generated by learners engaging in MOOCs, AI techniques have been proposed to improve our understanding of MOOC participants and enable MOOC practitioners to deliver better courses. These approaches have also greatly improved student experience and learning outcomes through constructing intelligent and personalized learning trajectories. In this paper, we first review the state-of-the-art AI research making an impact on MOOCs education, emphasizing on works which aim to enhance our understanding of student learning behaviours, improve student engagement, and improve learning outcomes. We then offer an overview of important future research to carry out in sub-fields of AI to enable MOOCs to reach their full potential.
{"title":"Artificial Intelligence Powered MOOCs: A Brief Survey","authors":"Simon Fauvel, Han Yu, C. Miao, Li-zhen Cui, Hengjie Song, L. Zhang, Xiaoming Li, Cyril Leung","doi":"10.1109/AGENTS.2018.8460059","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8460059","url":null,"abstract":"Massive Open Online Courses (MOOCs) have gained tremendous popularity in the last few years. Thanks to MOOCs, millions of learners from all over the world have taken thousands of high-quality courses for free. Artificial intelligence (AI) has played an important role in making MOOCs what they are today. By exploiting the vast amount of data generated by learners engaging in MOOCs, AI techniques have been proposed to improve our understanding of MOOC participants and enable MOOC practitioners to deliver better courses. These approaches have also greatly improved student experience and learning outcomes through constructing intelligent and personalized learning trajectories. In this paper, we first review the state-of-the-art AI research making an impact on MOOCs education, emphasizing on works which aim to enhance our understanding of student learning behaviours, improve student engagement, and improve learning outcomes. We then offer an overview of important future research to carry out in sub-fields of AI to enable MOOCs to reach their full potential.","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"467 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127542162","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}
Pub Date : 2018-07-01DOI: 10.1109/agents.2018.8460122
{"title":"IEEE ICA 2018: 2018 IEEE International Conference on Agents","authors":"","doi":"10.1109/agents.2018.8460122","DOIUrl":"https://doi.org/10.1109/agents.2018.8460122","url":null,"abstract":"","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131002269","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}
Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8460054
Y. Ohmoto, Shunya Ueno, T. Nishida
The human mental stance towards a virtual agent has an influence on the social relationship that exists between them. In this study, we focus on contingency, which is the behavior that occurs synchronously with the human action, and the icebreaker, which is a facilitation exercise that helps start an interaction. The aim of this study is to investigate whether an agent's contingent responses and icebreakers with the contingent agent are capable of inducing and maintaining an intentional stance. We conducted an experiment using an agent which provided contingent responses. In the experiment, participants first interacted with the contingent agent through an icebreaker. Afterwards, all participants performed a collaborative task. As a result, we conclude that the contingent responses and the icebreaker are capable of inducing and partially maintaining the intentional stance. In particular, the icebreaker designed based on interaction training techniques was effective to induce the intentional stance.
{"title":"Effect of Virtual Agent's Contingent Responses and Icebreakers Designed based on Interaction Training Techniques on Inducing Intentional Stance","authors":"Y. Ohmoto, Shunya Ueno, T. Nishida","doi":"10.1109/AGENTS.2018.8460054","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8460054","url":null,"abstract":"The human mental stance towards a virtual agent has an influence on the social relationship that exists between them. In this study, we focus on contingency, which is the behavior that occurs synchronously with the human action, and the icebreaker, which is a facilitation exercise that helps start an interaction. The aim of this study is to investigate whether an agent's contingent responses and icebreakers with the contingent agent are capable of inducing and maintaining an intentional stance. We conducted an experiment using an agent which provided contingent responses. In the experiment, participants first interacted with the contingent agent through an icebreaker. Afterwards, all participants performed a collaborative task. As a result, we conclude that the contingent responses and the icebreaker are capable of inducing and partially maintaining the intentional stance. In particular, the icebreaker designed based on interaction training techniques was effective to induce the intentional stance.","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131835515","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}
Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8460040
Cho Cho Hnin, Naw Naw, Aung Win
As social networks and online sites are growing rapidly, people can express their opinions in the form of comments and reviews. To analyze such opinionated reviews, the proposed system presents a linguistic approach to opinion mining. This system analyzes hotel user reviews written in Myanmar language and performs the opinion mining tasks at the aspect level. Finally, the system classifies the aspects/features contained in the reviews as positive, negative or neutral. The important task of aspect level opinion mining is identifying the relations between aspects and opinion words in the reviews. This detection is a big challenge because of informal writing styles of reviews. Especially, it is a difficult task of aspect level opinion mining on Myanmar reviews due to the nature of Myanmar language. Therefore, the proposed system mainly focuses on extracting the relevant pairs of aspects and opinion words from the user reviews using the syntactic patterns and some linguistic rules.
{"title":"Aspect Level Opinion Mining for Hotel Reviews in Myanmar Language","authors":"Cho Cho Hnin, Naw Naw, Aung Win","doi":"10.1109/AGENTS.2018.8460040","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8460040","url":null,"abstract":"As social networks and online sites are growing rapidly, people can express their opinions in the form of comments and reviews. To analyze such opinionated reviews, the proposed system presents a linguistic approach to opinion mining. This system analyzes hotel user reviews written in Myanmar language and performs the opinion mining tasks at the aspect level. Finally, the system classifies the aspects/features contained in the reviews as positive, negative or neutral. The important task of aspect level opinion mining is identifying the relations between aspects and opinion words in the reviews. This detection is a big challenge because of informal writing styles of reviews. Especially, it is a difficult task of aspect level opinion mining on Myanmar reviews due to the nature of Myanmar language. Therefore, the proposed system mainly focuses on extracting the relevant pairs of aspects and opinion words from the user reviews using the syntactic patterns and some linguistic rules.","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128902517","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}
Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8460083
Kazuya Aizawa, K. Tei, S. Honiden
Safety properties for systems are guaranteed under assumptions to an environment. If the assumptions are broken at runtime, the safety properties are no longer guaranteed. The system should adapt to the changes in order to guarantee the safety properties or relaxed safety properties. Our purpose is establishing techniques to identify the maximum level of safety properties that can be guaranteed in a changed environment. The technique should be efficient so that it is applicable to runtime usage. In this paper, we propose an efficient algorithm that identifies the maximum level of safety properties. Our idea is analyzing availability of each safety property guarantee at a time and restricting analysis only in changed part of the previous result, instead of analysis from the scratch. We extend an existing analysis algorithm based on two-player game to realize the difference analysis. We evaluate our algorithm in terms of (1) level of safety properties and (2) computational time through two case studies.
{"title":"Identifying safety properties guaranteed in changed environment at runtime","authors":"Kazuya Aizawa, K. Tei, S. Honiden","doi":"10.1109/AGENTS.2018.8460083","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8460083","url":null,"abstract":"Safety properties for systems are guaranteed under assumptions to an environment. If the assumptions are broken at runtime, the safety properties are no longer guaranteed. The system should adapt to the changes in order to guarantee the safety properties or relaxed safety properties. Our purpose is establishing techniques to identify the maximum level of safety properties that can be guaranteed in a changed environment. The technique should be efficient so that it is applicable to runtime usage. In this paper, we propose an efficient algorithm that identifies the maximum level of safety properties. Our idea is analyzing availability of each safety property guarantee at a time and restricting analysis only in changed part of the previous result, instead of analysis from the scratch. We extend an existing analysis algorithm based on two-player game to realize the difference analysis. We evaluate our algorithm in terms of (1) level of safety properties and (2) computational time through two case studies.","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115517240","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}
Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8460136
Elnaz Shafipour Yourdshahi, Thomas Pinder, Gauri Dhawan, L. Marcolino, P. Angelov
In complex environments, agents must be able to cooperate with previously unknown team-mates, and hence dynamically learn about other agents in the environment while searching for optimal actions. Previous works employ Monte Carlo Tree Search approaches. However, the search tree increases exponentially with the number of agents, and only scenarios with very small team sizes have been explored. Hence, in this paper we propose a history-based version of UCT Monte Carlo Tree Search, using a more compact representation than the original algorithm. We perform several experiments with a varying number of agents in the level-based foraging domain, an important testbed for ad-hoc teamwork. We achieve better overall performance than the state-of-the-art and better scalability with team size. Additionally, we contribute an open-source version of our system, making it easier for the research community to use the level-based foraging domain as a benchmark problern for ad-hoc teamwork.
{"title":"Towards Large Scale Ad-hoc Teamwork","authors":"Elnaz Shafipour Yourdshahi, Thomas Pinder, Gauri Dhawan, L. Marcolino, P. Angelov","doi":"10.1109/AGENTS.2018.8460136","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8460136","url":null,"abstract":"In complex environments, agents must be able to cooperate with previously unknown team-mates, and hence dynamically learn about other agents in the environment while searching for optimal actions. Previous works employ Monte Carlo Tree Search approaches. However, the search tree increases exponentially with the number of agents, and only scenarios with very small team sizes have been explored. Hence, in this paper we propose a history-based version of UCT Monte Carlo Tree Search, using a more compact representation than the original algorithm. We perform several experiments with a varying number of agents in the level-based foraging domain, an important testbed for ad-hoc teamwork. We achieve better overall performance than the state-of-the-art and better scalability with team size. Additionally, we contribute an open-source version of our system, making it easier for the research community to use the level-based foraging domain as a benchmark problern for ad-hoc teamwork.","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131684792","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}
Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8460133
Shohei Yamamoto, S. Matsubara
This paper proposes a method of working condition selection based on type identification of crowd workers. Here, the working condition selection means finding the values of working conditions that are suitable for individual workers. Multi-armed bandit techniques are promising, but it may happen that exploring various task settings for a single worker interferes with that worker, which deteriorates the quality of contributions. To solve this problem, we introduce the type identification test, i.e., we divide the entire period for a worker into a type identification phase and an execution phase and alternately handle the calculation at the individual level and at the aggregate level. Our method can find an appropriate task setting without exploring various settings for a worker, i.e., excessively interfering with the worker. Also, we provide a method of calculating the optimal type identification test to maximize the expected quality of contributions in the execution phase. Finally, we show our method outperforms conventional multi-armed bandit algorithms such as Softmax and UCB1 with data we collected on the Amazon Mechanical Turk and with a simulation.
{"title":"Adaptive Selection of Working Conditions for Crowdsourced Tasks","authors":"Shohei Yamamoto, S. Matsubara","doi":"10.1109/AGENTS.2018.8460133","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8460133","url":null,"abstract":"This paper proposes a method of working condition selection based on type identification of crowd workers. Here, the working condition selection means finding the values of working conditions that are suitable for individual workers. Multi-armed bandit techniques are promising, but it may happen that exploring various task settings for a single worker interferes with that worker, which deteriorates the quality of contributions. To solve this problem, we introduce the type identification test, i.e., we divide the entire period for a worker into a type identification phase and an execution phase and alternately handle the calculation at the individual level and at the aggregate level. Our method can find an appropriate task setting without exploring various settings for a worker, i.e., excessively interfering with the worker. Also, we provide a method of calculating the optimal type identification test to maximize the expected quality of contributions in the execution phase. Finally, we show our method outperforms conventional multi-armed bandit algorithms such as Softmax and UCB1 with data we collected on the Amazon Mechanical Turk and with a simulation.","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122575992","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}
Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8460084
P. Hsu, Chin-chiang Chou, Szu-Hao Huang, An-Pin Chen
Traditional professional traders and institutional investors utilized complex statistical models to price various derivative contracts and make trading decisions in the option and future markets. In recent years, with the rapid growth of algorithmic trading and program trading, the advanced information and communication technology has become an indispensable element for high-frequency traders, especially for the market makers. In addition, artificial intelligence and deep learning also plays an important role in novel financial technology (FinTech) research field. In this paper, we proposed a market making quotation strategy based on deep learning structure and practical finance domain knowledge. The proposed dual agents will simultaneously model the option prices and bid-ask spreads. The experiments demonstrate that our system can precisely estimate the value of options than famous financial engineering models. It also can be extended to develop proper market making quotation strategies to trade the options of Taiwan Stock Exchange Capitalization Weighted Stock Index(TAIEX).
{"title":"A Market Making Quotation Strategy Based on Dual Deep Learning Agents for Option Pricing and Bid-Ask Spread Estimation","authors":"P. Hsu, Chin-chiang Chou, Szu-Hao Huang, An-Pin Chen","doi":"10.1109/AGENTS.2018.8460084","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8460084","url":null,"abstract":"Traditional professional traders and institutional investors utilized complex statistical models to price various derivative contracts and make trading decisions in the option and future markets. In recent years, with the rapid growth of algorithmic trading and program trading, the advanced information and communication technology has become an indispensable element for high-frequency traders, especially for the market makers. In addition, artificial intelligence and deep learning also plays an important role in novel financial technology (FinTech) research field. In this paper, we proposed a market making quotation strategy based on deep learning structure and practical finance domain knowledge. The proposed dual agents will simultaneously model the option prices and bid-ask spreads. The experiments demonstrate that our system can precisely estimate the value of options than famous financial engineering models. It also can be extended to develop proper market making quotation strategies to trade the options of Taiwan Stock Exchange Capitalization Weighted Stock Index(TAIEX).","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127898042","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}
Pub Date : 2018-07-01DOI: 10.1109/agents.2018.8459978
{"title":"Copyright","authors":"","doi":"10.1109/agents.2018.8459978","DOIUrl":"https://doi.org/10.1109/agents.2018.8459978","url":null,"abstract":"","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116703047","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}
Pub Date : 2018-07-01DOI: 10.1109/AGENTS.2018.8459907
Yuchen Wang, F. Ren, Minjie Zhang
Social conventions have been used as an efficient mechanism to facilitate coordination among agents. Establishing a convention in a decentralised manner has attracted much attention in the literature. Existing techniques on convention emergence are not robust. These techniques may establish sub-conventions under particular network structures. The emergence of sub-conventions indicates that agents in a society fail to conform to a single convention. As a result, the coordination among these agents is negatively affected. In this paper, we propose a strategy to avoid sub-conventions under diverse network structures. The proposed strategy requires agents to only have local views. We prove that a convention can be established using the proposed strategy. We also give empirical studies on the speed of convention emergence with various experimental settings,
{"title":"A Cyclical Social Learning Strategy for Robust Convention Emergence","authors":"Yuchen Wang, F. Ren, Minjie Zhang","doi":"10.1109/AGENTS.2018.8459907","DOIUrl":"https://doi.org/10.1109/AGENTS.2018.8459907","url":null,"abstract":"Social conventions have been used as an efficient mechanism to facilitate coordination among agents. Establishing a convention in a decentralised manner has attracted much attention in the literature. Existing techniques on convention emergence are not robust. These techniques may establish sub-conventions under particular network structures. The emergence of sub-conventions indicates that agents in a society fail to conform to a single convention. As a result, the coordination among these agents is negatively affected. In this paper, we propose a strategy to avoid sub-conventions under diverse network structures. The proposed strategy requires agents to only have local views. We prove that a convention can be established using the proposed strategy. We also give empirical studies on the speed of convention emergence with various experimental settings,","PeriodicalId":248901,"journal":{"name":"2018 IEEE International Conference on Agents (ICA)","volume":"29 19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116540905","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：