The increasing number of ontologies of the semantic Web poses new challenges for ontology mapping. In the context of question answering there is a need for good mapping algorithms which efficiently can perform syntactic and semantic mappings between classes and class properties from different ontologies. Mapping algorithms without the help of human experts are in particular desirable when the answer comes from different domain specific databases or ontologies. One of the main problems with any mapping process is that it always has a certain degree of uncertainty associated with it. In this paper we propose a framework based on agents performing mappings and combining beliefs of each individual agent using the Dempster-Shafer rule of combination. We also discuss the problems which can be encountered if we have conflicting beliefs between agents in a particular mapping.
{"title":"Uncertain Reasoning in Multi-agent Ontology Mapping on the Semantic Web","authors":"M. Nagy, M. Vargas-Vera, E. Motta","doi":"10.1109/MICAI.2007.11","DOIUrl":"https://doi.org/10.1109/MICAI.2007.11","url":null,"abstract":"The increasing number of ontologies of the semantic Web poses new challenges for ontology mapping. In the context of question answering there is a need for good mapping algorithms which efficiently can perform syntactic and semantic mappings between classes and class properties from different ontologies. Mapping algorithms without the help of human experts are in particular desirable when the answer comes from different domain specific databases or ontologies. One of the main problems with any mapping process is that it always has a certain degree of uncertainty associated with it. In this paper we propose a framework based on agents performing mappings and combining beliefs of each individual agent using the Dempster-Shafer rule of combination. We also discuss the problems which can be encountered if we have conflicting beliefs between agents in a particular mapping.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115204760","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}
In this paper we exercise the genetic programming of a artificial neural network (ANN) that integrates sensor vision, path planning and steering control of a mobile robot. The training of the ANN is done by a simulation of the robot, its sensors, and environment. The results of each simulation run are then used to denote the ability for the tested network to operate the robot. After less than hundred evaluations we receive an ANN that is able to navigate the robot around obstacles better than a traditional implementation of sensor-based vision and navigation for the same robot.
{"title":"Genetic Evolution of a Neural Network for the Autonomous Control of a Four-Wheeled Robot","authors":"W. Elmenreich, G. Klingler","doi":"10.1109/MICAI.2007.13","DOIUrl":"https://doi.org/10.1109/MICAI.2007.13","url":null,"abstract":"In this paper we exercise the genetic programming of a artificial neural network (ANN) that integrates sensor vision, path planning and steering control of a mobile robot. The training of the ANN is done by a simulation of the robot, its sensors, and environment. The results of each simulation run are then used to denote the ability for the tested network to operate the robot. After less than hundred evaluations we receive an ANN that is able to navigate the robot around obstacles better than a traditional implementation of sensor-based vision and navigation for the same robot.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123521423","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. Guerra-Hernández, G. Ortiz-Hernández, W. A. Luna-Ramírez
This work deals with the problem of intentional learning in a multi-agent system (MAS). Smile (sound multi-agent incremental learning), a collaborative learning protocol which shows interesting results in the distributed learning of well known complex boolean formulae, is adopted here by a MAS of BDI agents to update their practical reasons while keeping MAS-consistency. An incremental algorithm for first-order induction of logical decision trees enables the BDI agents to adopt Smile, reducing the amount of communicated learning examples when compared to our previous non-incremental approaches to intentional learning. The protocol is formalized extending the operational semantics of AgentSpeak(L), and implemented in Jason, its well known Java-based extended interpreter.
{"title":"Jason Smiles: Incremental BDI MAS Learning","authors":"A. Guerra-Hernández, G. Ortiz-Hernández, W. A. Luna-Ramírez","doi":"10.1109/MICAI.2007.16","DOIUrl":"https://doi.org/10.1109/MICAI.2007.16","url":null,"abstract":"This work deals with the problem of intentional learning in a multi-agent system (MAS). Smile (sound multi-agent incremental learning), a collaborative learning protocol which shows interesting results in the distributed learning of well known complex boolean formulae, is adopted here by a MAS of BDI agents to update their practical reasons while keeping MAS-consistency. An incremental algorithm for first-order induction of logical decision trees enables the BDI agents to adopt Smile, reducing the amount of communicated learning examples when compared to our previous non-incremental approaches to intentional learning. The protocol is formalized extending the operational semantics of AgentSpeak(L), and implemented in Jason, its well known Java-based extended interpreter.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122101464","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}
In this paper we present a layered architecture for multirobot motion coordination. The purpose is to control and coordinate autonomous mobile robot with generic shapes and kinematics in a priori known environment. It is a centralized framework, where a leader robot (or a supervisor) plans the motion of all the robots and makes them moving synchronously. The architecture is layered and modularized, where each module is realized with concurrent threads. The underlying motion planner is based on an artificial potential fields method applied on a discretized C-space-time.
{"title":"A Multi-threads Architecture for the Motion Coordination of a Heterogeneous Multi-robot System","authors":"F. Marchese","doi":"10.1109/MICAI.2007.44","DOIUrl":"https://doi.org/10.1109/MICAI.2007.44","url":null,"abstract":"In this paper we present a layered architecture for multirobot motion coordination. The purpose is to control and coordinate autonomous mobile robot with generic shapes and kinematics in a priori known environment. It is a centralized framework, where a leader robot (or a supervisor) plans the motion of all the robots and makes them moving synchronously. The architecture is layered and modularized, where each module is realized with concurrent threads. The underlying motion planner is based on an artificial potential fields method applied on a discretized C-space-time.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125890163","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}
Evolutionary algorithms have been very successful at solving global optimization problems. Two competing goals govern the performance of evolutionary algorithms: exploration and exploitation. This paper proposes a new heuristic to keep population diversity: the shake and the regicide. The shake heuristic improves the exploration by perturbing the whole population. The regicide heuristic (kill the leader) reduces the risk of being, early, trapped by a local minimum. Experiments demonstrate that the Shake-Regicide heuristic improves significantly the precision of the results (in about 3 orders of magnitude) of standard differential evolution, genetic algorithm and evolution strategy.
{"title":"Shake – Regicide: A New Heuristic for the Diversity Control of Evolutionary Algorithms","authors":"J. Ramírez, M. Rivera, A. Hernandez-Aguirre","doi":"10.1109/MICAI.2007.9","DOIUrl":"https://doi.org/10.1109/MICAI.2007.9","url":null,"abstract":"Evolutionary algorithms have been very successful at solving global optimization problems. Two competing goals govern the performance of evolutionary algorithms: exploration and exploitation. This paper proposes a new heuristic to keep population diversity: the shake and the regicide. The shake heuristic improves the exploration by perturbing the whole population. The regicide heuristic (kill the leader) reduces the risk of being, early, trapped by a local minimum. Experiments demonstrate that the Shake-Regicide heuristic improves significantly the precision of the results (in about 3 orders of magnitude) of standard differential evolution, genetic algorithm and evolution strategy.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127250028","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}
Statistical methods have proven to be very effective when addressing linguistic problems, specially when dealing with machine translation. Nevertheless, statistical machine translation effectiveness is limited to situations where large amounts of training data are available. Therefore, the broader the coverage of a SMT system is, the better the chances to get a reasonable output are. In this paper we propose a method to improve quality of translations of a phrase-based machine translation system by extending phrase-tables with the use of translation paraphrases learned from a third language. Our experiments were done translating from Spanish to English pivoting through French.
{"title":"Using Translation Paraphrases from Trilingual Corpora to Improve Phrase-Based Statistical Machine Translation: A Preliminary Report","authors":"F. Herrera, L. Luna","doi":"10.1109/MICAI.2007.34","DOIUrl":"https://doi.org/10.1109/MICAI.2007.34","url":null,"abstract":"Statistical methods have proven to be very effective when addressing linguistic problems, specially when dealing with machine translation. Nevertheless, statistical machine translation effectiveness is limited to situations where large amounts of training data are available. Therefore, the broader the coverage of a SMT system is, the better the chances to get a reasonable output are. In this paper we propose a method to improve quality of translations of a phrase-based machine translation system by extending phrase-tables with the use of translation paraphrases learned from a third language. Our experiments were done translating from Spanish to English pivoting through French.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126914431","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}
In a real world, the emotions play a significant role in rational actions in human communication. In recent years, there has been growing interest in the study of emotions to improve the capabilities of current human-computer interaction. In this paper, we present an effective pattern recognition approach to improve the extraction features in the performance of emotion recognition from video sequences by combining the Nitzberg algorithm and statistics analysis by means of use of the first and second momentum.
{"title":"Algorithm for Affective Pattern Recognition by Means of Use of First Initial Momentum","authors":"R. Romero-Herrera, F. Funes, J. Y. Montiel-Pérez","doi":"10.1109/MICAI.2007.21","DOIUrl":"https://doi.org/10.1109/MICAI.2007.21","url":null,"abstract":"In a real world, the emotions play a significant role in rational actions in human communication. In recent years, there has been growing interest in the study of emotions to improve the capabilities of current human-computer interaction. In this paper, we present an effective pattern recognition approach to improve the extraction features in the performance of emotion recognition from video sequences by combining the Nitzberg algorithm and statistics analysis by means of use of the first and second momentum.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115159835","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}
In this paper, adaptive hierarchical fuzzy CMAC neural network controller (HFCMAC), for a certain class of nonlinear dynamical system is presented. The main advantages of adaptive HFCMAC control are: Better performance of the controller because adaptive HFCMAC can adjust itself to the changing enviroment and can be implemented in real time applications. The proposed method provides a simple control architecture that merges hierarchical structure, CMAC neural network and fuzzy logic. The input space dimension in CMAC is a time-consuming task especially when the number of inputs is huge this would be overload the memory and make the neuro-fuzzy system very hard to implement. This is can be simplified using a number of low-dimensional fuzzy CMAC in a hierarchical form. A new adaptation law is obtained for the method proposed, the overall adaptive scheme guarantees the global stability of the resulting closed-loop system in the sense that all signals involved are uniformly bounded. Simulation results for its applications to one example is presented to demonstrate the performance of the proposed methodology.
{"title":"Adaptive Hierarchical Fuzzy CMAC Controller with Stable Learning Algorithm for Unknown Nonlinear Systems","authors":"F. Ortiz, Wen Yu, M. Moreno-Armendáriz","doi":"10.1109/MICAI.2007.26","DOIUrl":"https://doi.org/10.1109/MICAI.2007.26","url":null,"abstract":"In this paper, adaptive hierarchical fuzzy CMAC neural network controller (HFCMAC), for a certain class of nonlinear dynamical system is presented. The main advantages of adaptive HFCMAC control are: Better performance of the controller because adaptive HFCMAC can adjust itself to the changing enviroment and can be implemented in real time applications. The proposed method provides a simple control architecture that merges hierarchical structure, CMAC neural network and fuzzy logic. The input space dimension in CMAC is a time-consuming task especially when the number of inputs is huge this would be overload the memory and make the neuro-fuzzy system very hard to implement. This is can be simplified using a number of low-dimensional fuzzy CMAC in a hierarchical form. A new adaptation law is obtained for the method proposed, the overall adaptive scheme guarantees the global stability of the resulting closed-loop system in the sense that all signals involved are uniformly bounded. Simulation results for its applications to one example is presented to demonstrate the performance of the proposed methodology.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"34 8","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132835754","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}
MDQL is an algorithm, based on reinforcement learning, for solving multiple objective optimization problems, that has been tested on several applications with promising results. MDQL discretizes the decision variables into a set of states, each associated with actions to move agents to contiguous states. A group of agents explore this state space and are able to find Pareto sets applying a distributed reinforcement learning algorithm. The precision of the Pareto solutions depends on the chosen granularity of the states. A finer granularity on the states creates more precise solutions but at the expense of a larger search space, and consequently the need for more computational resources. In this paper, a very important improvement is introduced into the original MDQL algorithm to incrementally refined the Pareto solutions. The new algorithm, called IMDQL, starts with a coarse granularity to find an initial Pareto set. A vicinity for each of the Pareto solutions in refined and a new Pareto set is founded in this refined state space. This process continues until there is no more improvement within a small threshold value. It is shown that IMDQL not only improves the solutions found by MDQL, but also converges faster. MDQL has also been tested on the solutions of dynamic optimization problems. In this paper, it is also shown that the adaptation capabilities observed in MDQL can be improved with IMDQL. IMDQL was tested on the benchmark problems proposed by Jin. Performance evaluation was made using the Collective Mean Fitness metric proposed by Morrison. IMDQL was compared against an standard evolution strategy with the covariance matrix adaptation (CMA-ES) with very promising results.
MDQL是一种基于强化学习的算法，用于解决多目标优化问题，已经在几个应用程序中进行了测试，结果很有希望。MDQL将决策变量离散为一组状态，每个状态都与将代理移动到连续状态的操作相关联。一组智能体探索这个状态空间，并能够应用分布式强化学习算法找到帕累托集。帕累托解的精度取决于所选择的状态粒度。更细的状态粒度创建更精确的解决方案，但代价是更大的搜索空间，因此需要更多的计算资源。本文对原来的MDQL算法进行了一个非常重要的改进，以逐步改进Pareto解。这个名为IMDQL的新算法从粗粒度开始寻找初始帕累托集。在此精炼状态空间中建立了每个Pareto解的邻域，并建立了一个新的Pareto集。这个过程一直持续，直到在一个小的阈值内没有更多的改进。结果表明，IMDQL不仅改进了MDQL找到的解，而且收敛速度更快。MDQL还对动态优化问题的解决方案进行了测试。本文还表明，在MDQL中观察到的自适应能力可以通过IMDQL得到改进。IMDQL在Jin提出的基准问题上进行了测试。采用Morrison提出的集体平均适应度(Collective Mean Fitness)指标进行绩效评价。将IMDQL与具有协方差矩阵自适应(CMA-ES)的标准进化策略进行了比较，得到了很好的结果。
{"title":"Incremental Refinement of Solutions for Dynamic Multi Objective Optimization Problems","authors":"C. E. Mariano-Romero, M.E.F. Morales","doi":"10.1109/MICAI.2007.47","DOIUrl":"https://doi.org/10.1109/MICAI.2007.47","url":null,"abstract":"MDQL is an algorithm, based on reinforcement learning, for solving multiple objective optimization problems, that has been tested on several applications with promising results. MDQL discretizes the decision variables into a set of states, each associated with actions to move agents to contiguous states. A group of agents explore this state space and are able to find Pareto sets applying a distributed reinforcement learning algorithm. The precision of the Pareto solutions depends on the chosen granularity of the states. A finer granularity on the states creates more precise solutions but at the expense of a larger search space, and consequently the need for more computational resources. In this paper, a very important improvement is introduced into the original MDQL algorithm to incrementally refined the Pareto solutions. The new algorithm, called IMDQL, starts with a coarse granularity to find an initial Pareto set. A vicinity for each of the Pareto solutions in refined and a new Pareto set is founded in this refined state space. This process continues until there is no more improvement within a small threshold value. It is shown that IMDQL not only improves the solutions found by MDQL, but also converges faster. MDQL has also been tested on the solutions of dynamic optimization problems. In this paper, it is also shown that the adaptation capabilities observed in MDQL can be improved with IMDQL. IMDQL was tested on the benchmark problems proposed by Jin. Performance evaluation was made using the Collective Mean Fitness metric proposed by Morrison. IMDQL was compared against an standard evolution strategy with the covariance matrix adaptation (CMA-ES) with very promising results.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"88 4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126307109","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}
We introduce causal agents, a methodology and agent architecture for modeling intelligent agents based on causality theory. We draw upon concepts from classical philosophy about metaphysical causes of existing entities for defining agents in terms of their formal, material, efficient and final causes and use computational mechanisms from Bayesian causal models for designing causal agents. Agent's intentions, interactions and performance are governed by their final causes. A semantic Bayesian causal model, which integrates a probabilistic causal model with a semantic layer, is used by agents for knowledge representation and inference. Agents are able to use semantic information from external stimuli (utterances, for example) which are mapped into the agent's causal model for reasoning about causal relationships with probabilistic methods. Our theory is being tested by an operational multiagents system implementation for managing research products.
{"title":"Modelling Intelligent Agents through Causality Theory","authors":"H. Ceballos, F. Cantú","doi":"10.1109/MICAI.2007.25","DOIUrl":"https://doi.org/10.1109/MICAI.2007.25","url":null,"abstract":"We introduce causal agents, a methodology and agent architecture for modeling intelligent agents based on causality theory. We draw upon concepts from classical philosophy about metaphysical causes of existing entities for defining agents in terms of their formal, material, efficient and final causes and use computational mechanisms from Bayesian causal models for designing causal agents. Agent's intentions, interactions and performance are governed by their final causes. A semantic Bayesian causal model, which integrates a probabilistic causal model with a semantic layer, is used by agents for knowledge representation and inference. Agents are able to use semantic information from external stimuli (utterances, for example) which are mapped into the agent's causal model for reasoning about causal relationships with probabilistic methods. Our theory is being tested by an operational multiagents system implementation for managing research products.","PeriodicalId":296192,"journal":{"name":"2007 Sixth Mexican International Conference on Artificial Intelligence, Special Session (MICAI)","volume":"342 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132608201","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：