Robotics 2019 Agent-embedded robots with machine intelligence - JongHwan Kim - Korea Advanced Institute of Science and Technology

Jong-Hwan Kim

摘要

To increase agent-embedded robots with system intelligence (MI), the design of intelligence working architecture (IOA) is needed for sensing, wondering and movement. One of the key modules in IOA is the reminiscence module for storing temporal occasion sequences of responsibilities, the mechanism of idea for reasoning, and movement making plans for execution, amongst others. This talk introduces how to increase agent embedded robots with MI primarily based on IOA, that specialize in long-term reminiscence for lively expertise acquisition and adaptive know-how application. The lengthyterm reminiscence is advanced as an integrated multireminiscence neural model, wherein episodic memory is designed the use of a Deep DRN (Developmental Resonance Network) neural model and semantic memory is constructed the usage of the DRN-tree. Procedural reminiscence is likewise designed using the context-based RNN (recurrent neural community) to keep the trajectories of the manipulators at the side of context facts and then retrieve them in step with the context without conscious thinking. Robots are taught both via human demonstration or symbolic description. A conduct suitable to the modern situation is selected with the aid of the mechanism of concept discovered through device intelligence studying, at the same time as a proper challenge is retrieved from the Deep DRN model. The behaviors are done safely and fast with the movement planning set of rules. The effectiveness of the agent-embedded robot improvement is demonstrated thru experiments with a humanoid robot, Mybot, evolved in the Robot Intelligence Technology Lab. At KAIST, Agent embedded Mybot is introduced specifically for herbal interactions such as VQA (Visual Question Answering) with human beings. In the remaining part, AI World Cup shall be delivered, which has three classes, AI Soccer, AI Commentator, and AI Reporter.??Robots are machines or mechanical human beings which are designed to help humans with arduous and complex duties. However, such robots aren't any greater simply mechanical layout rather they've emerge as smarter with time and development of technologies. AI developments have precipitated evolution and higher ability in robots. Even robotics and AI together can revolutionize nearly any industry for the more desirable. Artificial intelligence (AI) and robotics have turn out to be more and more hot subjects in the press and in academia. In October 2017, Bloomberg published a piece of writing claiming that artificial intelligence is probably to be the “most disruptive pressure in technology inside the coming decade” and caution that corporations which are sluggish to embody the technology may additionally risk extinction. Similarly, the subsequent month, the Financial Times declared?that the “robotic navy” is remodeling the worldwide place of work. This hobby is possibly due to the speedy gains that synthetic intelligence has been making in some programs, inclusive of photo recognition and abstract strategy games, and that advanced robotics has been making in labs, despite the fact that vast industrial applications can be lagging. Scholars were more and more interested in the monetary, social, and distributive implications of synthetic intelligence, robotics, and other kinds of automation. For instance, over the last 2?years, economists at the University of Toronto have convened meetings across the economics of artificial intelligence, which have been attended by means of a spectacular array of economics pupils from numerous factor of views inclusive of Nobel Prize winners Edmund Phelps, Paul Romer, Joseph Stiglit Some research has taken a morez, and others. There are some of well-attended meetings for felony, manufacturing, technical, and well known-interest groups which includes the World Conference on Robotics and Artificial Intelligence, We Robot, and AI Now. Organizational students are a chunk overdue to the sport and have only just started out to awareness on the organizational implications of artificial intelligence, robotics, and different kinds of advanced technology. However, as we describe on this primer, we believe that these technology present a completely unique opportunity for organizational pupils. Periods of high-quality technological exchange can result in first-rate development but also incredible turmoil. For example, while the steam engine caused superb monetary increase.??While many instructions can be drawn from previous episodes of automation, it's miles possible that synthetic intelligence and robotics may have precise results. Differences from previous episodes of automation encompass that the character of business hobby has shifted dramatically over the past decade such that many companies now depend on platform (i.e., 2sided marketplace) business models, artificial intelligence is likely to have an effect on white-collar people more so than blue-collar people (whilst perhaps robotics may also have an effect on blue-coll

机译：为了增加具有系统智能的代理嵌入式机器人（MI），需要对智能工作架构（IOA）的设计进行传感，想知道和运动。 IOA中的一个关键模块是用于存储职责的时间场合序列的Reminiscence模块，其推理的理念的机制，以及制定计划的行动，包括其他人。这次谈判介绍了如何利用MI增加Agent嵌入式机器人，主要基于IOA，专门从事长期回忆，以获得生动的专业知识和适应性专业知识。作为集成的多态性神经模型，LongeSym Reminiscence先进，其中，情节存储器设计使用深Drn（发展共振网络）神经模型和语义记忆构成DRN树的使用。程序回忆同样使用基于上下文的RNN（经常性神经社区）来设计，以将操纵器的轨迹保持在上下文事实的一侧，然后在没有意识的思考的情况下将它们与上下文中的上下文中的。通过人类示范或象征描述教授机器人。借助于通过设备智能研究发现的概念机制选择适合于现代情况的行为，同时从深DRN模型中检索适当的挑战。行为安全，快速完成运动规则规则。 Agent嵌入式机器人改善的有效性通过实验与人形机器人，MyBot进行了实验，演变在机器人智能技术实验室中。在卡斯特，代理嵌入式Mybot是专门用于草药互动，例如与人类的VQA（视觉问题回答）。在剩下的部分中，AI世界杯应送达，其中有三个班级，AI足球，AI评论员和AI记者.??ROBOS是机械或机械人类，旨在帮助人类具有艰巨和复杂的职责。然而，这种机器人并不是更简单的机械布局，而是随着技术的时间和发展，它们出现得更令人更聪明。 AI的发展在机器人中具有沉淀的演变和更高的能力。即使是机器人和AI也可以彻底彻底改变任何行业，以便更为理想。人工智能（AI）和机器人在新闻界和学术界的越来越多的热门科目。 2017年10月，布隆伯格发布了一份文件，声称人工智能可能是“未来十年技术中最具破坏性的压力”，并且谨慎对实施该技术的公司可能另外冒失而行。同样，随后的月份，金融时报宣称？“机械手海军”正在重塑全球工作地点。这种爱好可能是由于综合智能在某些计划中的快速增益，包括照片识别和抽象策略游戏，并且尽管大规模的工业应用可能滞后，但先进的机器人在实验室中已经制定了。学者对综合智能，机器人和其他自动化的货币，社会和分配含义越来越感兴趣。例如，在过去的2年中，多伦多大学的经济学家在人工智能经济学中召开会议，这已经通过了来自众多观点的巨大经济学学生，包括诺贝尔奖获奖者Edmund的众多观点Phelps，Paul Romer，Joseph Stiglit一些研究已经采取了更多的研究和其他研究。有些人参加了一些良好的会议，用于重罪，制造，技术和众所周知的兴趣集团，包括世界机器人和人工智能世界会议，我们现在的机器人和AI。组织学生对这项运动的逾期逾期，才刚刚开始了解人工智能，机器人和不同类型的先进技术的组织影响。然而，正如我们在这一底漆上描述的那样，我们认为这些技术为组织学生提供了一个完全独特的机会。高质量的技术交流时期可导致一流的开发，但也令人难以置信的动荡。例如，虽然蒸汽机引起卓越的货币增加。虽然许多指令可以从先前的自动化剧集中汲取许多指令，但这是综合智能和机器人的数英里可能具有精确的结果。与以往的自动化剧集的差异包括在过去十年中，商业爱好的特征已经大幅转移，使得许多公司现在依赖于平台（即2楼）的商业模式，人工智能可能对白领有影响比蓝领人民更多（虽然机器人）也可能对蓝色电池产生影响

Robotics 2019 Agent-embedded robots with machine intelligence - JongHwan Kim - Korea Advanced Institute of Science and Technology

摘要

著录项

相似文献

相关主题

期刊订阅