Cognitive Science Artificial Intelligence


Published on

A brief survey of approaches to using cognitive science artificial intelligence to achieve goals in both the cognitive science and artificial intelligence fields.

Published in: Technology, Spiritual
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Cognitive Science Artificial Intelligence

  1. 1. Cognitive Science Artificial Intelligence: Simulating the Human Mind to Achieve Goals Samantha Luber University of Michigan Ann Arbor, MI saluber@umich.eduAbstract: This paper provides a general overview of the representing the mind as a high-level collection of rules,interdisciplinary study of cognitive science, specifically symbols, and plans [3].the area of the field involving artificial intelligence. Inaddition, the paper will elaborate on current research for B. Cognitive Science in Artificial Intelligencecognitive science artificial intelligence, highlight theimportance of this research by providing specific In addition to simulating intelligence to model and study theexamples of its applications in present society, and human mind, artificial intelligence involves the study ofbriefly discuss future research opportunities for the cognitive phenomena in machines and attempts tooverlapping fields of cognitive science and artificial implement aspects of human intelligence in computerintelligence. programs. These programs can be used to address a variety of complex problems with the goal of doing so more I. An Overview of Cognitive Science and Artificial efficiently than a human. New theories in the cognitive Intelligence science field often influence improved artificial intelligence agents that better simulate the human thought process [2].Since ancient times, people have conducted countlessexperiments in attempts to better understand the human Achievements in cognitive science help improve artificialmind. These tests eventually lead to the development of simulation of the human mind. In turn, more accuratepsychology. In the late 1930’s, cognitive science emerged as artificial intelligence provides better models of the humanan extension of psychology topics; it is concerned with how mind for cognitive science researchers to use. Although theinformation is stored and transferred in the human mind. It is goals of cognitive science and artificial intelligence differ,an interdisciplinary science, linking psychology, linguistics, collaboration between the two fields is essential for theiranthropology, philosophy, neuroscience, sociology, and success. Cognitive science artificial intelligence refers to thelearning sciences [1]. A useful tool for cognitive researchers, interdisciplinary study that overlaps these areas in attempt toartificial intelligence is the branch of computer science achieve both cognitive science and artificial intelligenceconcerned with creating simulations that model human goals.cognition. In addition to serving as a research tool, artificialintelligence also contains a scientific aspect, focusing on II. Current Research in Cognitive Science Artificialstudying cognitive behavior of machines [2]. Developed to Intelligenceencapsulate the concept of both early cognitive science andintelligence simulated by machines, modern cognitive A fundamental goal of cognitive science artificialscience artificial intelligence focuses on how humans, intelligence is to use the power of computers to understandanimals, and machines store information associated with and supplement human thinking. In artificial intelligence, anperception, language, reasoning, and emotion. intelligent agent refers to a computer-simulated entity that interacts with its environment and works to achieve goals,A. Artificial Intelligence in Cognitive Science both simple and complex [3]. By observing which problems an intelligent agent can solve and how the computer programThe central principle of cognitive science is that a complete solves these problems, researchers in the cognitive scienceunderstanding of the mind cannot be obtained without field aim to develop theories about how the brain learns andanalyzing the mind on multiple levels. In other words, constructs logical rules, how intelligence arises within thenumerous techniques must be used to fully evaluate and brain, insights on which pieces of information humans willunderstand a process of the mind. Artificial intelligence is a forget and remember, and the kinds of resources the humanpowerful approach that allows researchers in cognitive mind uses [2].science to study behavior through computational modelingof the human mind [2]. There are numerous approaches to In addition to gaining a better insight into the nature of thesimulating how the mind is structured with approaches human mind, the ultimate goal of cognitive science artificialranging from creating and observing artificial neurons to intelligence is to eventually develop human-level, machine intelligence. At this level, the intelligent agent would not be
  2. 2. distinguished from human intelligence, a challenge known to solving problems is vital for modeling human levelas the Turing test [3]. Because intelligent agents often face intelligence.situations with incomplete information, encoding data for allpossible situations is a limited approach to simulate human Both the SOAR system and Zadeh’s research contribute tointelligence [4]. In other words, because there are infinitely achieving improved human-level intelligence simulations inmany situations that can arise in the real world, it is the cognitive science artificial intelligence field. Althoughimpossible to design an intelligent agent pre-programmed no artificial intelligence has come close to achieving the goalwith solutions to all of the problems it may face. Instead, an of human-level intelligence, intelligent agents areintelligent agent must be equipped with the ability to make consistently being re-evaluated and improved.decisions based on the information it has and re-evaluate itspast solutions to improve future decisions. Consequently, a III. Applications and the Importance of Cognitive Sciencemore fundamental understanding about how the human mind Artificial Intelligencelearns and solves problems is necessary to design anintelligent agent with the same intelligence. As seen in the goals of the previously mentioned researchers in the field, there are numerous, important, real world A. The SOAR Project applications of cognitive science artificial intelligence research. In our society, engineers and architects constantlyDr. Laird, a professor in computer science at the University face tasks, such as constructing a highway or designing aof Michigan, developed the SOAR system, a cognitive traffic light, that require optimizing a design despite physicalarchitecture programming structure with the goal of and financial limitations. For instance, in the traffic lightsimulating a human brain, as a unique, alternative approach example, an engineer must consider the tradeoff economicsto the traditional and restricted hard-coding data approach. between using stronger materials and the price of theseThe SOAR system stores information retrieved from the materials or calculate statistics on the large amounts ofintelligent agent’s environment in working memory [5]. traffic data available for the intersection to determine lightBecause immediate sensory data is sometimes insufficient timing. With the ability to consider large amounts offor decision-making in the real world, storing previous information and design considerations in a short period ofsituations is useful in differentiating between situations that time, advanced intelligence can be developed to solve thesewould otherwise appear identical to the intelligent agent at a types of complex logic problems [2]. In this way, the use ofspecific instance [6]. In short, maintaining memory of past artificial intelligence as a tool for engineers could make theevents “makes it possible to not only make correct decisions design process faster, more efficient, and more accurate.but to learn the correct decision” [6]. When the intelligentagent enters an impasse, the agent can search its memory for In addition, the creation of a human level intelligent agenta solution to the problem. If the problem is unique, the agent provides a “better mirror” of the human mind that is easier toremembers its actions in case the problem is encountered study than the human brain for cognitive science researchers.again [5]. In summary, the SOAR cognitive architecture By studying realistic simulations of human cognition,system relies on maintaining information from decisions and theories can be drawn about humans’ nature and cognitiveoutcomes in past experiences to improve future decisions in limitations. Furthermore, researchers can achieve specificsimulating human behavior. The SOAR system is a useful cognitive science goals, such as understanding howtool for using simulated human intelligence to solve intelligence develops in the brain or how damage to differentcomplex problems. parts of the brain affect cognition [2]. Progress in these areas can powerfully impact how the human mind is understood, B. Simulating Creativity with the potential of leading to improvements in present society. For instance, a better understanding of how theIn his papers on the simulation of human level intelligence human mind learns and retains information can lead toin the decision process, Dr. Zadeh emphasizes the improved learning methods implemented in schools toimportance of imitating creativity in intelligent agents. accelerate human progress. In the same way, improvedAlthough knowledge of past experiences is a useful tool in theories on how different areas of the brain affect behaviordecision-making, Dr. Zadeh acknowledges that “creativity is can help develop medical solutions for victims of braina gifted ability of human beings in thinking, inference, trauma [3]. The useful applications of pursuing research inproblem solving, and product development” [7]. In his cognitive science artificial intelligence continue to grow asformal definition of the unique ability, creativity is divided research in the field continues.into three categories: abstract, concrete, and artistic. Morerelevant to engineering applications, concrete creativity IV. The Future of Cognitive Science Artificial Intelligenceinvolves generating new, innovative solutions in anenvironment limited by goals and available conditions [7]. Although there have been many breakthroughs in theAiming to equip intelligent agents with the creative ability of cognitive science artificial intelligence field, researchers arethe human mind, Zadeh provides an outlined approach for continually working to improve intelligent agents.implementing the creative process in a computer program.The ability of an intelligent agent to create new approaches The human mind has the impressive capability of preforming numerous mental and physical tasks with little
  3. 3. mental strain [2]. On the other hand, computer simulated [5] Lehman, J.F., Laird, J., and Rosenbloom, P. (2006). Aintelligence is limited by the speed and capacity of hardware Gentle Introduction to SOAR, an Architecture forfor performing computations. The development of advanced Human Cognition: 2006 Update.nanotechnology to increase hardware speed and memorywill reduce this restraint on simulating human level [6] Laird, J.E., and Wang, Y. (2007). The Importance ofintelligence [8]. Furthermore, while theories of cognitive Action History in Decision Making and Reinforcementscience artificial intelligence have fostered improved Learning. Proceedings of the Eighth Internationalunderstanding of the human mind, advancements in the Conference on Cognitive Modeling. Ann Arbor, MI.psychology and cognitive science fields, that help betterunderstand human behavior, can be used to further improve [7] Zadeh, L. (2008). On Cognitive Foundations ofintelligent agents [3]. Finally, an issue more acknowledged Creativity and the Cognitive Process of the public than researchers in the field, include ethical Proceedings of the Seventh IEEE Internationaland organizational concerns with the coevolution of humans Conference on Cognitive Informatics: ICCI 2008 :and intelligent systems; these issues may one day have to be August 14-16, 2008, Stanford University, California,addressed [8]. For instance, society would have to address USA. [Piscataway, N.J.]: IEEE Xplore.restrictions on how a human-simulating robot can behave. [8] Jacobstein, N. (2005). The Prospects for AI. ITBecause intelligent agents are still far from achieving Conversations.artificial intelligence goals, such as passing the Turing test, < cognitive science goals, such as achieving human level ml>.intelligence or improving the present understanding of thehuman mind, there are still many opportunities for researchachievements in the cognitive science artificial intelligencefield. This room for growth shows great potential fordeveloping technology to increase the progress of mankind. V. ConclusionsArtificial intelligence is an extremely useful tool forcognitive science research of both fundamental and highlevel understanding of the human mind by simulating thehuman mind. In the same way, cognitive science theoriesprovide useful insight on human cognition that can beencoded into artificial intelligence. Cognitive scienceartificial intelligence is a powerful research area thataddresses the goals of both cognitive science and artificialintelligence. As research in the field continues, improvedintelligent agents will be developed with the ability tosimulate human-level intelligence in the final cognitivescience research goal of fully understanding the human mindor to address important, complex problems of mankindthrough artificial intelligence. VI. References[1] Thagard, Paul. (2009). Cognitive Science. The Stanford Encyclopedia of Philosophy (Fall 2008 Edition), Edward N. Zalta (ed.).[2] Simon, H. (2010). Cognitive Science: Relationship of AI to Psychology and Neuroscience. AAAI.[3] Wang, Y. (2008). Proceedings of the Seventh IEEE International Conference on Cognitive Informatics: ICCI 2008 : August 14-16, 2008, Stanford University, California, USA. [Piscataway, N.J.]: IEEE Xplore.[4] Bickhard, M., and Terveen, L. (1995). Foundational Issues in Artificial Intelligence and Cognitive Science. Elsevier Science Publishers.