Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

4190.408 2016-Spring

Intelligent Agents

Byoung-Tak Zhang

School of Computer Science and Engineering

Seoul National University

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

적응학습형에이전트An Adaptive Learning Agents

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

What is Agent?

• “Intelligent agents continuously perform three functions: perception of dynamic conditions in the environment; action to affect conditions in the environment; and reasoning to interpret perceptions, solve problems, draw inferences, and determine actions” [Hayes-Roth, 1995].

• “An autonomous agent is a system situated within and a part of an environment that senses that environment and acts on it, over time, in pursuit of its own agenda and so as to effect what it senses in the future” [Franklin and Graesser, 1995].

• “A hardware or (more usually) software-based computer system that enjoys the following properties: autonomy, social ability, reactivity, pro-activeness” [Wooldridge and Jennings, 1995].

• “Autonomous agents are computational systems that inhabit some complex dynamic environment, sense and act autonomously in this environment, and by doing so realize a set of goals or tasks for which they are designed” [Maes, 1995].

• “Intelligent agents are software entities that carry out some set of operations on behalf of a user or another program with some degree of independence or autonomy, and in so doing, employ some knowledge or representation of the user’s goals or desires” [IBM, tech. report.].

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

Indirect Manipulation Direct Manipulation

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

Characteristics of Agents

Reactivity

Autonomy

Collaborative behavior

Communication ability

Inferential capability

Temporal continuity

Personality

Mobility

Adaptivity

Learnability

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

[Gilbert et al., 1995]

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

[Nwana, 1996]

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

[Franklin and Graesser, 1996]

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

Classification of Agents

[Caglayan and Harrison, 1997]

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

Learning Agents

• “Agents that change its behavior based on its previous experience.”

• Learning Methods– Reinforcement Learning

– Decision Trees

– Bayesian Learning

– Neural Networks

– Genetic Algorithms

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

Reinforcement Learning Agents

• Generalized model learning for reinforcement learning on a humanoid robot: http://www.youtube.com/watch?v=mRpX9DFCdwI

• Autonomous spider learns to walk forward by reinforcement learning: http://www.youtube.com/watch?v=RZf8fR1SmNY&feature=related

• Reinforcement learning for a robotic soccer goalkeeper: http://www.youtube.com/watch?v=CIF2SBVY-J0&feature=related

(c)

2008

SNU

11

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)© 2010, SNU

Biointelligence Lab,

http://bi.snu.ac.kr/

12

Drivatar – Racing Game Agents

Probabilistically mimicking human driving in Forza 2, which is a racing game of Microsoft XBOX 360, through machine learning

Position Lane SpeedBreak /Accel.

Modelling of driving pattern based on probability

Segmenting all the paths Learning optimal paths which gamers choose

(Imitation Approach)

The Future of Racing Games http://www.youtube.com/watch?v=TaUyzlKKu-E

Generated infinite driving patterns through probabilistic modeling

Microsoft Research in Cambridge, UK

Driving Pattern in the Game

(참고: Thore Graepel, MS Research Cambridge) Whole-audience Control of a Racing Gamehttp://www.youtube.com/watch?v=NS_L3Yyv2RI

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

Personal Robots at Home and Office

© 2010, SNU Biointelligence Lab,

http://bi.snu.ac.kr/

13

http://www.youtube.com/watch?v=

mgHUNfqIhAc&feature=related

PR2 Robot Plays Pool

PR2 Robot Cleans Up

http://www.youtube.com/watch?v=g

Yqfa-YtvW4&feature=related

PR2 Robot of Willow Garage

(참고: Willow Garage)

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

Soccer Robots

• RoboCup 2000:

Beyond Human: Robot Soccer

• Humanoid Robot Soccer 2007:

RoboCup 2007 Final, Humanoid League

• RoboCup 2008:

CMDragons RoboCup 2008 SSL Highlights

• KondoCup Robot Soccer 2008:

12th KondoCup Robot Soccer: Cool Moves!

• RoboCup 2010

– https://www.youtube.com/watch?v=4wMSiKHPKX4

• RoboCup 2012

– https://www.youtube.com/watch?v=4B_sB0q4IDU

• RoboCup 2014

– https://www.youtube.com/watch?v=dhooVgC_0eY

(c) 2008 SNU Biointelligence Laboratory,

http://bi.snu.ac.kr/14

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

DARPA Grand Challenge Autonomous Driving Robots

© 2009, SNU

Biointelligence Lab,

http://bi.snu.ac.kr/

15

By applying machine learning methods to self-driving of unmanned cars, Stanford team won Grand Challenge in 2005, and was the second best in Urban Challenge of 2007.

Video

사람의운전패턴을학습

Car

Human

2005: Self-driving through 175 milesof desert area course in 10 hours

2007: Self-driving through 96 KMof urban area course in 6 hours

Prob.modelling

Terrain recognition using laser

Terrain recognition and planning

[Sebastian Thrun, Stanley & Junior, Stanford Univ.]

DARPA Grand Challenge:

Final Part 1

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

© 2008, SNU

Biointelligence Lab, 16

Autonomous Helicopter ControlAutomatically controlling RC helicopters through Reinforcement Learning

Automatic control through RLRC helicopters with accelerometer

자동제어를통한고난이도비행

Complex flight through auto. control

(참고: Andrew Ng, Stanford Univ.) Stanford Autonomous Helicopter - Airshow #2:

http://www.youtube.com/watch?v=VCdxqn0fcnE

Bio

Intelligence4190.408 Artificial Intelligence (2016-Spring)

DARPA Robotics Challenge

• focuses on disaster or emergency-response scenarios1. Drive a utility vehicle at the site.2. Travel dismounted across rubble.3. Remove debris blocking an entryway.4. Open a door and enter a building.5. Climb an industrial ladder and traverse an industrial walkway.6. Use a tool to break through a concrete panel.7. Locate and close a valve near a leaking pipe.8. Connect a fire hose to a standpipe and turn on a valve.

• 10 minute summary video: https://www.youtube.com/watch?v=TW3nD7ZwMWw