Finding aesthetic pleasure on the edge of chaos: A proposal for robotic creativity Ron Chrisley COGS Department of Informatics University of Sussex Workshop.

Slides:



Advertisements
Similar presentations
Implementing the Tech Standards Presenter: Eric Curts eTech|OHIO Tech Conference 2006.
Advertisements

Creative Development. 1: Explores different media and responds to a variety of sensory experiences. Engages in representational play. Scale points 1 –
Evolving concepts of creativity: A mirror, a tightrope and an inkblot Ron Chrisley COGS Department of Informatics University of Sussex Workshop on Evolving.
Cognitive Systems, ICANN panel, Q1 What is machine intelligence, as beyond pattern matching, classification and prediction. What is machine intelligence,
NEU Neural Computing MSc Natural Computation Department of Computer Science University of York.
16 key concepts.
Experience of a Learning Organization: How To Grow Beyond Blame.
Music Introduction to Humanities. Music chapter 9 Music is one of the most powerful of the arts partly because sounds – more than any other sensory stimulus.
Rationale To encourage all students to take a full part in the life of our school, college, workplace or wider community. To provide opportunities to enable.
Artificial Intelligence 13. Multi-Layer ANNs Course V231 Department of Computing Imperial College © Simon Colton.
MICHAEL MILFORD, DAVID PRASSER, AND GORDON WYETH FOLAMI ALAMUDUN GRADUATE STUDENT COMPUTER SCIENCE & ENGINEERING TEXAS A&M UNIVERSITY RatSLAM on the Edge:
By Kayla Paige Click to Begin!. Try Again! The relationship of width to height in a picture or shape is: A. Crop Crop B. Aspect Ratio Aspect Ratio C.
Brian Merrick CS498 Seminar.  Introduction to Neural Networks  Types of Neural Networks  Neural Networks with Pattern Recognition  Applications.
Midterm 1 Wednesday next week!. Synthesize the Big Picture Understanding Brain-wide neural circuits Extracranial electrophysiology EEG/MEG Metabolic Imaging.
Artificial Intelligence (CS 461D)
Painting an experience? How aesthetics might assist a neuroscience of sensory experience Ron Chrisley Centre for Research in Cognitive Science and School.
Working with your Head to build an effective Leadership team.
Reinforcement Learning in Real-Time Strategy Games Nick Imrei Supervisors: Matthew Mitchell & Martin Dick.
CS320n –Visual Programming Interactive Programs Mike Scott (Slides 5-1)
Laurent Itti: CS599 – Computational Architectures in Biological Vision, USC Lecture 7: Coding and Representation 1 Computational Architectures in.
October 7, 2010Neural Networks Lecture 10: Setting Backpropagation Parameters 1 Creating Data Representations On the other hand, sets of orthogonal vectors.
Teaching Gifted Students NATIONAL ASSOCIATION OF SPECIAL EDUCATION TEACHERS.
Design. Design is an important aspect of the world in which we live and our everyday lives. Design focuses on the generation of ideas and their realisation.
Creative Software for the Creative Industries Dr. Simon Colton Department of Computing Imperial College, London.
AJITESH VERMA 1.  Dictionary meaning of chaos- state of confusion lack of any order or control.  Chaos theory is a branch of mathematics which studies.
The Theme For March 2011 Creativity or Drama In Black and White.
Assessing employability through reflective diaries on teamwork CEC 202 A Sense of Place School of English Second Year Approved Module.
Introduction to digiCOACH Empowering Instructional Leaders Common Core Edition.
The Art Of Listening Take out a sheet of paper and write a paragraph about what you think the difference is between hearing and listening.
Presentation on Neural Networks.. Basics Of Neural Networks Neural networks refers to a connectionist model that simulates the biophysical information.
Effective Public Speaking Chapter # 3 Setting the Scene for Community in a Diverse Culture.
EVALUATION OF OUR RADIO ADVERT By Chris. We had to think of a genuine idea instead of a ridiculous one. We had to decide which sound effects to use and.
Learning Styles 1. Visual Learning Style You prefer using images, pictures, colors, and maps to organize information and communicate with others. You can.
Introduced by to. Group Pattern Language Project 2662 Alder St, Eugene OR 97405
Low Level Visual Processing. Information Maximization in the Retina Hypothesis: ganglion cells try to transmit as much information as possible about the.
Towards Cognitive Robotics Biointelligence Laboratory School of Computer Science and Engineering Seoul National University Christian.
Prediction in Human Presented by: Rezvan Kianifar January 2009.
© CROZ Authors: Mirela Grginčić Business Solution Consultant Krešimir Musa Director of Consulting and Implementation Services at CROZ Dealing with information.
Music Intelligence. What it is It is ability to think in pattern, rhythms and sound It is ability to appreciate the music It is ability to compose different.
Facilitating Emergence: Matrix Group Dialogue. Outcomes:  Understand what Matrix Group Dialogue is and what need it meets in today’s world.  Experience.
1 ECE 517: Reinforcement Learning in Artificial Intelligence Lecture 17: TRTRL, Implementation Considerations, Apprenticeship Learning Dr. Itamar Arel.
Collaboration Development through Interactive Learning between Human and Robot Tetsuya OGATA, Noritaka MASAGO, Shigeki SUGANO, Jun TANI.
Computer Go : A Go player Rohit Gurjar CS365 Project Presentation, IIT Kanpur Guided By – Prof. Amitabha Mukerjee.
Applications of Neural Networks in Time-Series Analysis Adam Maus Computer Science Department Mentor: Doctor Sprott Physics Department.
Evaluating Network Security with Two-Layer Attack Graphs Anming Xie Zhuhua Cai Cong Tang Jianbin Hu Zhong Chen ACSAC (Dec., 2009) 2010/6/151.
THE DIGITAL CLOCKWORK MUSE Rob Saunders; John S. Gero Key Centre of Design Computing and Cognition The University of Sydney, NSW 2007, Australia
Multiple Intelligences Ways to learn. 2 Yesterday, we took a test to determine our “learning style” Yesterday, we took a test to determine our “learning.
Communication and Language. Listening and attention: Children listen attentively in a range of situations. They listen to stories accurately anticipating.
Independent Enquirers Learners process and evaluate information in their investigations, planning what to do and how to go about it. They take informed.
Neural Networks Presented by M. Abbasi Course lecturer: Dr.Tohidkhah.
Dr.Abeer Mahmoud ARTIFICIAL INTELLIGENCE (CS 461D) Dr. Abeer Mahmoud Computer science Department Princess Nora University Faculty of Computer & Information.
Cost effectiveness in Human Services: simple complicated and complex considerations Towards a method that is useful and accurate Andrew Hawkins, ARTD Consultants.
TRIALS AND TRIBULATIONS Architectural Constraints on Modeling a Visuomotor Task within the Reinforcement Learning Paradigm.
Games in Practice Prepared by RLEF, July Why Games? It’s recommended that a players development in each training session should consist of learning.
How to do it right….  Enhance Understanding  Add Variety  Support Claims  Have a Lasting Impact.
Projection and the Reality of Routines – reflections of a computational modeller Bruce Edmonds Centre for Policy Modelling Manchester Metropolitan University.
MODELS PURPOSE: Predict the future, test outcomes of various scenarios, identify the important components or variables, and understand how the parts interact.
October 9, 2006 Section 4: Why do an RIA Designing an RIA.
EVALUATING FOR GENERATIVE CHANGE: CITIZEN PARTICIPATION FOR ACCOUNTABILITY Australasian Evaluation Society (AES), International Conference, Friday 2 Sept.
7 common challenges in using theory of change - and how to address them Professor Patricia Rogers BetterEvaluation Royal Melbourne Institute of Technology,
Maestro AI Vision and Design Overview Definitions Maestro: A naïve Sensorimotor Engine prototype. Sensorimotor Engine: Combining sensory and motor functions.
Robot Intelligence Technology Lab. Evolution of simple navigation Chapter 4 of Evolutionary Robotics Jan. 12, 2007 YongDuk Kim.
Artificial Intelligence (CS 370D)
MULTIPLE INTELLIGENCE SELF-ASSESSMENT
MULTIPLE INTELLIGENCE SELF-ASSESSMENT
Designing Neural Network Architectures Using Reinforcement Learning
View Inter-Prediction GAN: Unsupervised Representation Learning for 3D Shapes by Learning Global Shape Memories to Support Local View Predictions 1,2 1.
August 8, 2006 Danny Budik, Itamar Elhanany Machine Intelligence Lab
Patterns by Category Group Works Intention Context Context Faith
ECE 517: Reinforcement Learning in Artificial Intelligence Lecture 17: TRTRL, Implementation Considerations, Apprenticeship Learning November 3, 2010.
Presentation transcript:

Finding aesthetic pleasure on the edge of chaos: A proposal for robotic creativity Ron Chrisley COGS Department of Informatics University of Sussex Workshop on Computational Models of Creativity in the Arts Goldsmiths College, May 16th-17th 2006

Background Goal: Design a robot/environment system likely to exhibit creative behaviour: –Novel (at least for the robot) –Of (aesthetic) value (for humans, if possible) Engineering approach: –No direct modelling of human creativity –But exploit what is known about creativity in humans (and animals?), when expedient –Allow for possibility that insights into the human case may accrue anyway Manifesto only: No implementation yet –Set of "axioms" –Assume case of musical output for examples

Underlying architecture Key: Recurrent Connection (Copy) Full Inter-Connection Between Layers Of Units Action Expected Sensations Predicted State Previous Predicted State (Context Units) D-map T-map

Underlying architecture CNM: –Recurrent neural network –Forward model of environment Learns to anticipate/predict the sensory input it will receive if it performs a given action in a given context In conjunction with motivators can enable the robot to select actions that carry an expectation of "pleasure"

Main idea Add new motivators, corresponding to two dimensions of creativity: –Value –Novelty Axiom 1: If you make your robot pleasure-seeking, and make creativity pleasurable, you'll make your robot creative

Value: Appreciation Axiom 2: To be a good creator, it helps to be an appreciator –The CNM should evaluate the output of itself and others –That is, it should be able to feel pleasure upon experiencing outputs –Use this to guide its creative process (action selection)

Value: Reality Axiom 3: Let the robot experience output in the real world, as we do –Avoids the input bottleneck Robot can learn all the time Learns reality, not our edited version of it –Increases likelihood of consonance between what we value and what it values

Value: In our image Axiom 4: We won ’ t like what it likes unless it likes what we like –Built-in motivators should resemble ours –E.g., a preference for integer frequency ratios

Value: Sociability Axiom 5: An important motivator is the approval or attention of others –Indirect: Preference for human proximity/input –Direct: Buttons on robot that allow listeners to provide approval or disapproval feedback

From Saunders, 2001

Novelty: Complexity Axiom 6: Sometimes it is better not to try pursue novelty directly, but something that is correlated with it –Prefer outputs on the subjective "edge of chaos": That almost, but not quite, elude understanding of that agent at that time –Pleasure of an output is a hump-shaped function of the effort required to predict it –Result: Sing-song and white noise are boring, but catchy tunes are not

Novelty: Dynamics Axiom 7: Let dynamics play a role in appreciation –Process is temporally sensitive in several ways: 1.Pleasure associated with "getting it" depends on how much time it took to get there 2.Even if earlier portions are unpredictable (=> not pleasurable), work as a whole can be if it is coherent 3.Since the system learns, what it finds challenging, but possible, to predict (= pleasurable) will change over time

Novelty: Self-appreciation Axiom 8: Patterns in one's own states can be the objects of appreciation –Will only be a path to novelty if agent has limited access to its own processes Can only change internal states indirectly, by changing world Uses model of its processes to predict its own behaviour, rather than using those very processes themselves

Novelty: Embodiment Axiom 9: The best way to make outputs in the real world is to be embodied in the real world –Avoids the output bottleneck Robot doesn’t require intervention for it to generate and appreciate –Allows for serendipity, in the space between expected and actual outcomes –Imposes naturalness relation, making some transitions non-arbitrary (value)

Implementation issues Intended platform: –Two AIBO ERS-7s Solution: – Translate bodily movements into sound Problem: – Disembodied sound generation

Thank you! Thanks to: Maggie Boden Rob Clowes Simon Colton Jon Rowe Rob Saunders Aaron Sloman Dustin Stokes Mitchell Whitelaw for helpful comments and discussions