Cognitive Systems for Nature Inspired Creative Design

248 views

Published on

A short talk on

Published in: Design
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
248
On SlideShare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide
  • Our next step was to develop a software platform that addresses the issues of findability and recognizability. We call our platform Biologue. It is an online social bookmarking platform aimed at promoting the sharing of references to biology articles among the BID community. It is similar to Connotea and CiteULike in spirit. The only major difference is the Biologue comes with a built-in Structure-Behavior-Function ontology. In addition to keyword-based tagging, it allows users to semantically tag articles with domain-bridging abstractions.
    We have chosen SBF ontology because it a well established ontology and because it subsumes most of the higher-level categories mentioned earlier like functions, mechanisms, principles, etc. and decently capable of providing domain-bridging semantics because its roots are in analogical design. Plus its extensible, and in our control. We are using SBF just as an example to test our hypothesis, we can switch to a better ontology if one comes along.
    If you look at the interface, users can bookmark articles on the left-hand side (a). On the right hand side they can tag portions of the articles with functions, behaviors, etc (b and c). Multiple users can tag the same article. Through collaboration and aggregation of tagging activities of multiple users, a partially-structured SBF model of the system discussed in the article emerges.
  • Our next step was to develop a software platform that addresses the issues of findability and recognizability. We call our platform Biologue. It is an online social bookmarking platform aimed at promoting the sharing of references to biology articles among the BID community. It is similar to Connotea and CiteULike in spirit. The only major difference is the Biologue comes with a built-in Structure-Behavior-Function ontology. In addition to keyword-based tagging, it allows users to semantically tag articles with domain-bridging abstractions.
    We have chosen SBF ontology because it a well established ontology and because it subsumes most of the higher-level categories mentioned earlier like functions, mechanisms, principles, etc. and decently capable of providing domain-bridging semantics because its roots are in analogical design. Plus its extensible, and in our control. We are using SBF just as an example to test our hypothesis, we can switch to a better ontology if one comes along.
    If you look at the interface, users can bookmark articles on the left-hand side (a). On the right hand side they can tag portions of the articles with functions, behaviors, etc (b and c). Multiple users can tag the same article. Through collaboration and aggregation of tagging activities of multiple users, a partially-structured SBF model of the system discussed in the article emerges.
  • Our next step was to develop a software platform that addresses the issues of findability and recognizability. We call our platform Biologue. It is an online social bookmarking platform aimed at promoting the sharing of references to biology articles among the BID community. It is similar to Connotea and CiteULike in spirit. The only major difference is the Biologue comes with a built-in Structure-Behavior-Function ontology. In addition to keyword-based tagging, it allows users to semantically tag articles with domain-bridging abstractions.
    We have chosen SBF ontology because it a well established ontology and because it subsumes most of the higher-level categories mentioned earlier like functions, mechanisms, principles, etc. and decently capable of providing domain-bridging semantics because its roots are in analogical design. Plus its extensible, and in our control. We are using SBF just as an example to test our hypothesis, we can switch to a better ontology if one comes along.
    If you look at the interface, users can bookmark articles on the left-hand side (a). On the right hand side they can tag portions of the articles with functions, behaviors, etc (b and c). Multiple users can tag the same article. Through collaboration and aggregation of tagging activities of multiple users, a partially-structured SBF model of the system discussed in the article emerges.
  • Our next step was to develop a software platform that addresses the issues of findability and recognizability. We call our platform Biologue. It is an online social bookmarking platform aimed at promoting the sharing of references to biology articles among the BID community. It is similar to Connotea and CiteULike in spirit. The only major difference is the Biologue comes with a built-in Structure-Behavior-Function ontology. In addition to keyword-based tagging, it allows users to semantically tag articles with domain-bridging abstractions.
    We have chosen SBF ontology because it a well established ontology and because it subsumes most of the higher-level categories mentioned earlier like functions, mechanisms, principles, etc. and decently capable of providing domain-bridging semantics because its roots are in analogical design. Plus its extensible, and in our control. We are using SBF just as an example to test our hypothesis, we can switch to a better ontology if one comes along.
    If you look at the interface, users can bookmark articles on the left-hand side (a). On the right hand side they can tag portions of the articles with functions, behaviors, etc (b and c). Multiple users can tag the same article. Through collaboration and aggregation of tagging activities of multiple users, a partially-structured SBF model of the system discussed in the article emerges.
  • Cognitive Systems for Nature Inspired Creative Design

    1. 1. Cognitive Systems for Nature Inspired Creative Design Ashok K. Goel Design & Intelligence Laboratory, School of Interactive Computing, Center for Biologically Inspired Design IBM Cognitive Systems Webinar, December 2014
    2. 2. Cognitive Science Artificial Intelligence Human-Centered Computing Cognitive Systems 1985 1995 2005
    3. 3. Design Thinking Computational Creativity Systems Thinking Analogical Thinking Visual Thinking Meta Thinking Abductive Thinking
    4. 4. Research Faculty Rugaber Faculty Goel
    5. 5. Biomimicry or Biologically Inspired Design Problem-Driven Design: Design of the nose of Shinkansen 500, the Japanese bullet train, imitating Kingfisher’s beak
    6. 6. Solution-Based Design Example: Design of windmill turbines mimicking the tubercles on the pectoral flippers of humpback whales Frank Fish, Liquid Life Laboratory, West Chester University
    7. 7. DILab Research Methodology for Studying Biologically Inspired Design Information Services and Computational © Copyright 2013 Georgia Institute of Technology 7 4. Develop 4. Develop 4. Develop Information Information Services Services & Computational and Computational Tools Tools Tools 1. Observe Design Practices 2. Develop Cognitive Models 3. Develop Pedagogical Techniques 5. Evaluate (Formative, Situated) Cognitive Theory Design Theory
    8. 8. Situated Analogy 8
    9. 9. Eight Cognitive Challenges of Biologically Inspired Design 1. Communication across disciplines 2. Understanding biological and technological systems 3. Understanding the design processes 4. Specifying problems 5. Searching for biological analogies 6. Evaluating biological analogies 7. Analogical transfer 8. Evaluating design solutions
    10. 10. Cognitive Challenge Our Solution 1. Communicating across disciplines Shared language 2. Understanding systems Model schema 3. Understanding design processes Library of case studies 4. Specifying problems Problem schema 5. Searching for analogies Search engine for the web 6. Evaluating biological analogies Matching and mapping 7. Analogical transfer Design patterns 8. Evaluating design solutions Modeling, Simulation
    11. 11. SBF Model of Superhydrophobic Effect of Lotus Leafs Function: Self Clean Of: Lotus Leaf Contaminants ▪ Location: on leaf By-external stimulus: Drop of rain falling on the leaf By-function: Cause Superhydrophobic Effect Of: Leaf Water droplet ▪ Location: near contaminants ▪ Shape: spherical By-function: Make Water Droplet Roll over contaminants Of: Leaf By-function: Reduce area of contact b/w contaminants and leaf surface Of: Nano bumps By-structural-constraint : Nano bumps on leaf surface By-function: Absorb particle Of: Water droplet By-structural-constraint: (Force of absorption > Forces between particles and leaf surface) Contaminants ▪ Location: on water droplet By-function: Make water droplet roll beyond the edge Of: Leaf Water droplet ▪ Location: not on leaf Contaminants ▪ Location: not on leaf Function: Cause Superhydrophobic Effect Of: Leaf Water droplet ▪ Location: falling ▪ Shape: -NA- By-function: Make surface non-wettable Of: Nano bumps By-structural-constraint: Nano bumps on leaf surface By-domain-principle: Young’s equation Water droplet ▪ Location: on leaf ▪ Shape: spherical ▪ Contact angle (q ) > 120° gL,V gS,L q gS,V Function: Make Water Droplet Roll Of: Leaf Water droplet ▪ Location: x ▪ Inertial mass: M, Mass: m, ▪ Composite drag: D q x By-structural-constraint: Incline of the leaf By-structural-constraint: Spherical shape of water droplet By-principle: Laws of motion on inclined plane Water droplet ▪ Location: y v(y) = √(2y (-mg sinq - D) / M) y x Behavior: Self Cleaning Behavior
    12. 12. DANE (Design by Analogy to Nature) http://dilab.cc.gatech.edu/dane/ 105,000 hits, > 3400 unique users 1. Explicit representation of functions
    13. 13. Explicit Representation of Mechanism
    14. 14. Explicit Representation of Problem Decomposition
    15. 15. Finding biological analogues on the web: Challenges Findability Recognizability Understandability © Copyright 2013 Georgia Institute of Technology 15
    16. 16. Biologue !"#$ !%#$ !&#$
    17. 17. Findability © Copyright 2013 Georgia Institute of Technology 17 Results from Biologue: Findability
    18. 18. Recognizability © Copyright 2013 Georgia Institute of Technology 18 Results from Biologue: Recognizability
    19. 19. IBID (Intelligent Biologically Inspired Design) Accessing biology articles from the web via natural language processing
    20. 20. IBID (administration)
    21. 21. IBID (ontology)
    22. 22. BIDE: Interactive Design Environment Visual Sketchpad © Copyright 2013 Georgia Institute of Technology 22 Specifying Design Problems Biological Analogies Understanding Biological Systems Evaluating Analogies Evaluating Analogies Digital 4. Evaluation Notebook Modeling and Simulation Pattern Pattern Abstraction CCaassee S Sttuuddieiess Abstraction
    23. 23. Acknowledgements © Copyright 2013 Georgia Institute of Technology

    ×