I just read the three articles by Prof. Ashok Goel (Georgia Tech):
Goel, A. K., Vattam, S., Wiltgen, B., & Helms, M. (2012). Cognitive, collaborative, conceptual and creative—four characteristics of the next generation of knowledge-based CAD systems: a study in biologically inspired design. Computer-Aided Design, 44(10), 879-900.
As I read this article, the first thing I thought of was that I needed to add R1/XCON and intelligent configurators and CAD systems to the list of instances in the brief history of cognitive systems.
Second, the blending of next generation computer-aided design and cognitive assistants made me think of Tony Stark’s computer assistant (in the “Iron Man” science fiction movie) and this implementation by Elon Musk: http://www.saratechinc.com/future-of-design/
Third, this article made me think more about the question – what is the difference between a “good tool” and a cognitive assistant? CAD systems may be “good tools” but if the user or even better multiple users can talk and gesture with the systems understanding the input and changing what is displayed, and the system can learn by being shown or asked to refer to relevant examples in the literature, and the system can provide levels of confidence in proposed solutions (see discussion thread “A Very Brief History of Cognitive Assistants”) – then we have from “good tool” to more clearly a cognitive assistant for engineers working on bio-inspired designs.
Next, I liked this section in the paper….
“Finally, the fourth C is creativity. It has been said many times
in the design literature that design can be routine, innovative, or
creative , even though these categories often are imprecise.
Brown and Chandrasekaran , for example, suggested that (1) in
routine design, both the basic structure of the desired system and
the plans for selecting the parametric values of each component
were known, (2) in innovative design, only the structure of
the system was known and the plans for selecting component
parameter values were unknown, and (3) in creative design, the
structure of the design itself was unknown. In our own earlier work
on case-based design [31,32], we have proposed that (1) in routine
design, the modifications needed to adapt a known design into the
desired design are limited to values of parameters of components
in the design, (2) in innovative design, the needed modifications
pertain to the components of the design, and (3) in creative design,
the modifications entail changes to the topology of the design
I have talked about incremental, radical and super-radical innovation as: (1) incremental innovation change the numeric values – for example 50 miles per gallon, instead of 45 miles per gallon. (2) radical innovations change the combinations of units of measure used to understand the innovation – for example, bits/joule for mobile phone communications, and (3) super-radical innovations change the units of measure themselves – for example, when e-Bay developed a measure of reputation for people using their system to sell items to others. As instances of types of systems are designed or evolve over time, we can see incremental, radical, and super-radical innovation examples.
Finally, I like the discussion of SBF (Structure-Behavior-Function) and the DANE system very much. I think cognitive assistants in some ways mirror textbooks. Cognitive assistants must include concepts, relationships, case studies, problems-solutions/questions-answers. So a good cognitive assistant should be able to operate in a mode that allows it to help users receive certifications for demonstrated competencies and skills. A good cognitive assistant should also be able to act as a personal coach or mentor for learners trying to become more competent or master a domain of study and practice.
Vattam, SS & Goel, AK (2013) Biological Solutions for Engineering Problems: A Study in Cross-Domain Textual Case-Based Reasoning. In S.J. Delany and S. Ontañón (Eds.): ICCBR 2013, LNAI 7969, Springer-Verlag Berlin Heidelberg pp. 343–357.
Again, this one made me think of what is the difference between a “good tool” and a cognitive assistant. Perhaps search engines need to be added to the list of a brief history of cognitive assistants. Textual Case Based Reasoning systems have the challenges of findability, recognizability, and understandability. The Biologue interactive system was an interesting exploration of some of these challenges in web-based retrieval of documents.
Goel, A, Zhang G, Wiltgen B, Zhang Y, Vattam, S, Yen, J (203) The Design Study Library: Compiling, Analyzing and Using Biologically Inspired Design Case Studies. In Design Computing and Cognition DCC’14. J.S. Gero (ed),
Springer. pp. xx-yy.
Regarding the Design Study Library (DSL) an interactive system that provides access to a digital library of case studies of biologically inspired design, I liked the two-level design of projects and documents. T-charts are also a nice innovation for comparing the problem domain and biological solution domain. An impressive set of case studies for a textbook/cognitive assistant are also presented. Prof. Goel’s classes are a rich source of data on learning too!
After reading the papers, I decided to do some searches on O*NET Online (Occupation Network On Line).… I searched for these keywords, and found the number of occupations interesting:
Manage – appears in 715 occupation descriptions
Communicate – 428 occupations
Design – 415 occupations
Engineer – 339 occupations
Collaborate – 253 occupations
As we think about building cognitive assistants for all occupations, it will be important to have cognitive systems components related to design. Building cognitive assistants from textbooks that define concepts, relationships, enumerate important cases studies, and assemble problem-solution, question-answer pairs, both correct and incorrect will be helpful. Prof. Goel’s articles, courses and systems are a gold mine of information. Also, O*NET is a good source of information about the tasks that professionals in an occupation must be able to perform – see http://www.onetonline.org/