I_E | Insight Engine (2014)

Bill Seaman
research/learning system | demo of working prototype (shown above)
custom software written in HTML5/Javascript/PHP w/database scripted back-end.

technical description

Bill Seaman, principle investigator and media researcher
Todd Berreth, design consultant & primary developer
Olivier Perriquet, research consultant - mathematical modelling / computational linguistics




The Insight Engine seeks to draw on my long history as a media researcher designing new forms of interface and qualities of interactivity, and to expand this via a strong interdisciplinary collaboration that bridges Neuroscience, Computer Science, the Arts and Humanities at Duke as well as through international collaboration. Such a project reflects clearly the interdisciplinary goals of both DIBS (The Duke Institute for Brain Sciences who funded the work) and the research community at large, and in particular presents a multi-perspective approach to knowledge navigation and subsequent knowledge production. This research seeks to work toward the digital authorship of a tool to empower insight production, distributed interdisciplinary team-based research and to potentially enable bisociational processes as discussed by Arthur Koestler in The Act of Creation: "I have coined the term “bisociation” in order to make a distinction between the routine skills of thinking on a single “plane”, as it were, and the creative act, which, as I shall try to show, always operates on more than one plane…" "We learn by assimilating experiences and grouping them into ordered schemata, into stable patterns of unity in variety. They enable us to come to grips with events and situations by applying the rules of the game appropriate to them. The matrices which pattern our perceptions, thoughts, and activities are condensations of learning into habit […] The bisociative act connects previously unconnected matrices of experience…"

If we reverse engineer differing research communities across multiple disciplines we can assume that many researchers undertake similar practices— reading papers, viewing diagrams, exploring data sets, creating and viewing visualizations, annotating research materials, watching videos, and partaking in discussions among other activities. Interdisciplinary research also means crossing “linguistic” domains framing that research. Here the generation of shared language (developing bridging languages) is essential. Yet, could we make a new system that heightens the potential for insight and creative juxtaposition of essential ideas that cut across multiple research communities/domains?

The notion here is to explore Neuroscience/AI/Learning Systems/Neosentience (see Seaman and Rossler’s Book - Neosentience | The Benevolence Engine) through the associative “lens” of focused computational interactivity, functioning in the service of providing new insights and associations across interdisciplinary research fields, as well as exploring different concepts and foci from within individual research domains. In this instance artfully displayed interactive informatics represents the outermost level of the system. The web based interface enables a user-centric experience, “driving” the generation of a visual set of associative experiences —calling up different words, phrases, titles, images, videos, urls, and models as a network of potential associations that are brought into visual proximity. Such a work functions both on a local level in a visual installation, as well as on a laptop driven across the internet, forming a vast community of contributing researchers.

Outwardly, the initial experience is aesthetic and participatory in nature— the system is designed to be focused in different user-driven directions. Thus, though a network of “pre-seeded” choices (the first year of research) one can drive the system to focus on Neuronscience-only related topics of association. Alternately one can juxtapose texts and images from the arts and humanities — poetic texts, critical/social texts, texts related to ethics, or historical texts from multiple fields— this depends on the choices of the interactant. I describe this as a multi-perspective approach to knowledge production.

One can also select from a scrolling list of researcher/ topics. The media objects that populate the system’s database include applications, audio, databases, documents, drawings, images, video sections, quotations, models from multiple fields, and linked websites. Abstracts are also attached for each media object. Thus, we begin with a “seeded” database of diverse materials as contributed by this particular research community - Seaman contacted each of the researchers to participate in this first year of authoring the system, although as the system is opened up to others (and other research communities), these communities will grow more organically. This database can be added to in an ongoing manner. By digitally “scraping” the searched paper or key word references for media objects, the system can gather, store and enable different qualities of experience to be articulated, related to that information—the system enables instant access to the media objects. Thus the system functions as an iterative learning system. Multiple researchers with the same login can work in a distributed manner, making queries and or calling up past queries. One can create new “vortices” and store papers and/or media objects for future reference. A second menu system enables users to explore a series of lists of key words and concepts, as another associational trigger. These include key words drawn from the researcher’s own papers (in blue) as well as lists that Seaman developed to stimulate thought and insight. A random button lets the user generate random texts as derived from the lists. These key word / concepts can also be stored with the users profile. Thus all choices and use oof the system can be stored by individuals making up the research community.

All of the papers and media objects are statistically analyzed for word use and enable the the weighting and chooses that are derived through bisociational queries made by the user.

Thus the insight engine is a vehicle for developing a community of researchers who might not normally find each other, given particular publishing domains. It stimulates the generation of a community of communities and in particular enables research into ideas which cross disciplinary boundaries. To my mind this is where often the most exciting new forms of research take place.



 

download high-resolution video:
I_E_demo_video (h.264, 1728x1080) (12:45) (495.3 mb) (right click->Save Link As....)

screen captures of system (click for larger image) :

Copyright © 2014, Insight Engine by Bill Seaman. All Rights Reserved.