Synthetic Ecosystems for Learning about the Living World

Guide to the Project Proposal

Proposal description in HTML

Proposal descriptive text in Word format

Curriculum Vitae of Principal Investigators and Co-PIs

Board of Advisors

Other Participants

Preliminary Schedule and Budget

NSF LIS Call for Participation

Fig 1: Nerve Garden Prototype
see more views at Nerve Garden growth sequences
see our paper on the project

Synthetic Ecosystems for Learning about the Living World

When we as children were presented with the gift of a terrarium, aquarium, ant farm or chemical garden, we were given much more than just a toy. These were dresser-top universes, which, together with books, were complete learning systems that had the power to transform young minds, and inspired many to pursue careers in science or engineering.

An emerging confluence of technologies has made it possible to bring this kind of powerful learning environment to millions of students with access to the Internet. Three dimensional scene description through VRML (Virtual Reality Modeling Language), behaviors embodied in Java from neural networks to fuzzy logic systems, biological and ecological models from L-Systems, genetic programming, and cellular automata-based agents can now be combined to bring the virtual terrarium into existence.

This program will establish a Center for Collaborative Research on Learning Technologies (CRLT) to undertake a large collaborative development effort to bring the virtual terrarium and other environments we term synthetic ecosystems to the classroom and students at home. Projects within this CRLT will enable significant technology transfer from academic and commercial research to educators, and develop resources for curricula development including teacher training. Our own research on the pedagogical effectiveness of these novel environments will provide insight into what may become a 21^st Century learning system. Significant contributions and cost sharing by commercial entities will magnify NSF support and provide valuable program management input. Wide deployment of Synthetic Ecosystems by principal investigators from K-12 through college and university curricula will provide participating entities and the NSF a positive public benefit through parental and media awareness of the program.

Fig 2: L-Systems generated by Laurens Lapre

Project background, unique characteristics

This program will design and deploy a series of synthetic ecosystems and web-based resources to support learning about living systems. These simulated ecosystems will draw from the experience of many existing scientific systems such as Tierra, Sugarscape, and Polyworld as well as commercial systems such as SimAnt™, Creatures™ and William Latham's Organic Art. Synthetic Ecosystems will also open up new territory by:

• being delivered entirely through the Internet to commonly available computers at modem speeds,
• employing multi-user technologies, so that geographically separated students, teachers, researchers and the interested public can meet within the virtual environments to assess the results of experiments or design new ones, and by
• using a flexible and open interface to enable students to modify and share genetic structures, rewire behavioral models and set the parameters of the ecosystems they bring into being.

Acting as a repository for the collection of knowledge about form and function in Nature, scientists will be able to use Synthetic Ecosystems as a mechanism to deliver compelling research insights more quickly and clearly to a wider learning audience.

Fig 3: Tierra and Tierra in VRML
see the Tierra Home Page

Tools for Educators

For example, a biology teacher could supplement a lab experiment on bacterial colonies with a simulated population of Myxobacteria embodying the latest understanding of chemical signaling mechanisms for self organization. Using recursive L-system plant models, mathematics instructors could illustrate fractals and other forms of self-similarity. Fine arts teachers could employ genetic algorithms available through Synthetic Ecosystems to delve into forms of artistic beauty that find their source in the natural world. Computer Science teachers could excite students with projects in 3D virtual worlds while imparting principles of object-oriented design and messaging in multi-agent cellular automata systems. Social sciences teachers and economics instructors could use simple virtual ecosystems to demonstrate principles of resource scarcity, wealth accumulation, population dynamics, competition, and conflict.

Extensive on-line tutorials and experiment templates will be designed to allow teachers and students from K-12 through to colleges and university undergraduate levels to include Synthetic Ecosystems modules in their existing curricula. Formal survey and assessment methods will be used to evaluate the effectiveness of these systems in learning. In the end, students will be provided the opportunity to gain deeper understanding and increased curiosity about relationships between nature, science, and technology.

Fig 4: Mutating Organic Art by William Latham at Computer Artworks

A Collaborative Learning Space

Using standard protocols and common cross-platform technologies, all Synthetic Ecosystems would be available for collaborative learning experiments between classrooms, universities, corporate research environments, and students at home. Projects could be structured with the best resources and participants, rather than with what is available within one classroom or school. Learners will be able to collaborate in these virtual environments and not just be individual users. Student teams comprised of participants from different backgrounds (biology, computer science, mathematics, art, social studies) could work together to design experiments in different Synthetic Ecosystems. Students could reach out through the Internet to collaborate with students at other schools, professional designers of virtual environments, or university and corporate researchers. Students would thereby gain valuable project team and collaborative experience and build networks of contacts that would aid them in career path decisions.

As a CRLT center, Synthetic Ecosystems will draw from the experience of other centers, including: the High School Science Virtual Machine Laboratory at Boston University, the Center for Intelligent Multimedia Instructional Systems at the University of Massachusetts, Amherst, and Design-Based Learning of Vertebrate Biology, Comparative Anatomy and Physiology of Virtual Creatures at Stanford University. Synthetic Ecosystems will complement these programs, which concentrate on simulation for learning in mathematics, physics, chemistry and biological dissection, with additional virtual environments for biology, mathematics, social sciences, economics and the arts.

Fig 5: A-Life and CA Systems
see the Artificial Life Online

Suitability of Methods

A generation of children have been raised on navigation in 3D computer spaces and interaction with software objects exhibiting a variety of sophisticated behaviors. Even students with no prior experience of these environments tend to pick up the skills quickly. Since 1995, the emergence of a large number of multi-user Internet virtual worlds with a combined population of 320,000 users has built a tremendous infrastructure that can deliver a rich collaborative experience to desktop computers. Children have been featured highly in these environments, building 3D cityscapes containing millions of objects. The Biota.org special interest group of the Contact Consortium has been working to bring science education into these environments and has constructed a prototype virtual garden using VRML 2.0 and web-based delivery. It is on the basis of this experience and an extensive survey of the building blocks for synthetic ecosystems that we are now seeking to move ahead with this program.

Fig 8: User Visiting Nerve Garden L-System area

Synthetic Ecosystems will be based upon the research findings, software engineering experience, and interface methodologies from the existing work of its principal investigators. The ecosystems themselves will be created by plugging existing simulation engines into standard interfaces such as VRML and Java and constructing extensive web-based companion documents. The resulting educational assets will be served to schools, colleges, universities and the general Internet community and their effectiveness will also be evaluated on-line. In-person tutoring of teachers and students will be provided in some cases.

Taking Risks, Breaking New Ground

The high risk aspect of this program involves the novel combination of numerically and graphically-intensive computer software methods into an application which can be delivered to commercially available desktop systems through narrow bandwidth constraints. Working with our participating principal investigators and Co-PIs, members of our board of advisors, and commercial sponsors, we will work to overcome these hurdles.

A Broadly Interdisciplinary Effort

Principal Investigators and Co-PIs	Board of Advisors	Participating Companies
Bruce Damer, Contact Consortium, PI and Program Manager	Tom Ray, ATR	Silicon Graphics, Inc. Fujitsu Interactive, Inc.
Iain Miller University of Cincinnati, PI	Przemislaw Prusinkiewicz, University of Calgary	Live Picture, Inc., Onlive Technologies
Clark Guest UC San Diego, PI	Murray Turoff, NJIT	dFORM Internet Design
Carol A Kerney, PI, San Diego Schools, Apple Classrooms of Tomorrow	Roxanne Hiltz, NJIT	Construct Internet Design
Ajaz R. Rana and Jerry Fjermestad, NJIT, Co-PIs	Bruce Blumberg, MIT Media Laboratory	Paragraph International
Demetri Terzopoulos, PI University of Toronto		Intervista Corporation, British Telecom

The Contact Consortium is a non-profit research and educational corporation founded in 1995. Consortium member companies, universities, schools and other non-profit institutions are engaged in bringing powerful virtual environments to a wide Internet audience. Prior to submitting this pre-proposal, Consortium working groups developed prototypes to test the viability of Synthetic Ecosystems. Nonlinear dynamic systems embodied by neural networks (Damer), were combined with L-system formal methods (Prusinkiewicz) and geometric hierarchies employing inverse kinematics and proximity sensors as shown in fig. 1 below (Silicon Graphics, Intervista), to produce an environment called nerve garden. Nerve garden, shown in fig. 2, was placed onto the Internet in July 1996 and has served as an effective proof of concept for this program.

NSF support would enable us to integrate fuzzy dynamical maps (Guest), methods from the digital evolution models of Network Tierra (Ray), models and behaviors of artificial animals (Terzopoulos), Ethologically inspired models of action selection and learning (Blumberg), artificial intelligence and expert systems database methods (Fujitsu Interactive), and design of communication structures to be used for collaboration by a large interdisciplinary group (Turoff) into a compelling cross-disciplinary educational environment. Improving Internet infrastructure, the ubiquity of Java and VRML, and increasing computing power of end user systems over the period of this program will all converge to bring compelling and powerful experiences of Synthetic Ecosystems to millions of students.

Equally important to the formal scientific and technological methods will be the development of effective Synthetic Ecosystems course materials, integration with existing curricula and the evaluation of the effectiveness of the environments as learning systems. NSF support will provide resources to our investigators and participating organizations to introduce the environments and web-based course materials (dFORM Internet Design) at the college and university undergraduate level (Miller), in K-12 education (Kerney), and the construction of an on-line survey system for the evaluation of the effectiveness and impact of the environments on learning (Fjermestad, Rana, Turoff, Hiltz). Kerney brings significant test bed facilities to this program including an Apple Classrooms of Tomorrow center and other existing school projects. In addition, through Rocky Harris and other teachers in Portland, Oregon schools, high school students will be involved as summer and yearlong interns.

In summary, this broadly interdisciplinary program will draw on research from both academic and corporate research serving to enrich both communities while providing a unique and powerful new learning environment to a large number of educational institutions in the United States and worldwide.

Cost Sharing and other Benefits of Participation by Commercial Entities

Prototyping efforts and technology developed in 1995-96 constitute a $200,000 value that will be contributed to this program. Additional cost sharing has been committed to this program by commercial enterprises who will augment NSF's support with finances, development resources, equipment and managerial support. Involvement of commercial entities will multiply the benefits of the NSF support. Wide commercial backing will provide critical program management resources while the success of Synthetic Ecosystems will yield a positive public image benefit back to participating commercial entities.

Fig 6: Flocking and swarming behaviors
see Craig Reynold's Boids

The Investigators and Their Qualifications

This program is based on a four-part foundation of specialties in biology, computer science, networked education and online virtual environments. Members of this program include research scientists from academic institutions, instructors from secondary schools, colleges and university undergraduate programs, and, working with the Program Manager, several commercial entities and independent developers of Internet virtual environments. Coordinating the program will be a director of the Contact Consortium, a non-profit corporation which serves as a primary forum for multi-user virtual environments on the Internet. The Consortium has thirty corporate members and sponsors, and hosts numerous research and educational programs including an annual conference. Biota.org is a special interest group of the Consortium and constitutes the internal organizing force behind this proposal.

Fig 7: PolyWorld Artificial Life System and Computational Ecology
see the PolyWorld home page

The Effect of the Infrastructure on Science, Engineering and Education

Synthetic Ecosystems could serve as a proving ground for other collaborative scientific and educational environments which are considering using 3D virtual environments on the Internet. The merits of the two-dimensional, document-based World Wide Web were proven by part of the scientific community and later adopted by virtually all of science, education, and engineering. In the same way, this program could prove the scientific and educational utility of collaborative three dimensional spaces. In addition, the combination of genetic programming, neural networks, fuzzy logic systems and other methods with Java and VRML may well provide valuable spin-offs for scientific and commercial applications of 3D virtual spaces.

Synthetic Ecosystems will yield insight into how students form and apply multiple cognitive representations of complex systems, and how they are subsequently used to tackle real world problems. Large freshman classes in Biology and smaller K-12 classes will provide the opportunity for a comparative statistical evaluation of the pedagogical effectiveness of the new virtual learning environments.

We believe that Synthetic Ecosystems will give future direction to other virtual world laboratories for learning and scientific uses. By enabling scientists to disseminate their findings directly to learners in a shared, collaborative space, Synthetic Ecosystems creates the foundation of future virtual repositories and laboratories. One day, researchers may be able to use these environments to work with models and behaviors which would be difficult to represent, understand or manipulate in any other medium.

Preliminary Project Schedule

Year 1: Design of Synthetic Ecosystems, prototype implementation, evaluation system design

Year 2: Implementation and deployment of working systems, and on-line evaluation system

Year 3: Refinement and evaluation of deployed systems, determination of necessity for continuance

Funding Requirements

We are seeking approximately $800,000 annually for three years to support the design, deployment and evaluation of Synthetic Ecosystems. Cost sharing by participating commercial entities will augment this with another $300,000 annually in contributed equipment, finances, engineering, and managerial hours. Funding will be divided among participating PIs and the program managing organization, which is responsible for final integration of PI knowledge bases, methods and technologies, and promotion of the use of the environments in education. A preliminary budget breakdown is presented here.

Table 1: Preliminary Budget
*Budget items will be further augmented by cost sharing from participating commercial entities. Areas indicated will be augmented up to or exceeding $300,000 annually.

Fig 9: Overhead view of digital city built in 3D by 100,000 users

End of Grant Pre-proposal

Cambrian Organisms Courtesy of Prof. Kerry Clark, Florida Inst of Technology, © K B Clark 1995. Visit the Metazoa website Burgess photo courtesy Yoho-Burgess Shale Research Foundation Burgess image courtesy Hooper Virtual Paleontological Museum