IPv6 and the Future of Learning
By Dr. Henry Kelly
President, Federation of American Scientists

There's no hiding the challenges facing education systems in the US. The definition of a "basic education" keeps expanding, the people needing educational services are increasingly diverse, and funds are increasingly tight. Innovation is essential to meet these challenges and modern information technologies, powered by IPv6, have a key role to play. These technologies make it possible to provide more people expertise in more areas at lower cost – and do it in a way that is much more personalized, engaging and compatible with the frantic lifestyles of modern Americans. These new learning technologies will not emerge under "business as usual." New public and private institutions must be invented to exploit them.

Modern international research networks feed on each other, creating powerful positive feedback that continuously expands the scope and power of what we know. Modern engineering converts this knowledge into devices and systems of staggering complexity. But, this new knowledge and these new devices and systems are useful only if the knowledge can be assimilated and properly applied to address relevant needs.

In addition to sophisticated technical skills, employers increasingly demand people capable of working effectively across disciplinary lines and innovating products and services. Employers seek employees who can think strategically, manage multiple resources and solve complex problems. Today's workers must be able to track down information from unfamiliar sources, work with teams, analyze information and make decisions based on incomplete information. It's difficult to see how the need to develop workers with these skills can be met affordably without dramatic changes in the way information and experience are acquired.

Fortunately it seems possible to use the same information technology tools that many modern service companies have used to improve the quality of their services to meet these new needs. Powerful simulations developed for science and engineering and for entertainment and video games make it possible to create synthetic environments, avatars and synthetic equipment that can be used to facilitate both education and training concepts. Software tools designed to continuously track and evaluate users of websites can develop sophisticated measures of a student's progress and capabilities. Systems designed to provide help and instruction for consumers and company employees have developed capabilities that can be used to address a learner's questions rapidly and accurately.

These new technology tools seem tailor made to implement the insights of modern research on learning:

• Bridging theory to practice – simulations and virtual environments allow players to build, experiment, operate equipment, explore and practice without consequence of failure
• Developing high performance teams – collaborative environments permit team training for distributed, multi-operator problems and allow individuals and teams to practice interacting and coordinating tasks
• Meaningful evaluation and certification – the test measures something obviously important and embodies the challenge
• Access to expertise – inquiry and question asking stimulates learners to ask questions and provide fast, appropriate response; collaborative environments connect learners to peers, teachers and experts
• Complex problem-solving & strategic thinking – simulations allow learners to train for rapidly evolving, ambiguous scenarios that require managing large amounts information quickly and efficiently

Building learning systems capable of implementing these approaches to learning, however, is neither cheap nor easy. A hypothetical system to provide these services is outlined in Figure 1. Such a learning system will require pulling together a heterogeneous set of information and human connections on very short notice. A student or worker asking a question may get an automated answer that might include a video clip, an animation, or a connection to a live instructor. A large number of students, experts, tutors and multimedia data must be easily identified and connected to achieve this. And the learning system must support continuous dynamic changes.

Figure 1

The difficulty of building and maintaining such learning systems is daunting. It is increasingly likely that future systems will not only operate dynamically but will be built dynamically. This will require defining a rough set of strategies and interfaces that a heterogeneous collection of developers can agree to use. It would, for example, be necessary to agree on a common format for simulation interoperability, student recordkeeping and strategies for dispatching question and answers.

Figure 2

Given the right incentives, a collection of tools and associated learning content can be contributed by individuals and organizations and assembled quickly into instructional systems. A complex simulation could, for example, be created by assembling components built by different groups. This type of distributed, dynamic learning system would depend essentially on the ability to move large numbers of copies of large files around the world and on sophisticated networks of heterogeneous systems that can be assembled quickly. These would be greatly facilitated by the IPv6 systems.
A hypothetical system for achieving this is outlined in Figure 2.

A number of collaborative efforts of this type have been extraordinary successes – systems like Linux, the Wikipedia and the commercial site Second Life. But a number have also failed. There is much to be learned about how to provide the right rewards and incentives to keep such systems growing and operating. And many questions remain to be addressed, including the mixture of freely contributed materials and protected intellectual property and ways to reduce development costs and facilitate interoperability. To address these issues, we will need partnerships involving all the stakeholders – industry, academia and non-profits.

A key question, of course, is whether such learning systems will actually improve learning outcomes and meet our workforce development needs. Unfortunately, few groups have been able to assemble the financial and human resources needed to undertake the scale of efforts needed to build such powerful systems. But the results of the early investments are tantalizing. Dr. Wes Regian, Air Force Labs, reports average improvement of one standard deviation, compared to classroom instruction (equivalent to one grade level improvement). Carnegie Mellon University reports that their Intelligent Tutoring Systems (ITS) for algebra and geometry require 1/3 less instruction time than comparable classroom studies and result in a one standard deviation performance improvement. The Air Force has developed an IT based system to train technicians to repair avionics and reports that a student spending 20 hours with the system demonstrates expertise equivalent to four years of on-the-job experience.

The need for increased federal support for research, demonstration and testing in this critical area is obvious and has attracted strong bipartisan interest in the Congress. The benefits of such an effort are national and not regional and only the federal government is in a position to mount an effort at an adequate scale. Once systems are developed and tested, however, it should be easy to take proven systems and tools and adapt them to specific regional or corporate needs. But a concerted effort is needed now to launch the process.