Biography: Dr. Paul Werbos is best known (and most cited) for the original discovery of backpropagation, and for the theorem establishing its validity, as part of his PhD thesis in Applied Mathematics for Harvard in 1974. Even before 1974, he had developed backpropagation as one element of a more general approach to reinforcement learning (http://vixra.org/abs/1902.0046), which combined a new way to learn to approximate dynamic programming with key insights from Freud’s theory of how learning works in neurons of the brain. He inaugurated the field which we now know of as RLADP, Reinforcement Learning and Approximate Dynamic Programming, building on this earlier work, his later papers, and on the research area of Adaptive and Intelligent Systems at NSF which he led from 1988 to 2015. Dr. Werbos has been very active for decades in National Science Foundation and IEEE USA and made numerous contributions to the research society. |
|
Title: How to use the "new AI" in social networks to prevent human extinction Abstract Massive changes have already happened in the global internet. These changes have excited a wide variety of people in the policy community, the regulatory community and the investment community, who tend to depend on advisors who do not really understand the underlying principles which have only just started a massive change in human life. Any “expert” who claims to know whether these changes will raise humanity to a higher level, or cause the extinction of our species, is lying, because no one knows this yet. The best and deepest information now available, drawing on the best sources available to IEEE societies, clearly shows that both extremes are possible. The outcome will depend on choices made by those people who are willing and able to bridge the gap between more advanced, new solid, concrete systems design and the larger social networks and systems. We need to build our own social networks to do justice to this design task, developing the necessary pieces and -- above all -- the necessary new connections and integration, not only between apps but between people. This talk will try to provide a roadmap of the most crucial pieces, needs, and connections. The website build-a-world.org posts a few of the extensive new materials already available this year from IEEE, NSF and other deep scientific efforts on:
For example, the climate section builds on www.werbos.com/E/GridIOT.pdf, which described how a new approach to market design has demonstrated the value of a radically new approach, integrating optimization programs and new computer hardware with rational microeconomics and human players. This kind of Sustainable Intelligent Market Design can be extended to all sectors of the new Internet of Things (cyberphysical global internet), if the IEEE community can rise to the new challenges for fundamental crossdisciplinary research and development required. (See http://www.werbos.com/How_to%20Build_Past_Emerging_Internet_Chaos.htm ). This will also require new computer hardware, like the new quantum RLADP (optimization) technology discussed in my abstract for WCCI2022, posted at build-a-world.org. Most important for SMC, it will require deeper integration of human feedback and human potential into a new type of social network with more recurrence, integrating RLADP and natural language capabilities, and drawing much more on a new understanding of human “qi” and neuroscience, drawing on our new understanding of how brains work (Werbos and Davis), on the “Nerves of Government” (Deutsch), on new research and on emerging new technologies in physics. As the internet/AGI/IOT becomes the master integrator of all new modes of production, there are many modes of fatal instability possible, which can be avoided only by new work on integration and understanding, creating harmony between aspects of human life whose separation was sustainable only in olden times. It will of course require developing an upgrade to build-a-world.org, with more of the supporting material, open international access and structured but open dialogue. |
|