Frequently Asked Questions   What does Korns Associates do? What is an agent? So what is an intelligent agent? What advanced technologies will appear in agents? How are Agent Information Server? agents different from Java agents? Who are Korns Associates' principals? Research Publications using AIS

Korns Associates is a privately held applied research company that develops sophisticated agent technology, development tools, and applications. It was founded in June 1993, and has pioneered the use of "intelligent agents" for securities investing using a business model wherein its research is self-funding. In 1999 Korns Associates created Investment Science Corp to support, incubate, and sell commercial applications of the Korns Associates technology. For a more detailed description of Korns Associates' technology, click here.

There are many different definitions of "agent" that are being used, and they are not consistent. Even more confusingly, many companies use the word "agent" to describe any program that does anything for anyone.

For our purposes, an "agent" is a program building block that is capable of at least a rudimentary level of autonomy. It contains more than just raw binary computer code. It contains information about itself, about the goals it plans to achieve, about the knowledge it has accumulated along the way. It has some rudimentary mechanism for being informed when an outside source attempts to communicate with it. It can travel from one database server to another across LAN's, WAN's, intranets, and the Internet.

In Agent Information Server?, "agents" are to intelligent systems what "cells" are to living systems - basic, self-contained building blocks, which form together in communities to form larger more complex systems. While each individual agent in an application may not be extremely sophisticated, a community of agents can quickly become very complex. Behavior may emerge from such a community of agents which is adaptive. Such emergent behavior may even appear "intelligent".

The term "intelligent agent" is no better defined than "agent" itself. It implies that the agent is using some reasonably sophisticated technology to guide itself and make decisions. In practice, other companies sometimes use the term simply to accentuate the presumed power of their programs (i.e., "it does something really interesting, so it's an intelligent agent").

Korns Associates is continuously phasing in new advanced technologies to its agent research, such as learning algorithms and negotiation protocols. As agents develop increasing sophistication, many agents will have more of a claim to calling themselves "intelligent".

4. What "advanced" technologies will appear in agents?

As mentioned above, Korns Associates plans on incorporating advanced technologies into its research:

• Learning algorithms, for agents that will adapt themselves to user routines and user preferences;
• Planning algorithms, for agents to automatically construct a sequence of steps that accomplish some high-level goal;
• Expert system technology, to allow programmers to describe how an agent should behave when it encounters certain situations on the Web;
• Negotiation protocols, that make it possible for agents to reach agreements on sharing resources, on buying and selling, and on scheduling (e.g., setting up a meeting for their human owners).

First, Agent Information Server? is a new style of distributed database for dynamic data and knowledge representation. Java is a general purpose Web programming language and not a database.

Currently, Agent Information Server?'s agents are written in an agent oriented dialect of Lisp and exist inside the Agent Information Server? repository at each distributed location on the Internet or Intranet.

Future development plans include support for Java and Smalltalk compilers within the Agent Information Server? repository. At that time presumably our "agents" would be Lisp, Java, and Smalltalk agents existing inside the Agent Information Server? repository as differentiated from Java "agents" existing outside our repository.

Korns Associates was founded by Michael F. Korns, and Gilda Cabral, who between them have over 45 years of experience in the software industry, as well as over 9 years of experience in research on agents.

Michael F. Korns serves as President of Korns Associates. He started his career working at IBM in Advanced Engineering. He has been Vice President Information Sciences at Tymshare Transactions Corporation, and Vice President Chief Scientist of Xerox Imaging Corporation. For over 35 years, Michael Korns has been an expert in converting academic research into commercial applications.

Analytic Information Server has aided a number of scientists in valuable peer reviewed research.
Here are the links to a number of peer reviewed scientific research papers for which AIS has provided either subject material, tools, or valuable assistance. All of these papers are published in peer reviewed scientific journals or peer reviewed books. These papers have receives hundreds of citations.
(Note: these downloads are all drafts only as we are not allowed to distribute the final printed papers. Please purchase the print versions from publishers to support continuing research.)

1. Korns, Michael F., 2015. Trading Volatility Using Highly Accurate Symbolic Regression. In Ryan, et. al., Handbook Of Genetic Programming Applications, New York, New York, USA. Springer.
2. Korns, Michael F., 2012. Predicting Corporate Forward 12 Month Earnings, 2012. Theory and New Applications of Swarm Intelligence, ISBN 978-953-51-0364-6, edited by Rafael Parpinelli and Heitor S. Lopes, InTech Academic Publishers.
3. Korns, Michael F., 2012. A Baseline Symbolic Regression Algorithm. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice X, New York, New York, USA. Springer.
4. Korns, Michael F, 2010. Abstract Expression Grammar Symbolic Regression. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice VIII, New York, New York, USA. Springer.
5. Korns, Michael F, 2011. Accuracy in Symbolic Regression. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice IX, New York, New York, USA. Springer.
6. Korns, Michael F. 2006. Large-Scale, Time-Constrained Symbolic Regression. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice IV, New York, New York, USA. Springer.
7. Korns, Michael F. 2007. Large-Scale, Time-Constrained Symbolic Regression-Classification. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice V, New York, New York, USA. Springer.
8. Korns, Michael F, 2009. Symbolic Regression of Conditional Target Expressions. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice VII, New York, New York, USA. Springer.
9. Korns, Michael F, 2009. Symbolic Regression Using Abstract Expression Grammars. In Proceedings of GECCO Genetic and Evolutionary Computation Conference, Montreal, July 2009. Association for Computing Machinery.
10. Korns, Michael F., 2015. Highly Accurate Symbolic Regression for Noisy Training Data. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XIII, New York, New York, USA. Springer.
11. Korns, Michael F., 2016. An Evolutionary Algorithm for Big Data Multiclass Classification. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XIV, New York, New York, USA. Springer.
12. Korns, Michael F., 2017. Evolutionary Linear Discriminant Analysis for Multiclass Classification Problems. In GECCO Conference Proceedings ‘17, July 15-19, Berlin Germany 2017.
13. Korns, Michael F., 2017. Genetic Programming Symbolic Classification: A Study. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XV, New York, New York, USA. Springer.
14. Korns, Michael F., May, Tim, 2019. Strong Typing, Swarm Enhancement, and Deep Learning Feature Selection in the Pursuit of Symbolic Regression-Classification. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XVI, New York, New York, USA. Springer.
15. Korns, Michael F., May, Tim, 2019. Feature Discovery with Deep Learning Algebra Networks. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XVIII, New York, New York, USA. Springer.
16. Korns, Michael F., and Truscott, Philip, 2011. Detecting Shadow Economy Sizes with Symbolic Regression. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice IX, New York, New York, USA. Springer.
17. Truscott, Philip, Korns, Michael F., 2015. Predicting Product Choice with Symbolic Regression and Classification. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XIII, New York, New York, USA. Springer.
18. Korns, Michael F., 2013. Extreme Accuracy in Symbolic Regression. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XI, New York, New York, USA. Springer.
19. Korns, Michael F, 2009. Mutation and Crossover with Abstract Expression Grammars. In Proceedings of World Summit on Genetic and Evolutionary Computation, Shanghai, June 2009. Association for Computing Machinery.
20. Korns, Michael F, 2009. Symbolic Regression of Conditional Target Expressions. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice VII, New York, New York, USA. Springer.
21. Truscott, Philip, Korns, Michael F., 2014. Explaining Unemployment Rates and Symbolic Regression. In Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice XI, New York, New York, USA. Springer.
22. Korns, Michael F., Nunez, Loreyfel, 2008. Profiling Symbolic Regression Classification. In Riolo, Rick, L, Soule, Terrance, and Wortzel, Bill, editors, Genetic Programming Theory and Practice VI, New York, New York, USA. Springer.

Korns Associates Korns Associates is a privately held R&D company that develops and uses sophisticated agent technology to build artificial intelligence applications for securities analysis and stock ranking. We were founded in June 1993, and have engaged in continuous software research and development of securities analysis software.