By John Rogers
Fledgling New-York firm Intelligenesis Corp has developed a radical software architecture that it claims will change the face of artificial intelligence and help turn the internet into an organized, collective mind of sorts. The company’s Java-based Webmind product, due for commercial launch early next year, is a server-side, platform-independent software system that employs intelligent agents to absorb and analyze information from intranets, data feeds and the internet at large.
Webmind supports thin clients and Windows-based GUIs. It is touted as the first technology to enable the utilization of unstructured text-based data in quantitative analysis. The result is an artificial intelligence architecture that can essentially organize and manage any desired information on the internet, creating a World Wide Brain, as the company describes it, that can process infinite amounts of text and numerical data in the same way a human mind would.
Webmind is designed to pull networked data into its internal structure and transform it into a dynamic, continuously evolving semantic network, Intelligenesis says. Even as that may sound like science fiction, Webmind could prove invaluable to financial institutions that struggle to synthesize market news, data and research into an effective strategy for trading securities or commodities. While many current trading systems make use of search engines or data mining technology to crunch numbers, Webmind’s text-crunching ability will take quantitative analysis to the next level, according to Intelligenesis.
The privately-funded company was formed in early 1997 by academic Dr Ben Goertzel and Wall Street analyst Lisa Pazer after Pazer came across a web site outlining Goertzel’s research in the area of artificial intelligence and his vision of applying it to the internet. The combination offered the tantalizing opportunity to take Goertzel’s ideas and test them in the no-nonsense environment of Wall Street where, if the technology succeeded in an automated trading model it could yield great financial rewards for bankers and Intelligenesis alike.
The value proposition is that Webmind offers the combined functionality of text and document retrieval, data mining, OLAP, data-modeling, knowledge management and financial analytics. The company recently offered a glimpse of its technology at the Financial Technology Expo in New York as it attempted to drum up interest among potential customers. The underlying idea is that firms will be happy to pay for anything that promises an edge in the financial markets. And if Webmind can establish itself at a handful of firms, Pazer says, it could soon be perceived by Wall Street as something crucial for keeping up with the pack. In early testing, the performance of five optimized trading models using only conventional numeric indicators was compared to the output of the same systems boosted by the textual dataset Webmind offers.
Intelligenesis claims that the addition of its technology enhanced profitability rose anywhere from 300% to over 1,100%. It also sharply reduced the overall number of trades while increasing the number of winning trades from an average of 57% to roughly 84%. Intelligenesis expects to have Webmind ready for commercial use by the end of January.
Pazer says the company has canceled plans for a shrink-wrapped knowledge management product, noting nothing ‘knowledge management’ is making money anyway, and will instead follow a partnership strategy in which it hopes to become the dominant supplier of software intelligence to the internet economy. Under the new business plan, Intelligenesis will supply Webmind to partners such as financial services firms, online information vendors, and network management services providers, who will in turn supply the information, technology platform, customer base, and support infrastructure.