Project ISIS (Intelligent Speech for Information Systems) is the first project conducted under the Joint Center for Intelligence Engineering (JCIE) between the Peking University and The Chinese University of Hong Kong. It strives to address research issues at the technological forefront of today's information economy. We choose to focus on the ease-of-access to information and information services by anyone, from anywhere and at anytime.  In this regard, speech is obviously a preferred interface between human and computer.  We plan to develop speech technologies for the languages of Beijing and Hong Kong — Putonghua, Cantonese and English.  Inclusion of the English language shall also bring our work beyond parochial borders.  In addition, we propose to develop within the financial information domain, due to the abundance of real-time and dynamic information sources — a context well-suited for showcasing the benefits of speech technology.  This domain relates our work closely with electronic commerce, and may lead to opportunities for technology transfer to industry.  More importantly, we hope that our work may have widespread beneficial implications for our local economy.

Project ISIS also attempts to pave the way for further initiatives underneath the JCIE umbrella.  The wealth of financial information often inundates the information consumer.  This situation opens doors for active research in intelligent financial data processing, data-mining, distributed decision-making, etc., which are among the key research directions identified by the Academic Committee.  Intelligent information processing is in turn conducive to the further development of a highly habitable human-computer speech interface — one which provides the user with ease-of-access to relevant, decision-critical information, in a timely and precise manner, and in a digestible form.

At the implementational level, Project ISIS hopes to develop a human-computer speech interface for accessing electronic financial information.  Users will be able to request updated financial information by speaking over the telephone.  Access to private financial information will be secured via speaker authentication.  The appropriate audio response will be automatically generated and delivered in real time.  The telephone interface is desirable due to the high penetration of telephones in Beijing and Hong Kong, in both fixed and mobile networks, as well as roaming services for the latter.  Furthermore, Internet telephony holds significant promise for the future.  The speech interface is an attractive alternative to the modus
operandi — complex, touch-tone menus which often require the user to navigate slowly through.  A single spoken request can bypass long sequences of menus to accomplish an intended task directly.  As such, the human-computer speech interface shall empower both the computer-savvy and the computer-naive with ease-of-access to computing, communication and content.