SearchPad - Version 1.61 SearchPad is a desktop-based Internet "Genie" designed to enhance the speed and accuracy of web searches. It integrates a powerful neural network engine with an advanced AI component that dramatically increases the precision and simplicity of searches. Typically, search engines return only document titles and short summaries, leaving end-users with the rather tedious task of examining each search result individually. SearchPad takes searching to document content level by automatically collecting titles, their URLs, summaries and indexes that match the original query. SearchPad uses its strength of multi- threading capabilities to concurrently query multiple search engines making it a very powerful meta-search tool. Its intelligent back-end then summarizes these search results that have been located. While browsing the Internet, we often find that to obtain `useful' and `relevant' search results, it is expected of the user to provide intelligent and precise queries. Most users find this a difficult task. If the query is too specific then the number of search results is inadequate and if the query too general then too many documents are returned. SearchPad allows the user to formulate `the right' query by supporting an extended set of Boolean operators. The unique feature offered by SearchPad is however the two levels of filtering mechanism it comes armed with. Specifying `accept' and `reject' rules provide the first level of filtering. SearchPad looks for documents that satisfy the query, & the `accept' and the `reject' rules. This allows SearchPad to `find' and not just search. Today there is no other tool that allows a user to train the tool by `rating' documents based on specific requirements. With SearchPad a user can provide `feedback' on a document by specifying its degree of relevance and a rationale for the same. SearchPad allows the user to provide feedback in two modes. A user can give keywords from a document as feedback. The second mode is that a user can highlight a portion of the document text. A powerful and proprietary context extractor with text mining capabilities then comes up with a suggested set of keywords from the highlighted text providing the user the option of selecting those that are relevant from the set and submit them as feedback. Feedback given by the user is used to train SearchPad and is utilized to develop a `personal' usage profile that `remembers' the user's tastes and interests. This provides the second level of information filtering that is applied for automatic classification. Features ? Multiple search engines ( 30+) accessed concurrently ? Automatic elimination of duplicate documents and broken links. ? Automatic grouping of results by domain. ? Built in url tracker to keep track of currency of a document. ? Supports advanced Boolean queries (AND, OR, ANDNOT & NEAR) ? Allows searching on multiple queries simultaneously ? Allows `crawling' and `banning' of specific websites ? Uses feedback to develop a knowledge base (called topic) of usage patterns ? Searches can be scheduled to run at a specific time ? Automatically keeps track of changing search engine formats and updates application ? Search results can be viewed off-line ? Accept and Reject rules for information filtering ? Topic & query structure allow results to be shared across teams ? Allows sorting of results based on Search engine ranking, Domain and Cache ? Query wizard to facilitate formulation of search queries ? Allows specification of search preferences ? Progressive display of results as they are obtained Benefits ? Personal productivity improvement: Very often in an Internet session experience a user is unable to quantify what the user learned or discovered. The user spends a lot of time but learns nothing specific. SearchPad's document classification algorithm allows the user to sort, rank and cluster information, quantifying personal tastes and interests. ? Information sharing & feedback: SearchPad facilitates sharing of search profiles (topics) with others in the team. Knowledge can be institutionalized. Thus saving time and effort. ? Power of organized information: By classifying results, SearchPad categorizes and prioritizes documents, for better information access & management. ? Guided search & content research: SearchPad behaves like an intelligent research assistant (something we always wanted but could not afford). SearchPad remembers a user's past behavior in information navigation and thus works as an information guide, and not just as static compass alone. SearchPad allows you to create your own clones (topics) that mimic your logic (and illogic). ? Build virtual communities: Parents can create valuable knowledge enriched capsules (topics) for their children. Homework tutors can be built to help students in their reports and projects. Interesting SearchPad topics can be shared with other members of the family. ? Start from where you left, maintaining continuity: Unlike any other browser, SearchPad allows a user to carry queries over to next session, thereby maintaining continuity in searches. Thus SearchPad builds up a usage profile over time, which is used to facilitate repeated searches. ? Experience the `wait-less'ness: An advanced information filtering mechanism facilitates getting relevant documents, fast. Saves online time & money. Market Drivers ? Information overload and vocabulary mismatch: Too much information is as good as none. The Web is too slow and good information is hard to find. In addition, each user has different tastes and they vary with time. The diversity of expertise of user population and vocabularies in different functional domains compounds the overload phenomenon. ? Inability to articulate the query exactly: Very often users are unable to formulate queries Intelligently, and the impact of information overload gets multiplied by an order of magnitude because of fuzziness in query. ? Personalization at Desktop: The era of "End User Computing" demands more control at the desk top as opposed to a centralized search engine approach. ? Increasing User expectations: Initially, Netizens had a simple desire to find new fun (for no particular reason). Now, they all want to find something that is useful (to their work or personal interests). For instance a Market Researcher could keep track of his competition in terms of new products, new variants, price changes, feature changes etc. by using the URL tracker facility and neatly organizing them in classified folders. ? Phenomenal Growth of Internet: With the explosion of Internet, the information available is also growing phenomenally. This drives the need for intelligent agents to meet personal information requirements.