Abstract: Additional documents are automatically located that are relevant to an original document, such as a document being read by a user, and also potentially relevant to personal information of the user. The additional documents may be located based on descriptive information that includes personal information of the user and content information of the document being read. The additional documents, or links to the additional documents, may be incorporated into the document being read. In some implementations, the additional documents may be presented in-line with the document being read, such as through an in-link link or text snippet. The user can thus be efficiently presented with additional information that is relevant to the original document being read.

Abstract: User profile information for a user may be determined by (a) determining initial user profile information for the user, (b) inferring user profile information for the user, and (c) determining the user profile information for the user using both the initial user profile information and the inferred user profile information. Initial user profile information for the user may be determined using past search queries submitted by the user, and/or past document selections by the user. User profile information for the user may be inferred by (a) defining a node for each of a number of documents and the user, (b) adding edges between nodes if there is an association between the nodes to define a graph, and (c) inferring user profile information for the user using a topology of the graph and user profile information of other documents.

Abstract: Methods and apparatus consistent with the invention allow a user to submit an ambiguous search query and to receive relevant search results. Queries can be expressed using character sets and/or languages that are different from the character set and/or language of at least some of the data that is to be searched. A translation between these character sets and/or languages can be performed by examining the use of terms in aligned text. Probabilities can be associated with each possible translation. Refinements can be made to these probabilities by examining user interactions with the search results.

Abstract: A method of displaying heterogeneous information includes executing an application program to process one or more activities. Each activity has information nuggets and next step links, and each information nugget is extracted from one or more heterogeneous information sources. Also, each next step link is associated with one or more information nuggets. One of the activities is selected, and an information nugget of the selected activity is displayed in a first region of the graphical user interface. A next step link associated with the displayed information nugget is displayed in a second region of the graphical user interface.

Abstract: A system and method are provided that overcome the previously described limitations of prior art systems. The computer system, in fact, uses a set of information objects to interface with heterogeneous information sources. Information objects can be combined to form tasks that can be added, deleted and modified without requiring any modifications to the underlying program. In addition, the computer system allows users to navigate through a task and automatically generates links to related tasks based on the contextual information in which each step in the task is performed. As a result, links to related tasks do not have to be defined ahead of time, but rather can be dynamically generated at run time. This is accomplished by associating a semantic type or multiple semantic types with the information generated during each step. This allows the system to present links to tasks that utilize the same semantic types generated during the previous step.

Abstract: Some embodiments of the invention include methods for identifying clusters in a database, data warehouse or data mart. The identified clusters can be meaningfully understood by a list of the attributes and corresponding values for each of the clusters. Some embodiments of the invention include a method for scalable probabilistic clustering using a decision tree. Some embodiments of the invention, perform linearly in the size of the set of data and only require a single access to the set of data. Some embodiments of the invention produce interpretable clusters that can be described in terms of a set of attributes and attribute values for that set of attributes. In some embodiments, the cluster can be interpreted by reading the attribute values and attributes on the path from the root node of the decision tree to the node of the decision tree corresponding to the cluster. In some embodiments, it is not necessary for there to be a domain specific distance function for the attributes.

Abstract: Some embodiments of the invention include methods for identifying clusters in a database, data warehouse or data mart. The identified clusters can be meaningfully understood by a list of the attributes and corresponding values for each of the clusters. Some embodiments of the invention include a method for scalable probabilistic clustering using a decision tree. Some embodiments of the invention, perform linearly in the size of the set of data and only require a single access to the set of data. Some embodiments of the invention produce interpretable clusters that can be described in terms of a set of attributes and attribute values for that set of attributes. In some embodiments, the cluster can be interpreted by reading the attribute values and attributes on the path from the root node of the decision tree to the node of the decision tree corresponding to the cluster. In some embodiments, it is not necessary for there to be a domain specific distance function for the attributes.

Abstract: A text classifier and building the text classifier by determining appropriate parameters for the text classifier.

Abstract: A technique, specifically a method and apparatus that implements the method, which through a probabilistic classifier (370) and, for a given recipient, detects electronic mail (e-mail) messages, in an incoming message stream, which that recipient is likely to consider "junk". Specifically, the invention discriminates message content for that recipient, through a probabilistic classifier (e.g., a support vector machine) trained on prior content classifications. Through a resulting quantitative probability measure, i.e., an output confidence level, produced by the classifier for each message and subsequently compared against a predefined threshold, that message is classified as either, e.g., spam or legitimate mail, and, e.g., then stored in a corresponding folder (223, 227) for subsequent retrieval by and display to the recipient.