Abstract: An adaptive probabilistic computer-controlled method and system determines computer-controlled actions that are expected to reduce uncertainties that are embodied as probabilities and then updates the probabilities in accordance with the results of the computer-controlled actions. The updated probabilities inform the determination of subsequent computer-controlled actions. Expected values of information may also be applied in determining the subsequent computer-controlled actions. The computer-controlled actions may comprise analysis of content and generating recommendations that are delivered to users, and inferences from the analysis of content and usage behaviors may inform the updating of the probabilities. Statistical learning functions and/or neural networks may be applied in determining the computer-controlled actions.

Abstract: A self-referential semantic-based method, system, and device links semantic chains comprising subject, predicate, and objects of the predicate and optionally, an associated probability, with behavioral-based chains comprising a computer-implemented system, predicate, and an object of the predicate, and generates self-referential communications based on the linked chains. The self-referential communications may include metaphorical and/or imaginative constructs, and may convey a sense of confidence or what has been learned over time in accordance with weightings associated with the semantic and/or behavioral chains. The self-referential semantic-based device may take the form of a self-propelled apparatus, including an apparatus in humanoid form.

Abstract: A directed expertise level-based discovery system, method, and device infers an expertise level associated with a system user from a plurality of usage behaviors and delivers one or more computer-implemented objects to the user that are selected in accordance with the inferred expertise level and a directionally distinct relationship between computer-implemented objects. The directed expertise level-based discovery system, method, and device may infer the expertise level from textual information and/or in accordance with calibration information such as test results. Geographic location awareness associated with a portable device may inform the selection of the computer-implemented objects to be delivered to the user. An explanation may be delivered to the user that references information that is used to inform the selection of the computer-implemented objects to be delivered to the user.

Abstract: A content-based expertise level inferencing system and method analyzes content such as text-based content that is associated with a user and infers an expertise level of the user that is associated with a topic. Language-based analytic techniques may be applied in inferring the expertise level, including the application of domain-specific vocabulary and frequency analyzes. The expertise level may be further inferred based on usage behaviors of the user and/or other users. Calibration information may be applied in generating the inference of the expertise level. Recommendations may be generated based on the inferred expertise level, and explanations for the recommendation and/or the inference may be generated for delivery to a recommendation recipient.

Abstract: A personalized activity stream system, method, and device delivers a recommendation that is contextualized basis a selected item in an activity stream and that is personalized for a recommendation recipient consistent with an inference of the recommendation recipient's interests. Contextualization and personalization may be based on an automatic analysis of usage behaviors and/or content. The recommendation may be informed by an inference of an expertise level, and an explanation for the recommendation may be delivered to the recommendation recipient. The recommendation may be in accordance with an automatically determined geographic location associated with a location-aware portable device, as well as environmental conditions proximal to the portable device.

Abstract: A computer-implemented adaptive experimentation method and system is described that automatically selects and executes information gathering actions. The adaptive experimentation method and system integrates value of information considerations, experimental design, and inferences from experimental results. The experimental results may include behaviors of users of a computer-based system. The process enables an automatic, adaptive process for attaining additional information and applying the attained information in making subsequent experiment decisions.

Abstract: A physical object-based people matching method, system, and device recommends a first person to a second person based on the geographic proximity of a location associated with the first person to one or more physical objects that are associated with the second person. The geographic proximity may be determined from the use of a location-aware portable device. The recommendation may be further informed by a mobility inference. The first person is enabled to send an expression of interest to the second person in response to the recommendation, and the second person may be enabled to send an expression of interest to the first person in response to receiving the expression of interest. An explanation for the recommendation may be delivered to the first person that references a mobility inference or a computer-implemented map.

Abstract: A location-based matching system and method matches users of a computer-implemented system in accordance with inferences of mutual interest between the users, whereby the inferences of mutual interest are based on mutual computer-implemented relationships and user location information. Inferences of mutual interest may also be in accordance with a physical object. The inferences of mutual interest may be with respect to interests by users in being physically proximal to each other or being physically proximal to a location associated with one of the users. Information associated with user matches may be delivered to a matched user in the form of a computer-implemented map.

Abstract: A mutual expressions of interest people matching system and method recommends people to other people in accordance with inferences of mutual interests from usage behaviors. A person who receives a recommendation can express an interest to a recommended person, and the expression of interest can be an explicit indication of interest in establishing a computer-implemented relationship or the establishing of a directional distinct relationship between the recommended person and the recommendation receiver. The recommended person can in response express an interest to the recommendation receiver by establishing a computer-implemented relationship between the parties. The relationship between the parties may be a subscription-based relationship and explanations for the recommendations may be provided to recommendation receivers.

Abstract: An integrated search and adaptive discovery system and method integrates contents-based indexing and behavioral-based indexing of collections of computer-implemented objects to generate contextualized and/or personalized recommendations. The degree to which contextualization and/or personalization criteria are applied in generating recommendations can be tuned by the recommendation recipient. Personalization functions are applied that are informed by inferred interests and/or expertise, and personalization vectors can be transformed into a format executable by a search engine. Explanations may be provided to recommendation recipients as to why they received recommendations.

Abstract: A comparative expertise level-based discovery system infers an expertise level associated with a system user from a plurality of usage behaviors and compares this expertise level with expertise levels associated with computer-implemented objects. One or more of the computer-implemented objects are selected for delivery to the user in accordance with a comparison of the inferred expertise level of the user and the expertise levels associated with the computer-implemented objects. The expertise levels associated with the computer-implemented objects may be inferred from information within the objects such as text and/or may be inferred from usage behaviors. Inferences of interests of the user may further inform the selection of one or more of the computer-implemented objects for delivery to the user. Explanations for the delivery of the computer-implemented objects may be provided to the user.

Abstract: A multi-social network discovery system and method enables users to select multiple social networks comprising activity streams and to receive recommendations sourced from the selected multiple social networks. The multi-social network discovery system and method generates affinity values based on inferred affinity levels between topics, which may be associated with tags, and objects that are associated with the activity streams. The inferred affinity values may be derived from usage behaviors such as tagging and/or based on an automatic analysis of the content of the objects. The affinity values inform the selection of the objects to be recommended to a user, and the selection may be further informed by inferred affinity levels between the user and the topics. The recommendations may be tuned by the user through application of a recommendation tuning control.

Abstract: A computer-implemented transformational protocol is applied to contextually transform an originating computer-implemented structure into a fuzzy network-based structure. The originating structure may be a hierarchy, for example. The fuzzy network-based structure is adaptively modified based on behavioral information, and recommendations delivered to users may be generated in accordance with the fuzzy network-based structure and inferences from a plurality of usage behaviors. The recommendations may be generated in accordance with a network neighborhood within the fuzzy network-based structure that is based, at least in part, on a navigational context within the originating structure.

Abstract: A computer-implemented transforming function contextually transforms multiple originating computer-implemented structures into a composite fuzzy network-based structure. The composite fuzzy network-based structure serves as a contextual basis for recommendations. The recommendations may be generated in accordance with inferences of preferences and interests based on usage behaviors. The recommendations may further be generated in accordance with a network neighborhood within the composite fuzzy network-based structure that is based, at least in part, on a navigational context within an originating structure and a user-established navigational neighborhood scope.

Abstract: A computer-implemented adaptive expertise clustering system and method generates expertise cohorts based on inferences from behavioral information and applies the expertise cohorts in generating recommendations for delivery to users. The expertise clustering may be informed by the evaluation of the contents of computer-implemented objects. Behaviors associated with members of specific expertise cohorts after specific events may inform the recommendations. Recommendations may comprise, for example, other users or process steps. Explanations for the recommendations may be provided to recommendation recipients.

Abstract: A computer-implemented discovery system applies a contextual neighborhood scope with respect to a fuzzy network-based structure, and generates recommendations in accordance with the resulting contextual neighborhood and inferences from behavioral information. The contextual neighborhood scope may be established by a user, and the basis for the contextual neighborhood may be in accordance with an object that is accessed by a user. The inferences may be embodied in affinity vectors, and may include inferences of user preferences and levels of expertise.

Abstract: A computer-implemented activity stream-based recommendations system delivers recommendations in accordance with a selected item of an activity stream, inferences of interests based on usage behaviors, and a contextual neighborhood of objects. In addition, or alternatively, the recommendations may be generated in accordance with an inference of expertise. The contents of the objects in the activity stream may be generated by humans or automatically by a processor-based device. Explanations for the recommendations may be delivered to recommendation recipients.

Abstract: A computer-implemented portable inferred interests and expertise profiling system generates an inferred user profile embodied as an inference vector, which is generated based on inferences from behavioral information and in accordance with standardized tags that are associated with topical areas in a first system. The inferences may be of, for example, interests or levels of expertise. Recommendations are then generated based on the inference vector and a different set of topics. Multiple inference vectors may be combined into a composite inference vector, which can serve as a basis for recommendations generated with respect to different sets of topics than those that informed the generation of the composite inference vector.

Abstract: A computer-implemented serendipitous recommendations system and method generates recommendations for delivery to system users in accordance with settings of desired levels of serendipity, including serendipity levels established through use of serendipity tuning controls operable by users. The recommendations are informed by an interest affinity anomaly function that identifies contrasting interest affinities between recommendation recipients and other users. Explanations may be generated that provide reasons as to why a recommendation was delivered to a user, and the explanation may include a selection of phrases that are influenced by a serendipity level setting, and may include an expression of a level of confidence with regard to the recommendation.

Abstract: A computer-implemented expertise discovery method and system generates recommendations of people in accordance with inferences of levels of expertise from behavioral information. The generation of the recommendations may be in accordance with comparisons of two or more expertise vectors and associated correlations between the vectors. The inferences of expertise may be with respect to specified topical neighborhoods. Explanations as to why recommendations were made are delivered to the recommendation recipients, and may include references to the inferred levels of expertise, the behavioral bases for the inferences, and/or the confidence in the inferences.