Abstract: Advertiser repository information (e.g., a list of HTML pages) may be used to automatically generate a page-ad and associated ad information for each page using a subset of features of the page. Subsequently, a page-ad index for all pages from different advertisers may be created. After creating a page-ad index, if a client issues a query (e.g., from a search engine), candidate page-ads may be selected using the page-ads index. If necessary (e.g., if there are more candidate page-ads than requested ads), the candidate page-ads may be entered into an arbitration (e.g., auction). Then, the arbitration may determine a set of one or more winning page-ads to be served for rendering.

Abstract: A system receives a request from a client and provides an aerial image to the client in response to the request. The aerial image includes an advertisement superimposed on the aerial image.

Abstract: A fired heater is adapted for increasing the output of a plant where the furnace capacity is considerably improved without corresponding increase in the pressure drop. The described technique utilizes a parallel thermal path in contrast to the conventional series thermal path for heating a hydrocarbon fluid. The fluid is divided into at least two paths where the fluid in the first path is heated primarily by radiation heat transfer mechanism and the fluid in the second path is heated primarily by convection heat transfer mechanism. The at least two fluid streams may then be combined to continue with other desired processing of the fluid.

Abstract: The present invention relates to a system and methodology providing layered probabilistic representations for sensing, learning, and inference from multiple sensory streams at multiple levels of temporal granularity and abstraction. The methods facilitate robustness to subtle changes in environment and enable model adaptation with minimal retraining. An architecture of Layered Hidden Markov Models (LHMMs) can be employed having parameters learned from stream data and at different periods of time, wherein inferences can be determined relating to context and activity from perceptual signals.

Abstract: Methods and apparatus, including systems and computer program products, to provide clustering of users in which users are each represented as a set of elements representing items, e.g., items selected by users using a system. In one aspect, a program operates to obtain a respective interest set for each of multiple users, each interest set representing items in which the respective user expressed interest; for each of the users, to determine k hash values of the respective interest set, wherein the i-th hash value is a minimum value under a corresponding i-th hash function; and to assign each of the multiple users to each of the respective k clusters established for the respective user, the i-th cluster being represented by the i-th hash value. The assignment of each of the users to k clusters is done without regard to the assignment of any of the other users to k clusters.

Abstract: Disclosed is a system, method, and program storage device of aggregating opinions comprising consolidating a plurality of expressed opinions on various dimensions of topics as discrete probability distributions, generating an aggregate opinion as a single point probability distribution by minimizing a sum of weighted divergences between a plurality of the discrete probability distributions, and presenting the aggregate opinion as a Bayesian network, wherein the divergences comprise Kullback-Liebler distance divergences, and wherein the expressed opinions are generated by experts and comprise opinions on sentiments of products and services. Moreover, the aggregate opinion predicts success of the products and services. Furthermore, the experts are arranged in a hierarchy of knowledge, wherein the knowledge comprises the various dimensions of topics for which opinions may be expressed upon.

Abstract: A promotional offer may be generated by (i) accepting information concerning at least one of (A) a search query entered, at a client device, by a user, (B) an item or establishment which is the subject of a search result selected by a user using a client device, (C) one or more items or establishments which are elements of a shopping session summary provided to a user via a client device, (ii) determining a promotional offer to serve using at least the accepted search query information, and (iii) determining terms of the promotional offer using at least one of (A) a location of the client device, (B) a distance from the client device to an establishment associated with the promotional offer, and (C) a distance from the client device to an establishment competing with the establishment associated with the promotional offer, (D) an inventory, at an establishment associated with the promotional offer, of the goods which the promotional offer concerns, (E) a capacity, at an establishment associated with the promo

Abstract: The present invention relates to a system and methodology to facilitate machine learning and predictive capabilities in a processing environment. In one aspect of the present invention, a Mutual Information Model is provided to facilitate predictive state determinations in accordance with signal or data analysis, and to mitigate classification error. The model parameters are computed by maximizing a convex combination of the mutual information between hidden states and the observations and the joint likelihood of states and observations in training data. Once the model parameters have been learned, new data can be accurately classified.

Abstract: The present invention relates to a system and methodology to facilitate machine learning and predictive capabilities in a processing environment. In one aspect of the present invention, a Mutual Information Model is provided to facilitate predictive state determinations in accordance with signal or data analysis, and to mitigate classification error. The model parameters are computed by maximizing a convex combination of the mutual information between hidden states and the observations and the joint likelihood of states and observations in training data. Once the model parameters have been learned, new data can be accurately classified.

Abstract: A system, method and program storage device implementing a method for modeling a data generating process, wherein the modeling comprises observing a data sequence comprising irregularly sampled data, obtaining an observation sequence based on the observed data sequence, assigning a time index sequence to the data sequence, obtaining a hidden state sequence of the data sequence, and decoding the data sequence based on a combination of the time index sequence and the hidden state sequence to model the data sequence. The method further comprises assigning a probability distribution over time stamp values of the observation sequence, wherein the decoding comprises using a Hidden Markov Model. The method further comprises using an expectation maximization methodology to learn the Hidden Markov Model.

Abstract: Disclosed is a system, method, and program storage device of aggregating opinions comprising consolidating a plurality of expressed opinions on various dimensions of topics as discrete probability distributions, generating an aggregate opinion as a single point probability distribution by minimizing a sum of weighted divergences between a plurality of the discrete probability distributions, and presenting the aggregate opinion as a Bayesian network, wherein the divergences comprise Kullback-Liebler distance divergences, and wherein the expressed opinions are generated by experts and comprise opinions on sentiments of products and services. Moreover, the aggregate opinion predicts success of the products and services. Furthermore, the experts are arranged in a hierarchy of knowledge, wherein the knowledge comprises the various dimensions of topics for which opinions may be expressed upon.

Abstract: Computer based apparatus and method automates gene prediction in a subject genomic sequence. A plurality of expert systems provide preliminary or intermediate gene predictions. A Bayesian network combiner combines the intermediate gene predictions and forms a final gene prediction. The final gene prediction accounts for dependencies between individual expert systems and dependencies between adjacent parts of the subject genomic sequence.

Abstract: Computer based apparatus and method automates gene prediction in a subject genomic sequence. A plurality of expert systems provide preliminary or intermediate gene predictions. A Bayesian network combiner combines the intermediate gene predictions and forms a final gene prediction. The final gene prediction accounts for dependencies between individual expert systems and dependencies between adjacent parts of the subject genomic sequence.

Abstract: A fired heater is adapted for increasing the output of a plant where the furnace capacity is considerably improved without corresponding increase in the pressure drop. The described technique utilizes a parallel thermal path in contrast to the conventional series thermal path for heating a hydrocarbon fluid. The fluid is divided into at least two paths where the fluid in the first path is heated primarily by radiation heat transfer mechanism and the fluid in the second path is heated primarily by convection heat transfer mechanism. The at least two fluid streams may then be combined to continue with other desired processing of the fluid.

Abstract: The present invention relates to a system and methodology to facilitate machine learning and predictive capabilities in a processing environment. In one aspect of the present invention, a Mutual Information Model is provided to facilitate predictive state determinations in accordance with signal or data analysis, and to mitigate classification error. The model parameters are computed by maximizing a convex combination of the mutual information between hidden states and the observations and the joint likelihood of states and observations in training data. Once the model parameters have been learned, new data can be accurately classified.

Abstract: The present invention relates to a system and methodology providing layered probabilistic representations for sensing, learning, and inference from multiple sensory streams at multiple levels of temporal granularity and abstraction. The methods facilitate robustness to subtle changes in environment and enable model adaptation with minimal retraining. An architecture of Layered Hidden Markov Models (LHMMs) can be employed having parameters learned from stream data and at different periods of time, wherein inferences can be determined relating to context and activity from perceptual signals.

Abstract: Computer based apparatus and method automates gene prediction in a subject genomic sequence. A plurality of expert systems provide preliminary or intermediate gene predictions. A Bayesian network combiner combines the intermediate gene predictions and forms a final gene prediction. The final gene prediction accounts for dependencies between individual expert systems and dependencies between adjacent parts of the subject genomic sequence.