Abstract: An automated speech recognition (ASR) model includes a first encoder, a first encoder, a second encoder, and a second decoder. The first encoder receives, as input, a sequence of acoustic frames, and generates, at each of a plurality of output steps, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The first decoder receives, as input, the first higher order feature representation generated by the first encoder, and generates a first probability distribution over possible speech recognition hypotheses. The second encoder receives, as input, the first higher order feature representation generated by the first encoder, and generates a second higher order feature representation for a corresponding first higher order feature frame. The second decoder receives, as input, the second higher order feature representation generated by the second encoder, and generates a second probability distribution over possible speech recognition hypotheses.

Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.

Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.

Abstract: A sparsifier is generated from a union of multiple spanners of a graph. The edges of the sparsifier are weighted based on a measure of connectivity called robust connectivity. The robust connectivity of a node pair is the highest edge sampling probability at which a distance between the pair is likely to drop below a specified length. Each spanner is generated from a subgraph of the graph that is generated using a decreasing sampling probability. For the weight of each edge, a spanner is determined where an estimated distance between the nodes associated with the edge is greater than a threshold distance. The sampling probability of the subgraph used to generate the spanner is an estimate of the robust connectivity of the edge. The weight of the edge is set to the inverse of the estimated robust connectivity.

Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.

Abstract: Architecture that provides a data structure to facilitate personalized ranking over recommended content (e.g., documents). The data structure approximates the social distance of the searching user to the content at query time. A graph is created of content recommended by members of the social network, where the nodes of the graph include content nodes (for the content) and recommending member nodes (for members of the social network who recommended the content). If a member recommends content, an edge is created between the member node and the content node. If a member is a “friend” (tagged as related in some way) of another member, an edge is created between the two member nodes. Each node is converted to a lower dimensional feature set. Feature sets of the content are indexed and the feature set of the searching user is utilized to match and rank the search results at query time.

Abstract: Multiple data prediction strategies are received. Each data prediction strategy may predict a next data value in a sequence of data values with a corresponding confidence value. Rather than rely on a single prediction strategy, the predictions of each of the data prediction strategies are linearly combined to generate a single prediction that is more accurate and has a lower overall loss than any of the individual prediction strategies. Further, a deviation is calculated based on the values in the sequence of values that have been observed so far using a weighted sum that favors more recent values in the sequence over less recent values in the sequence. A prediction of the next value in the sequence is generated based on the combined strategies and the calculated deviation.

Abstract: User accounts in a social networking application are divided into highly-connected accounts and regular accounts. A mapping of the highly-connected accounts to their friends, and a mapping of accounts to documents endorsed by the users associated with the accounts are stored on index servers of a search engine. When a query is received by a front-end server of the search engine, the front-end server determines if an account associated with the query is a highly-connected account. If it is, only an identifier of the account is sent to the index servers with the query. If it is not, however, then identifiers of all of the accounts that are friends with the account are sent with the query. The index servers then determine the documents that are endorsed by the friends of the account, and consider the determined documents when selecting documents that are responsive to the query.

Abstract: To facilitate the estimation of relatedness between nodes of a graph, implementations estimate relatedness between nodes in a graph by pre-computing for a subset of sample nodes (e.g., center nodes) a plurality of transition probabilities between each sample node and each of the other nodes in the graph, and then later when queried the implementations calculate in real-time the resultant estimated transition probability between the first node and the second node through the at least one sample node based on the pre-computed transition probabilities.

Abstract: A sparsifier is generated from a union of multiple spanners of a graph. The edges of the sparsifier are weighted based on a measure of connectivity called robust connectivity. The robust connectivity of a node pair is the highest edge sampling probability at which a distance between the pair is likely to drop below a specified length. Each spanner is generated from a subgraph of the graph that is generated using a decreasing sampling probability. For the weight of each edge, a spanner is determined where an estimated distance between the nodes associated with the edge is greater than a threshold distance. The sampling probability of the subgraph used to generate the spanner is an estimate of the robust connectivity of the edge. The weight of the edge is set to the inverse of the estimated robust connectivity.

Abstract: Sketches are generated for each node in a graph. For undirected graphs, each sketch for a node may include an indicator of a node from a seed set of nodes and the shortest distance between the node and the indicated node. When a request is received for the shortest distance between two nodes of the graph, the sketches for each of the two nodes are retrieved, and nodes that are indicated in both of the sketches are determined. The distances between each of the two nodes and a determined node as indicated in the sketches is summed for each of the determined nodes, and the sum having the least distance is selected as the estimated shortest distance between the two nodes.

Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.

Abstract: Architecture that provides a data structure to facilitate personalized ranking over recommended content (e.g., documents). The data structure approximates the social distance of the searching user to the content at query time. A graph is created of content recommended by members of the social network, where the nodes of the graph include content nodes (for the content) and recommending member nodes (for members of the social network who recommended the content). If a member recommends content, an edge is created between the member node and the content node. If a member is a “friend” (tagged as related in some way) of another member, an edge is created between the two member nodes. Each node is converted to a lower dimensional feature set. Feature sets of the content are indexed and the feature set of the searching user is utilized to match and rank the search results at query time.

Abstract: User accounts in a social networking application are divided into highly-connected accounts and regular accounts. A mapping of the highly-connected accounts to their friends, and a mapping of accounts to documents endorsed by the users associated with the accounts are stored on index servers of a search engine. When a query is received by a front-end server of the search engine, the front-end server determines if an account associated with the query is a highly-connected account. If it is, only an identifier of the account is sent to the index servers with the query. If it is not, however, then identifiers of all of the accounts that are friends with the account are sent with the query. The index servers then determine the documents that are endorsed by the friends of the account, and consider the determined documents when selecting documents that are responsive to the query.

Abstract: A graph is generated based on a social networking application with a node for each user account, and one or more edges representing the social networking relationships between the user accounts (e.g., friends). A sketch is generated for each node in iterations where edges are removed from the graph and a set of reachable nodes is determined for the node. A representative node is then selected from the set of reachable nodes and added to the sketch as a dimension. The generated sketches for two nodes are used to calculate an affinity score between the accounts associated with each of the two nodes.

Abstract: Multiple data prediction strategies are received. Each data prediction strategy may predict a next data value in a sequence of data values with a corresponding confidence value. Rather than rely on a single prediction strategy, the predictions of each of the data prediction strategies are linearly combined to generate a single prediction that is more accurate and has a lower overall loss than any of the individual prediction strategies. Further, a deviation is calculated based on the values in the sequence of values that have been observed so far using a weighted sum that favors more recent values in the sequence over less recent values in the sequence. A prediction of the next value in the sequence is generated based on the combined strategies and the calculated deviation.

Abstract: To facilitate the estimation of relatedness between nodes of a graph, implementations estimate relatedness between nodes in a graph by pre-computing for a subset of sample nodes (e.g., center nodes) a plurality of transition probabilities between each sample node and each of the other nodes in the graph, and then later when queried the implementations calculate in real-time the resultant estimated transition probability between the first node and the second node through the at least one sample node based on the pre-computed transition probabilities.

Abstract: Relevant search results for a given query may be determined using click data for the query and the number of times the query is issued to a search engine. The number of clicks that a result receives for the given query may provide a feedback mechanism to the search engine on how relevant the result is for the given query. The frequency of a query along with the associated clicks provides the search engine with the effectiveness of the query in producing relevant results. Edges in a graph of queries versus results may be weighted in accordance with the click data and the efficiency to rank the search results provided to a user.

Abstract: The rank of nodes in a graph may be inferred from a calculated probability that each node in the graph appears in a single random walk of the graph. Short random walks may be generated for each node in the graph. The generated random walks may be combined to form a longer single random walk. Multiple single random walks may be generated in this fashion. The probability of each node appearing in a single random may be calculated from the generated single random walks. The rank of each node may then be inferred from the calculated probabilities.

Abstract: A technique for automatically detecting and correcting configuration errors in a computing system. In a learning process, recurring event sequences, including e.g., registry access events, are identified from event logs, and corresponding rules are developed. In a detecting phase, the rules are applied to detected event sequences to identify violations and to recover from failures. Event sequences across multiple hosts can be analyzed. The recurring event sequences are identified efficiently by flattening a hierarchical sequence of the events such as is obtained from the Sequitur algorithm. A trie is generated from the recurring event sequences and edges of nodes of the trie are marked as rule edges or non-rule edges. A rule is formed from a set of nodes connected by rule edges. The rules can be updated as additional event sequences are analyzed. False positive suppression policies include a violation-consistency policy and an expected event disappearance policy.