Fujitsu Corporation

Servicenow, Inc.

Agilent Technologies, Inc.

Huawei Technologies Co., Limited

Sandisk Technologies Inc.

Facebook Technologies, LLC

Meta Platforms Technologies, LLC

Jun Wang

Kanji Uchino

Baskar Jayaraman

Chitrabharathi Ganapathy

Aniruddha Madhusudan Thakur

Ruwan T Kurulugama

Feng Q Pan

Drew Eric Wingard

Bo Lei

Kannan Govindarajan

Gregor T Overney

Neeraj Upasani

Wojciech Stefan Powiertowski

Gregory Ehmann

Handong Ye

Shankar Guhados

Yuko Okubo

Chen Tian

Miguel Angel Guerrero

Gang Zou

Dinesh Kumar Kishorkumar Surapaneni

George Stafford

Xuefeng Han

Marco Brambilla

Tao Feng

Takuro Watanabe

Reza Kakoee

Sneha Sharma

Minli Lin

Karthik Rao

Zongfang Lin

Ziang Hu

Systems and methods involving data structures for efficient management of paragraph vectors for textual searching are described. A database may contain records, each associated with an identifier and including a text string and timestamp. A look-up table may contain entries for text strings from the records, each entry associating: a paragraph vector for a respective unique text string, a hash of the respective unique text string, and a set of identifiers of records containing the respective unique text string. A server may receive from a client device an input string, compute a hash of the input string, and determine matching table entries, each containing a hash identical to that of the input string, or a paragraph vector similar to one calculated for the input string. A prioritized list of identifiers from the matching entries may be determined based on timestamps, and the prioritized list may be returned to the client.

Data structures for efficient storage and updating of paragraph vectors

Adjusting an operating frequency of a CPU includes setting the operating frequency for a current operating cycle based on a CPU load in a prior operating cycle and a target CPU load. A current CPU load associated with the current processing cycle is detected. The CPU operating frequency is adjusted to a new operating frequency based on a difference between the target CPU load and the current CPU load. The operating frequency is adjusted based on minimizing the difference between the target CPU load and the detected load. A CPU load error is determined based on the current CPU load and the target CPU load. The target CPU load is adjusted based on the determined CPU load error and a threshold load error. A prediction is generated on whether to perform a new adjustment of the operating frequency of the CPU prior to expiration of a threshold time duration.

Energy efficiency adjustments for a CPU governor

According to an aspect of an embodiment, operations may include obtaining multiple electronic documents and obtaining a theme text. The method may also include selecting a seed text based on a semantic similarity between the seed text and the theme text. The method may also include changing a seed weight included in a weight vector that is used in identification of topics of the multiple electronic documents. The changed seed weight may bias the identification of topics of the plurality of electronic documents in favor of the seed text as compared to one or more other text strings of the weight vector. The method may also include generating, a representation of a topic model for display to a user, the topic model may be based on the multiple electronic documents and the weight vector.

Analysis of theme coverage of documents

A computer-implemented method includes obtaining a plurality of textual records divided into clusters and a residual set of the textual records, where a machine learning (ML) clustering model has divided the plurality of textual records into the clusters based on a similarity metric. The method also includes receiving, from a client device, a particular textual record representing a query and determining, by way of the ML clustering model and based on the similarity metric, that the particular textual record does not fit into any of the clusters. The method additionally includes, in response to determining that the particular textual record does not fit into any of the clusters, adding the particular textual record to the residual set of the textual records. The method can additionally include identifying, by way of the ML clustering model, that the residual set of the textual records contains a further cluster.

Clustering and dynamic re-clustering of similar textual documents

The disclosure describes techniques for interrupt and inter-processor communication (IPC) mechanisms that are shared among computer processors. For example, an artificial reality system includes a plurality of processors; an inter-processor communication (IPC) unit comprising a register, wherein the IPC unit is configured to: receive a memory access request from a first processor of the processors, wherein the memory access request includes information indicative of a hardware identifier (HWID) associated with the first processor; determine whether the HWID associated with the first processor matches an HWID for the register of the IPC unit; and permit, based on determining that the HWID associated with the first processor matches the HWID for the register of the IPC unit, the memory access request to indicate a communication from the first processor to at least one other processor.

Artificial reality system with inter-processor communication (IPC)

A method may include obtaining multiple electronic documents and multiple topics associated with the electronic documents. The method may further include determining a similarity between a first topic and a second topic. The first topic may be associated with a first set of electronic documents. The method may further include refining the multiple topics based on the similarity between the first topic and the second topic by associating the first set of the electronic documents with the second topic and removing the first topic from the multiple topics. The method may further include building a document-classifier model by applying machine learning to at least one electronic document associated with each of the refined topics. The method may further include obtaining an electronic document and classifying the electronic document into one of the refined topics using the document-classifier model.

Classification of electronic documents

A method includes storing a research paper and a set of candidate resources including media content. The method further includes encoding each of one or more first content fields in the research paper into a first vector based on a first field type associated with each of the one or more first content fields. The method further includes encoding each of one or more second content fields in each of the parsed set of candidate resources into a second vector, based on a second field type associated with each of the one or more second content fields. The method further includes comparing the first vector with the second vector to determine a final set of resources based on the comparison. The method further includes controlling a display screen to output the determined final set of resources and the research paper.

Extraction of multi-modal online resources associated with research papers

A method of automatically estimating the difficulty level of online content. The method includes receiving, by one or more processors, electronic content which is accessible via an online network and which is at least temporarily stored in at least one non-transitory storage media, extracting, by the one or more processors, structural and non-linguistic information from the electronic content. The method further includes extracting, by the one or more processors, linguistic information from the electronic content. The method further includes generating, by the one or more processors, a difficulty estimation for the content based on the structural and non-linguistic information in addition to the linguistic information. The method further includes generating, by the one or more processors, a recommendation of a subset of the electronic content to a user based on the difficulty estimation.

System and method for automatic difficulty level estimation

This disclosure describes hardware-based mutexes that employ software process authentication to prevent a software process from releasing the lock of a mutex locked by another software process. For example, systems are described in which a mutex controller receives a request from a process to lock a mutex. The mutex controller locks the mutex, writing a process key and process identifier to one or more hardware registers associated with the mutex. If the mutex controller receives a request to release the lock on the mutex, the mutex controller determines if the key received with the request matches the process key written in the one or more hardware registers of the mutex and, if so, releases the lock on the mutex.

Artificial reality system having hardware mutex with process authentication

Word vectors are multi-dimensional vectors that represent words in a corpus of text and that are embedded in a semantically-encoded vector space. Word vectors can be used for sentiment analysis, comparison of the topic or content of sentences, paragraphs, or other passages of text or other natural language processing tasks. However, the generation of word vectors can be computationally expensive. Accordingly, when a set of word vectors is needed for a particular corpus of text, a set of word vectors previously generated from a corpus of text that is sufficiently similar to the particular corpus of text, with respect to some criteria, may be re-used for the particular corpus of text. Such similarity could include the two corpora of text containing the same or similar sets of words or containing incident reports or other time-coded sets of text from overlapping or otherwise similar periods of time.

Persistent word vector input to multiple machine learning models

A method includes seed expansion using Graph Neural Network (GNN). The method includes retrieving a first plurality of seed accounts based on a first user input indicative of domain-specific information associated with the first plurality of seed accounts. The method further includes determining a second plurality of seed accounts based on the first plurality of seed accounts associated with the domain-specific information, for training a GNN. The method further includes updating the retrieved first plurality of seed accounts and the determined second plurality of seed accounts based on the trained GNN. The method further includes generating a final set of seed accounts associated with the domain-specific information, based on an application of a predefined ranking criteria on the updated first plurality of seed accounts and the second plurality of seed accounts. The method further includes controlling a display screen to output the generated final set of seed accounts.

Seed expansion in social network using graph neural network

In an embodiment, operations including extracting a first set of sub-graphs from a first graph. A set of network motifs is extracted from the first graph based on first statistical information associated with the extracted first set of sub-graphs, and a set of network motif instances of the set of network motifs is located. Second statistical information of the set of network motifs is determined based on the set of network motif instances. The second statistical information includes network motif information of a node or edge associated with the set of network motif instances. A graph explanation model is applied on the first graph based on the second statistical information. A first explanation sub-graph for a first class is extracted from the first graph, based on the application of the graph explanation model and a set of regularization constraints. An output representation indicative of the first explanation sub-graph is displayed.

Explanation of graph-based predictions using network motif analysis

In an embodiment, operations include crawling a set of web pages and labeling one or more items of a first web page based on user input. Each item corresponds to a node in a first tree data structure of the first web page. The operations further include generating a first extraction rule to extract a first item from the one or more first items. The first extraction rule includes a first path, in the first tree data structure, for a first node of the first item, and includes first visual information of each node in the first path. The operations further include comparing the first visual information in the first path with second visual information of each of a plurality of candidate nodes in a second tree data structure of a second web page and further refining the first extraction rule to generate a second extraction rule.

Enhanced information extraction from web pages

A method of collecting and organizing online resources may include selecting an item from an electronic document. The method may include extracting data related to the item from the electronic document. The method may include identifying whether a first list that includes the item is associated with the electronic document. The method may include refining the first list in response to an identification of the first list. The method may include generating a second list that includes the item in response to a failure to identify the first list. The method may include verifying that the first list is consistent with a user interest. The method may include saving the first list to an electronic data storage device when the first list is consistent with the user interest. The method may include selecting a second item when the first list is inconsistent with the user interest.

Growing community of inventors

San Jose, CA, United States of America

Jun Wang