Abstract: A journal-based database may include items, e.g., organized in a table, associated with a journal. The journal may include a hash-chained set of blocks individually including data representing previous and/or current versions of the individual items. The database may receive a request to redact specific data from the database. The database may search the blocks to determine whether the specific data exists. When the specific data exists in a block of the journal, the database may redate the specific data in the identified block. The database may retain existing metadata of the block such that cryptographic verifiability of the hash-chained blocks may be preserved.

Abstract: A document management system can include an artificial intelligence-based document manager that can perform one or more predictive operations based on characteristics of a user, a document, a user account, or historical document activity. For instance, the document management system can apply a machine-learning model to determine how long an expiring agreement document is likely to take to renegotiate and can prompt a user to begin the renegotiation process in advance. The document management system can detect a change to language in a particular clause type and can prompt a user to update other documents that include the clause type to include the change. The document management system can determine a type of a document being worked on and can identify one or more actions that a corresponding user may want to take using a machine-learning model trained on similar documents and similar users.

Abstract: Systems and methods for determining neural network brittleness are disclosed. For example, the system may include one or more memory units storing instructions and one or more processors configured to execute the instructions to perform operations. The operations may include receiving a modeling request comprising a preliminary model and a dataset. The operations may include determining a preliminary brittleness score of the preliminary model. The operations may include identifying a reference model and determining a reference brittleness score of the reference model. The operations may include comparing the preliminary brittleness score to the reference brittleness score and generating a preferred model based on the comparison. The operations may include providing the preferred model.

Abstract: A log sampling and storage system reduces volumes of stored log information. A log storage engine receives a series of log messages responsive to a request. The engine compares that series to various patterns previously generated from historical log data. This comparison can reference •a name of the request; •a total time to process the request; •an item count; •a message length; and/or •a message hash code. Comparing incoming log series with the predefined patterns, results in generation of a similarity score. If the score falls outside a tolerance, the existence of an outlier log message series is indicated, and that outlier log message series is stored in its entirety. However, if the similarity score falls within a tolerance, similarity to an existing predefined pattern is indicated. Rather than storing the (non-outlier) log message series, the engine directs storing a link to a sample log message series reflecting the pattern.

Abstract: Document embedding vectors for each document of a corpus may be generated by combining embedding vectors for document subparts, thereby yielding a final embedding vector for the document. A machine learning model is trained using a query corpus and the document corpus, where the model generates a ranking score for a given (query, document) pair. During training, rankings scores are generated using the model, such that the training dataset is further refined using the generated ranking scores. For example, top documents and a negative document may be determined for a given query and subsequently used as training data. Multiple negative documents may therefore be determined for a given query. A negative document for a given query may be determined from the negative documents using noise-contrastive estimation. Such determined negative documents may be evaluated using a loss function during model training, thereby yielding a more robust model for search processing.

Abstract: Systems and methods are disclosed. An example method may be performed by one or more processors of a system and include retrieving case data indicating, for each respective case of a number of cases, one or more documents retrieved to assist a system user associated with the respective case, generating, from the case data, a case matrix including a plurality of rows each corresponding to a respective case of the number of cases and a plurality of columns each corresponding to the documents retrieved to assist the system user associated with the respective case, and identifying groups of similar cases among the plurality of cases based on a clustering process performed on at least a portion of the case matrix.

Abstract: An error management system can provide search results based on a received stack trace. For example, the error management system receives, from a client device, a search query including a stack trace. The error management system parses the search query to identify at least a first class and one or more errors associated with the first class, yielding a parsed search query. The error management system searches a class mapping table based on the parsed search query. The class mapping table includes a listing of classes and corresponding links to source code. The error management system searches an error discussion database based on the parsed search query. The error discussion database includes one or more user generated postings regarding programming errors. The error management system provides, to the client device, search results resulting from searching the class mapping table and the error discussion database.

Abstract: An electronic computing system receives a document and identifies steps for processing the document. Each step is associated with a respective machine learning model trained to generate an output. The system subdivides content of the document into a plurality of chunks. For each step and for each chunk, the system applies the respective machine learning model to the chunk to generate an output and a confidence value. The system compares the confidence value to a threshold value. In accordance with a determination that the confidence value exceeds the threshold value, the system stores the output as final output for the chunk and the step. In accordance with a determination that the confidence value does not exceed the threshold value, the system requests user input for the chunk and the step, and the system stores received user input as final output for the chunk and step.

Abstract: Techniques are provided for storage tier verification checks. A determination is made that a mount operation of an aggregate of a set of volumes stored within a multi-tier storage environment has completed. A first metafile and a second metafile are maintained to track information related to the storage of objects of a volume of the aggregate within a remote object store that is a tier of the multi-tier storage environment. A distributed verification is performed between the first metafile and the second metafile to identify an inconsistency. Accordingly, the first metafile and the second metafile are reconciled to address the inconsistency so that storage information within the first metafile and the second metafile are consistent.

Abstract: A highly accurate document search, particularly a search for a document relating to intellectual property, is achieved with an easy input method. A document search system includes a processing portion. The processing portion has a function of extracting a keyword included in text data, a function of extracting a related term of the keyword from words included in a plurality of pieces of first reference text analysis data, a function of giving a weight to each of the keyword and the related term, a function of giving a score to each of a plurality of pieces of second reference text analysis data on the basis of the weight, a function of ranking the plurality of pieces of second reference text analysis data on the basis of the score to generate ranking data, and a function of outputting the ranking data.

Abstract: The present invention discloses a method for disambiguating between authors with a same name on basis of network representation and semantic representation.

Abstract: Methods and apparatus related to determining points of interest based on content of one or more sources associated with a user. Relevance scores are determined for the points of interest based on additional information associated with the user and the content. Point of interest information that is related to the points of interest is provided based on the determined relevance scores.

Abstract: A mass file retrieval method and apparatus based on an AI training platform, and a device. The method includes: an AI training platform acquiring a retrieval task issued by a user, the AI training platform generating a retrieval thread flow according to the retrieval task, and controlling a service logic of a retrieval process according to the retrieval thread flow; and the AI training platform, performing sequential encoding on files in a database in units of folders, generating ordered queue folders, extracting retrieval key words from the retrieval task, and performing keyword retrieval on each ordered queue folder by combining a binary search and depth-first traversal.

Abstract: An adaptive document curation method. A machine learning model is trained on a first library of documents to assign staleness scores to documents, each indicating a confidence that the document to which it is assigned should be archived. If a document has a staleness score at least equal to a threshold, the document is identified as a candidate to be archived or deleted.

Abstract: Systems and methods described herein comprise a user interface for searching, analyzing, and interpreting documents obtained from computer databases. Exemplary systems and methods receive a user input and automatically identify, analyze, and interpret unmet technical needs and/or technical problems in specific areas of technology based on that input. Other exemplary systems and methods automatically identify, analyze, and interpret unmet technical needs and/or technical problems across numerous areas of technology based on similar user input. Other exemplary systems and methods receive user input and automatically identify, analyze, and interpret documents associated with a company to determine one or more technical capabilities of that company.

Abstract: A computer-implemented method is disclosed that includes operations of receiving document to be classified, performing pre-processing operations on the document resulting in generation of a tokenized document, performing word embedding operations on the tokenized document resulting in generation of a vectorized document, performing text similarity operations on the vectorized document and each of one or more vectorized topics resulting in a set of one or more similarity scores, wherein a first similarity score indicates a level of similarity between the vectorized document and a first vectorized topic, and wherein each vectorized topic represents one of a predetermined set of topics and classifying the document into one of the predetermined set of topics based on the set of one or more similarity scores. Performing the word embedding operations includes mapping each token of the remaining subset to a multi-dimensional vector, with each multi-dimensional vector representing a semantic meaning of a token.

Abstract: A computer-implemented method and a corresponding system for expanding an initially given input list of items from an application domain by similar or related items. The method operates on a computer-retrievable document repository containing a plurality of electronic documents. The method computes an expanded list of items from the initially given seed list by combining (i) methods for extracting candidate items that co-occur with already established items in selected lists contained in documents retrieved from said document store with (ii) knowledge-based inference tasks involving domain-specific logical facts and, in some embodiments also at least one domain-specific logical rule related to said application domain. By this combination, an improved item set expansion is obtained, which takes the particularities of the application domain into account, and which leads to a list expansion that may better fit a user's expectation.

Abstract: Systems and methods are disclosed. An example method may be performed by one or more processors of a system and include retrieving case data indicating, for each respective case of a number of cases, one or more documents retrieved to assist a system user associated with the respective case, generating, from the case data, a case matrix including a plurality of rows each corresponding to a respective case of the number of cases and a plurality of columns each corresponding to the documents retrieved to assist the system user associated with the respective case, and identifying groups of similar cases among the plurality of cases based on a clustering process performed on at least a portion of the case matrix.

Abstract: A management server obtains rating information indicating a second user's rating of a dish or restaurant, generates a first evaluation value of the first user's sense of taste based on a measurement test regarding the sense of taste, and generates a second evaluation value of the second user's sense of taste based on a measurement test regarding the sense of taste. One search filter for filtering dishes or restaurants based on rating results of the dishes or the restaurants rated by a second user who has a sense-of-taste characteristic similar to that of the first user is provided on a dish search screen. When the first user uses the search filter to filter dishes or restaurants, and when a first difference between the first and second evaluation values is in a first predetermined range, the second user's rating results are used to filter the dishes or the restaurant.

Abstract: Transaction-enabled methods for providing provable access to a distributed ledger with a tokenized instruction set for polymer production processes are described. A method may include accessing a distributed ledger comprising an instruction set for a polymer production process and tokenizing the instruction set. The method may further include interpreting an instruction set access request and providing a provable access to the instruction set. The method may further include providing commands to a production tool of the polymer production process and recording the transaction on the distributed ledger.