Abstract: A system and method for correcting image data is disclosed. The method includes receiving one or more documents from one or more electronic mediums. The method further includes determining a primary character and one or more alternate characters corresponding to the mis-captured character image, extracting one or more confident instances of the primary character and the one or more alternate characters from the one or more documents and generating one or more scores corresponding to the primary character and the one or more alternate characters. Further, the method includes predicting a correct character corresponding to the mis-captured character image by using a trained image prediction-based ML model and automatically replacing the mis-captured character image with the predicted correct character.

Abstract: A system and method for or facilitating correction of data in documents is disclosed. The method includes receiving one or more documents from and scanning the received one or more documents by using a document processing system for obtaining one or more mis-captured data fields. The method further includes obtaining a historical correction data and determining one or more deltas based on the obtained historical correction data by using a trained data correction-based ML model. Further, the method includes parsing the determined one or more deltas into one or more datasets, generating one or more correct data fields corresponding to the one or more mis-captured data fields and automatically replacing the one or more mis-captured data fields with the generated one or more correct data fields based on one or more predefined rules.

Abstract: Novel tools and techniques are provided for implementing automatic document pattern recognition and analysis. In various embodiments, a computing system might receive one or more first documents containing first remittance information. The computing system might analyze the one or more first documents to extract first remittance information from the first document by obtaining one or more historical documents and correlating the one or more first documents to the one or more historical documents. Next, based on the correlation of the one or more first documents to the one or more historical documents, the computing system might determine at least one historical document that correlates to the one or more first documents. Additionally, the computing system might extract the first remittance information from the one or more first documents based on the at least one historical document.

Abstract: A system and method for or facilitating correction of data in documents is disclosed. The method includes receiving one or more documents from and scanning the received one or more documents by using a document processing system for obtaining one or more mis-captured data fields. The method further includes obtaining a historical correction data and determining one or more deltas based on the obtained historical correction data by using a trained data correction-based ML model. Further, the method includes parsing the determined one or more deltas into one or more datasets, generating one or more correct data fields corresponding to the one or more mis-captured data fields and automatically replacing the one or more mis-captured data fields with the generated one or more correct data fields based on one or more predefined rules.

Abstract: Novel tools and techniques are provided for implementing identification and removal of noise from documents, and, more particularly, to methods, systems, and apparatuses for implementing identification and removal of noise from financial documents using one or more machine learning algorithms. In various embodiments, computing system might receive a document. The computing system might detect, using one or more machine learning algorithms, that noise exists in the document. Based on the detection that noise exists in the document, the computing system might remove the noise from the document. Once the noise is removed from the document, the computing system might generate a copy of the document with the noise removed while retaining important or useful information contained in the document.

Abstract: Machine learning (ML) systems and methods for fact extraction and claim verification are provided. The system receives a claim and retrieves a document from a dataset. The document has a first relatedness score higher than a first threshold, which indicates that ML models of the system determine that the document is most likely to be relevant to the claim. The dataset includes supporting documents and claims including a first group of claims supported by facts from more than two supporting documents and a second group of claims not supported by the supporting documents. The system selects a set of sentences from the document. The set of sentences have second relatedness scores higher than a second threshold, which indicate that the ML models determine that the set of sentences are most likely to be relevant to the claim. The system determines whether the claim includes facts from the set of sentences.

Abstract: In implementations of generating descriptions of image relationships, a computing device implements a description system which receives a source digital image and a target digital image. The description system generates a source feature sequence from the source digital image and a target feature sequence from the target digital image. A visual relationship between the source digital image and the target digital image is determined by using cross-attention between the source feature sequence and the target feature sequence. The system generates a description of a visual transformation between the source digital image and the target digital image based on the visual relationship.

Abstract: A technique is described herein for effectively exposing annotation information in a document for use by various applications. The technique involves generating a tag tree data structure that identifies a collection of tag elements associated with a document. The technique also generates an overlay data structure that identifies a collection of annotations associated with the document. The overlay data structure also links the annotations to corresponding parts identified in the tag tree data structure. The technique then uses the tag tree data structure and the overlay data structure to provide information to a document-consuming component that conveys an order in which one or more annotations appear in the document relative to one or more parts in the document. According to one illustrative aspect, at least one annotation described by the overlay data structure is an active annotation, corresponding to a transient annotation that has not been saved.

Abstract: In implementations of generating descriptions of image relationships, a computing device implements a description system which receives a source digital image and a target digital image. The description system generates a source feature sequence from the source digital image and a target feature sequence from the target digital image. A visual relationship between the source digital image and the target digital image is determined by using cross-attention between the source feature sequence and the target feature sequence. The system generates a description of a visual transformation between the source digital image and the target digital image based on the visual relationship.

Abstract: A technique is described herein for effectively exposing annotation information in a document for use by various applications. The technique involves generating a tag tree data structure that identifies a collection of tag elements associated with a document. The technique also generates an overlay data structure that identifies a collection of annotations associated with the document. The overlay data structure also links the annotations to corresponding parts identified in the tag tree data structure. The technique then uses the tag tree data structure and the overlay data structure to provide information to a document-consuming component that conveys an order in which one or more annotations appear in the document relative to one or more parts in the document. According to one illustrative aspect, at least one annotation described by the overlay data structure is an active annotation, corresponding to a transient annotation that has not been saved.

Abstract: Techniques are disclosed for making an electronic document easier to use based on prior interactions with the same or a similar document by other users. An electronic document is presented to one or more users in an interactive environment. Interactions between the users and the document can be recorded as usage data. The usage data may represent one or more operations performed on the electronic document by the users. Based on the usage data, an enhanced user interaction feature associated with the document is configured. The electronic document and the enhanced user interaction feature are then presented to another user in another interactive environment. The enhanced user interaction feature makes using the document easier than it would be if the feature was not present, particularly for users who are unfamiliar with the document.

Abstract: A computer implemented method and apparatus for detecting, validating, and correlating form-fields in a scanned document. The method comprises displaying a plurality of interactive form-fields associated with a scanned document, wherein each interactive form-field in the plurality of form-fields is defined by a location in the document where one or more previous users entered information on the scanned document, and a data type for the entered information.

Abstract: A computer-implemented technique can include receiving, at a server including one or more processors, a source word in a source language. The technique can include determining, at the server, one or more potential translations for the source word in a target language different than the source language. The technique can include determining, at the server, one or more synonyms for each of the one or more potential translations to obtain a plurality of potential translations. The technique can include determining, at the server, one or more translation clusters using the plurality of potential translations and a clustering algorithm. Each translation cluster can contain all of the plurality of potential translations that have a similar denotation and each of the plurality of translations that have a similar denotation can be included in a specific translation cluster. The technique can also include outputting, at the server, the one or more translation clusters.

Abstract: Techniques are disclosed for making an electronic document easier to use based on prior interactions with the same or a similar document by other users. An electronic document is presented to one or more users in an interactive environment. Interactions between the users and the document can be recorded as usage data. The usage data may represent one or more operations performed on the electronic document by the users. Based on the usage data, an enhanced user interaction feature associated with the document is configured. The electronic document and the enhanced user interaction feature are then presented to another user in another interactive environment. The enhanced user interaction feature makes using the document easier than it would be if the feature was not present, particularly for users who are unfamiliar with the document.

Abstract: A computer implemented method and apparatus for detecting, validating, and correlating form-fields in a scanned document. The method comprises displaying a plurality of interactive form-fields associated with a scanned document, wherein each interactive form-field in the plurality of form-fields is defined by a location in the document where one or more previous users entered information on the scanned document, and a data type for the entered information.

Abstract: A computer-implemented technique can include receiving, at a server including one or more processors, a source word in a source language. The technique can include determining, at the server, one or more potential translations for the source word in a target language different than the source language. The technique can include determining, at the server, one or more synonyms for each of the one or more potential translations to obtain a plurality of potential translations. The technique can include determining, at the server, one or more translation clusters using the plurality of potential translations and a clustering algorithm. Each translation cluster can contain all of the plurality of potential translations that have a similar denotation and each of the plurality of translations that have a similar denotation can be included in a specific translation cluster. The technique can also include outputting, at the server, the one or more translation clusters.