Systems and methods for generation and use of document analysis architectures are disclosed. A model builder component may be utilized to receiving user input data for labeling a set of documents as in class or out of class. That user input data may be utilized to train one or more classification models, which may then be utilized to predict classification of other documents. Trained models may be incorporated into a model taxonomy for searching and use by other users for document analysis purposes.

A method including: analyzing, by a computing device, a plurality of portions of a document; determining, by the computing device and based on the analyzing, a concept of each of the portions of the document; comparing, by the computing device, a title of the document with the concept of each of the portions of the document; determining, by the computing device and based on the comparing, an alignment of the concept of each of the portions of the document with the title; generating, by the computing device and based on the alignment, a propensity score for each of the portions of the document; and reordering, by the computing device and based on the propensity scores, the portions of the document from most aligned with the title to least aligned with the title.

A system, method and a computing device for performing a method for data augmentation allowing for document classification of a plurality of documents are disclosed. The system, method and computing device including a processor configured to convert the plurality of documents into images, a memory configured to store the images, the processor configured to obtain a vector representation for each page included in the plurality of documents, the processor configured to create a plurality of clusters from the images based on similarity, where each cluster of the plurality of clusters represents a distinct page format, the processor configured to select one image from each cluster of the plurality of clusters, the processor configured to compile the selected one image from each cluster of the plurality of clusters to create a logically complete document, the memory configured to store the logically complete document, and the processor configured to train the classification based on the complete document.

Methods, apparatus, systems, and articles of manufacture are disclosed that determine related content. An example apparatus includes processor circuitry to generate a segment-level graph by sampling segment-level edges among segment nodes representing text segments, the segment-level graph including segment node embeddings representing features of the segment nodes; cluster the text segments to form entities by applying a first GAN based model to the segment-level graph to update the segment node embeddings; generate a multi-level graph by (a) generating an entity-level graph including hypernodes representing the entities and sampled entity edges connecting ones of the hypernodes, and (b) connecting the segment nodes to respective ones of the hypernodes using relation edges; generate hypernode embeddings by propagating the updated segment node embeddings using a relation graph; and cluster the entities by product by applying a second GAN based model to the multi-level graph, the multi-level graph to generate updated hypernode embeddings.

A recognition method includes the following steps. A text is analyzed by a language recognition network to generate an entity feature, a relation feature and an overall feature. An input image is analyzed by an object detection network to generate candidate regions. Node features, aggregated edge features and compound features are generated by an enhanced cross-modal graph attention network according to the entity feature, the relation feature, the candidate regions and the overall feature. The entity feature and the relation feature are matched to the node features and the aggregated edge features to generate the first scores. The overall feature is matched to the compound features to generate second scores. Final scores corresponding to the candidate regions are generated according to the first scores and the second scores.

The invention relates to a computer-implemented method for automatically extracting data from a digital image comprising a graphical representation of quantitative data. The method comprises: Basic graphical objects are detected and structural primitives determined comprising grouping the basic graphical objects based on geometric relations. A semantic label is assigned to each of the structural primitives. A spatial data region of the graphical representation is determined using the semantic labels of the structural primitives. Quantitative data values are extracted which are represented by structural primitives within the data region which are assigned with first semantic labels identifying the respective structural primitives to represent quantitative data. The extracted quantitative data values are provided in units of pixels according to an image coordinate system. The extracted quantitative data values are transformed from the image coordinate system to a coordinate system of physical units of the quantitative data represented by the graphical representation.

A computing device may access visually rich documents comprising an image and metadata. A graph, based on the image or metadata, can be generated for a visually rich document. The graph's nodes can correspond to words from the visually rich document. Features for nodes can be determined by the device. The device may generate model labeled graphs by assigning a pseudo-label to nodes using a pretrained model. The device may generate a plurality of graph labeled graphs by assigning a pseudo-label to nodes by matching a first node from a first graph to at least a second node from a second graph. The device may generate a plurality of updated graphs by cross referencing labels from the model labeled graphs and the graph labeled graphs. Until a change in labels is below a threshold, a model can be trained to perform key-value extraction using the updated graphs.

Methods, apparatus, and processor-readable storage media for character-based representation learning for table data extraction using artificial intelligence techniques are provided herein. An example computer-implemented method includes identifying, from unstructured documents comprising tabular data, items of text and corresponding document position information using artificial intelligence-based text extraction techniques; generating an intermediate output by implementing character embedding with respect to the unstructured documents using an artificial intelligence-based encoder; determining structure-related information for the unstructured documents using one or more artificial intelligence-based graph-related techniques by inferring columns from the tabular data; generating a character-based representation of the unstructured documents using an artificial intelligence-based decoder by converting the inferred columns into one or more line items; classifying portions of the character-based representation using artificial intelligence-based statistical modeling techniques; and performing one or more automated actions based on the classifying.

Generating a cybersecurity risk model using sparse data is disclosed, including: obtaining signals associated with a cybersecurity risk, wherein the obtained signals include technographic signals and query derived signals obtained from queries; generating pseudo signals based at least in part on a priori factors relating to the cybersecurity risk; and combining the pseudo signals and the obtained signals into a Bayesian model indicating the cybersecurity risk.

The invention relates to a computer-implemented method for automatically extracting data from a digital image comprising a graphical representation of quantitative data. The method comprises: Basic graphical objects are detected and structural primitives determined comprising grouping the basic graphical objects based on geometric relations. A semantic label is assigned to each of the structural primitives. A spatial data region of the graphical representation is determined using the semantic labels of the structural primitives. Quantitative data values are extracted which are represented by structural primitives within the data region which are assigned with first semantic labels identifying the respective structural primitives to represent quantitative data. The extracted quantitative data values are provided in units of pixels according to an image coordinate system. The extracted quantitative data values are transformed from the image coordinate system to a coordinate system of physical units of the quantitative data represented by the graphical representation.