Abstract: The disclosed computer-implemented method may include determining a taxonomy of an object from its textual description and also standardized attributes of the object from the description and the taxonomy using a language model, according to embodiments. The method may also include building a graph data structure by using the standardized attributes for a node and connecting the node to other nodes using edges for common attributes. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Techniques are disclosed for altering tokenizer training data causing a tokenizer to generate an improved library of tokens from the altered training data for tokenizing new source data during training of a machine learning model e.g., for natural language processing. A system retrieves, from a source database, a set of original text data that includes characters. The server system identifies a plurality of strings included in the set of original text data. The system alters strings included in the original text data to generate altered text data by selecting current characters included in the strings to be altered. The system trains, using the altered text data, a tokenizer by assigning a threshold to the tokenizer and inputting the altered text data into the tokenizer, where the tokenizer generates tokens for strings in the altered text data based on the threshold. The system stores the trained tokenizer.

Abstract: Techniques are disclosed for altering tokenizer training data causing a tokenizer to generate an improved library of tokens from the altered training data for tokenizing new source data during training of a machine learning model e.g., for natural language processing. A system retrieves, from a source database, a set of original text data that includes characters. The server system identifies a plurality of strings included in the set of original text data. The system alters strings included in the original text data to generate altered text data by selecting current characters included in the strings to be altered. The system trains, using the altered text data, a tokenizer by assigning a threshold to the tokenizer and inputting the altered text data into the tokenizer, where the tokenizer generates tokens for strings in the altered text data based on the threshold. The system stores the trained tokenizer.

Abstract: Methods and systems are presented for providing explainable artificial intelligence for a deep-learning model on a global level. Multiple surrogate models are generated based on characteristics of the deep-learning model, where each surrogate model is configured to mimic a behavior of the deep-learning model with respect to one of the output dimensions associated with the deep-learning model. Simulations are performed on the surrogate models. Based on the simulation results, importance scores are calculated for each input feature of the deep-learning model. An importance score calculated for an input feature represents an extent to which the input feature contributes to a corresponding one of the output dimensions associated with the deep-learning model. The importance scores may then be used to modify the characteristics of the deep-learning model or other downstream machine learning models.

Abstract: The disclosed computer-implemented method may include determining a taxonomy of an object from its textual description and also standardized attributes of the object from the description and the taxonomy using a language model, according to embodiments. The method may also include building a graph data structure by using the standardized attributes for a node and connecting the node to other nodes using edges for common attributes. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Techniques are disclosed pertaining to classifying an event. A computer system may receive event information identifying event features of the event. The computer accesses a graph neural network model trained based on modified graph data describing a graph data structure having a plurality of event nodes representing events and a plurality of feature nodes representing combinations of event features. The graph data structure includes a virtual linking node that links a first node group to a second node group to enable a first set of event nodes of the first group to influence a second set of event nodes of the second group during a training phase in which the graph neural network model is trained based on the modified graph data. The computer system generates an event node representation of the event and then classifies the event based on the event node representation and the graph neural network model.

Abstract: A graph network of a service provider is accessed. The graph network includes a plurality of nodes interconnected by a plurality of edges. A plurality of sub-graphs is generated. Each of the sub-graphs corresponds to a different portion of the graph network. Each of the sub-graphs includes a different subset of the plurality of nodes. A directional flow for information exchanges is defined between the nodes of each of the sub-graphs. A graph neural network (GNN) model is trained based on the defined directional flow. The trained GNN model is utilized to generate one or more predictions.

Abstract: Techniques are disclosed for altering tokenizer training data causing a tokenizer to generate an improved library of tokens from the altered training data for tokenizing new source data during training of a machine learning model e.g., for natural language processing. A system retrieves, from a source database, a set of original text data that includes characters. The server system identifies a plurality of strings included in the set of original text data. The system alters strings included in the original text data to generate altered text data by selecting current characters included in the strings to be altered. The system trains, using the altered text data, a tokenizer by assigning a threshold to the tokenizer and inputting the altered text data into the tokenizer, where the tokenizer generates tokens for strings in the altered text data based on the threshold. The system stores the trained tokenizer.

Abstract: Methods and systems are presented for providing explainable artificial intelligence for a deep-learning model on a global level. Multiple surrogate models are generated based on characteristics of the deep-learning model, where each surrogate model is configured to mimic a behavior of the deep-learning model with respect to one of the output dimensions associated with the deep-learning model. Simulations are performed on the surrogate models. Based on the simulation results, importance scores are calculated for each input feature of the deep-learning model. An importance score calculated for an input feature represents an extent to which the input feature contributes to a corresponding one of the output dimensions associated with the deep-learning model. The importance scores may then be used to modify the characteristics of the deep-learning model or other downstream machine learning models.