Abstract: In some implementations, a system may receive a data stream input. The system may identify one or more sets of content parameters associated with the data stream input, each set of content parameters being associated with a different entity. The system may determine a ruling set of content parameters based on the one or more sets of content parameters. The system may selectively modify the data stream input based on a determination of whether a condition indicated by the ruling set of content parameters is satisfied in the data stream input. The system may provide a data stream output resulting from the selective modification of the data stream input.

Abstract: A method and related system for generating document embeddings within an embedding space based on a set of structured documents by determining (i) a first vector based on a first segment of a first document and (ii) a second vector based on a second segment of the first document and updating association vectors indicating the second segment based on a distance between the first and second vectors. The method also includes generating a document embedding based on the association vectors, generating a candidate vector based on a candidate document, and determining a result indicating that a second distance between the candidate vector and a first document embedding satisfies a document embedding distance threshold. The method may also include generating a new document by providing, to a text generation model, a portion of the candidate document and a portion of the second segment of the first document.

Abstract: Systems described herein may provide responses to chatbot prompts that correspond to both a user's preferences and accepted views of society. A chat recommendation server may receive a prompt from a user device. The chat recommendation server may determine a general Overton window and a user-specific Overton window associated with the prompt. The chat recommendation server may generate a plurality of candidate response using the first machine learning model, input the prompt and the plurality of candidate responses to a second machine learning model, and receive, as output from the second machine learning model, a polarization score for each of the plurality of candidate responses. Based on the polarization scores, a recommended response may be selected which minimizes a distance between the user-specific Overton window and the general Overton window. Accordingly, the recommended response may be displayed on the user device.

Abstract: A processor of a card may determine a predetermined voltage associated with the card. The processor may receive an indication of a voltage received from a terminal, wherein the card is powered by the voltage received from the terminal. The processor may determine that the voltage received from the terminal is not equal to the predetermined voltage associated with the card. The processor may reject, based on the determination that the voltage received from the terminal is not equal to the predetermined voltage associated with the card, a request from the terminal to provide payment information associated with the card.

Abstract: Systems and methods are disclosed herein for generating updated descriptions of items based on analyzing candidate embeddings of semantic representations of item descriptions. The system may obtain a text file describing an item. The system may provide the text file to a generative language model to generate semantic representations of the text file. The system may generate, based on the text file, candidate embeddings in an embedding space. The system may obtain embeddings associated with existing items. The system may determine subsets of the embeddings within a threshold distance. The system may compare the subsets. The system may determine attributes associated with a candidate embedding based on the comparison. The system may generate an updated text file based on the attributes.

Abstract: In some aspects, a computing system may generate uninformative features that may be added to a dataset of real features to use as a baseline for determining the quality of an explanation of model output. The uninformative features may be features that do not correlate with what a model is tasked with predicting (e.g., the uninformative features may be random values), and the real features may be informative and correlate with what the model is tasked with predicting (e.g., variables of a dataset sample). A machine learning model may be trained on a dataset that includes both the real features and the uninformative features. The computing system may generate feature attributions for model output, which may include feature attributions for the uninformative features and the real features in the dataset.

Abstract: Systems and methods for triggering token alerts. In some aspects, the system, after determining that the probability that an authentication request from an authentication token is associated with a malicious activity is above a threshold, determines whether a user device associated with the authentication token is within a threshold distance of the authentication token. In response to determining that the authentication token is not within the threshold distance of the user device, the system declines the authentication request and transmits an alert request to the authentication token to emit an audio signal from a speaker included in the authentication token.

Abstract: In some aspects, a computing system may aggregating multiple counterfactual samples so that machine learning explanations can be generated for sub-populations. In addition, methods and systems described herein use machine learning and counterfactual samples to determine text to use in an explanation for a model's prediction. A computing system may also train machine learning models to not only determine whether a request to perform an action should be accepted, but also to generate output that is consistent with output generated by previous machine learning models. Further, a computing system may generate counterfactual samples based on user preferences. A computing system may obtain preferences and then apply a penalty or adjustment parameter such that when a counterfactual sample is created, the computing system is forced to change one or more features indicated by the preferences to create the counterfactual sample.

Abstract: In some aspects, a computing system may use a surrogate machine learning model to detect whether a production or other machine learning model has a tendency to generate different output depending on which subpopulation a particular sample belongs to. The surrogate machine learning model may be trained using features/outputs that are not included in the data used by the production model. For example, by using demographic information in lieu of the original labels of a dataset that was used to train a production model, a surrogate model may be used to detect whether the production model is able to discern one or more characteristics associated with but not present in a sample using other features of the dataset. Output of the surrogate machine learning model may be clustered to detect whether certain subpopulations are treated differently by the production model.

Abstract: Methods and systems are described herein for generating recommendations for counterfactual explanations to computer alerts that are automatically detected by a machine learning algorithm. The methods and systems use an artificial neural network architecture that trains a hybrid classifier and autoencoder. For example, one model (or artificial neural network), which is a classifier, is trained to make predictions. A second model (or artificial neural network), which is an autoencoder, is trained to reconstruct its inputs. As the second model is trained to reconstruct its inputs means, the second model is implicitly trained to determine what in-sample data looks like. By combining these networks and train them jointly, the system generates predictions (e.g., counterfactual explanations) that are in-sample.

Abstract: A computing system may generate a first set of importance metrics (e.g., scores or values) for a model. The importance metrics may be generated using an explainable artificial intelligence technique, and an individual importance metric may indicate how influential a corresponding feature is for a decision made by a model. The computing system may determine an important feature and create a modified dataset by removing the important feature from the dataset. The computing system may train the model on the modified dataset and evaluate the performance of the model to determine the effect of removing the feature (e.g., which may indicate how important the feature is to output generated by the model). This process may be repeated for additional features and additional performance metrics may be obtained.

Abstract: A computer-implemented method for semantically interpreting a real-world environment may include: receiving, via a user device, multimodal input that includes a plurality of modalities of data regarding an environment associated with a user of the user device; standardizing the plurality of modalities into a uniform data format to generate uniform multimodal context data; generating an embedding of the uniform multimodal context data; and determining a content entry predicted to be relevant to the environment associated with the user based on the generated embedding.

Abstract: In some implementations, a system may receive a data stream input. The system may identify one or more sets of content parameters associated with the data stream input, each set of content parameters being associated with a different entity. The system may determine a ruling set of content parameters based on the one or more sets of content parameters. The system may selectively modify the data stream input based on a determination of whether a condition indicated by the ruling set of content parameters is satisfied in the data stream input. The system may provide a data stream output resulting from the selective modification of the data stream input.

Abstract: Disclosed embodiments may include a system for providing automated bill splitting. The system may receive speech data. The system may identify, from the speech data and using natural language processing, one or more users. The system may determine, from the speech data and using natural language processing, orders of the one or more users. The system may determine, from the speech data and using natural language processing, rules for the orders of the one or more users. The system may process one or more payments for the orders based on the rules and one or more credentials associated with the one or more users.

Abstract: Methods and systems are described herein for updating a transformer model to identify key events. In some embodiments, a request to authorize a user may be received including a sequence of events representing interactions of the user with a server. The sequence of events can be provided to a first artificial intelligence model, trained for a particular use case, to obtain a classification result indicating whether the request should be granted. In addition to being provided to the first artificial intelligence model, the sequence of events may be provided to a second artificial intelligence model trained to identify a subset of events from the sequence of events that most heavily contribute to the prediction by the first artificial intelligence model of the classification result.

Abstract: Systems and methods for visual manipulation and execution of machine learning models rendered in a three-dimensional space. In some aspects, the system receives configuration data representing a machine learning model and generates a three-dimensional representation of the machine learning model by (1) generating virtual objects corresponding to nodes and edges of the model and (2) configuring values of virtual object parameters for virtual objects based on associated weight matrices from the configuration data. The system detects a user gesture that indicates a command to perform a modification of the machine learning model and, responsive to detecting the user gesture, causes execution of a modified machine learning model. The system generates a new three-dimensional representation of the modified machine learning model.

Abstract: Methods and systems are described herein for building and executing an artificial intelligence model that predicts whether a computer code update is likely to cause an issue. In particular, the system may receive a potential code update and identify update parameters associated with the potential code update. The system may then input the potential code update and the update parameters into a machine learning model to receive a prediction about the potential code update to be displayed to a user.

Abstract: Methods and systems described herein relate to the creation of a digital repository of artificial intelligence models that allows users to determine their individual fairness metric. More specifically, the methods and systems provide this digital repository by storing it on a blockchain network and tracking any changes made to the model and/or its fairness metric.

Abstract: In some implementations, a verification device may receive a request for an electronic exchange of information. The verification device may receive geographic location information associated with the request for the electronic exchange of information. The verification device may determine, using a machine learning model and based on the geographic location information, a validity of the request for the electronic exchange of information. The verification device may determine whether to execute the electronic exchange of information based on the determination of the validity of the request for the electronic exchange of information. The verification device may perform one of executing the electronic exchange of information or rejecting the request for the electronic exchange of information based on the determination of whether to execute the electronic exchange of information.

Abstract: Systems and methods are disclosed herein for generating updated descriptions of items based on analyzing candidate embeddings of semantic representations of item descriptions. The system may obtain a text file describing an item. The system may provide the text file to a generative language model to generate semantic representations of the text file. The system may generate, based on the text file, candidate embeddings in an embedding space. The system may obtain embeddings associated with existing items. The system may determine subsets of the embeddings within a threshold distance. The system may compare the subsets. The system may determine attributes associated with a candidate embedding based on the comparison. The system may generate an updated text file based on the attributes.