Abstract: This specification provides a training method of a hybrid graph neural network model. The hybrid graph neural network model includes an encoding function and a decoding function. The method includes the following: using instances corresponding to all targets in training samples and several nearest neighbors of the instances as nodes in a graph, a graph representation vector of each instance is generated by using the encoding function based on graph data of all the instances. t rounds of training are performed on a decoding parameter; and in each round, bs targets are extracted from training samples, a predicted quantity of each target is generated by using the decoding function based on the graph representation vector of the instance corresponding to each target and non-graph data corresponding to each target, and the decoding parameter is optimized based on a loss quantity of the current round that is determined by the predicted quantities and label quantities of the bs targets in the current round.

Abstract: A method for generating files, in particular text and audio files as well as files for computer games or videos. The processing of very large quantities of texts, which differ in content and structure, is thereby ensured. For this purpose, existing data is filtered and processed (cleaned) in a first step, and subsequently a training corpus is generated from the processed data, which is adjusted to the desired results. The filtered and processed data is standardized.

Abstract: Systems and techniques are provided for performing object detection using a machine learning model with a transformer architecture. An example method can include receiving a plurality of tokens corresponding to segmented sensor data; identifying, by a halting module within the machine learning model, at least one halted token from the plurality of tokens, wherein the at least one halted token is excluded from a plurality of non-halted tokens provided as input to a subsequent layer during inference of the machine learning model; and detecting, by the machine learning model, at least one detected object based at least on the plurality of non-halted tokens.

Abstract: Provided is a system for detecting an anomaly in a multivariate time series that includes at least one processor programmed or configured to receive a dataset of a plurality of data instances, wherein each data instance comprises a time series of data points, determine a set of target data instances based on the dataset, determine a set of historical data instances based on the dataset, generate, based on the set of target data instances, a true value matrix, a true frequency matrix, and a true correlation matrix, generate a forecast value matrix, a forecast frequency matrix, and a forecast correlation matrix based on the set of target data instances and the set of historical data instances, determine an amount of forecasting error, and determine whether the amount of forecasting error corresponds to an anomalous event associated with the dataset of data instances. Methods and computer program products are also provided.

Abstract: A method (700) for predicting, for each device included in a set of devices, whether the device will change state at a particular future point in time. The method includes, for a first device within the set of devices, obtaining a first state value indicating the current state of the first device. The method also includes, for a second device within the set of devices, obtaining a second state value indicating the current state of the second device. The method also includes forming an input vector, the input vector comprising the first state value, the second state value, and a temporal feature (e.g., a set of one or more time values indicating the current time). The method also includes inputting the input vector into a trained machine learning (ML) model.

Abstract: Disclosed herein is a method and device for absolute average deviation (AAD) pooling for a convolutional neural network accelerator. AAD utilizes the spatial locality of pixels using vertical and horizontal deviations to achieve higher accuracy, lower area, and lower power consumption than mixed pooling without increasing the computational complexity. AAD achieves 98% accuracy with lower computational and hardware costs compared to mixed pooling, making it an ideal pooling mechanism for an IoT CNN accelerator.

Abstract: An operating method for a neural processing unit is provided. The method includes determining, by a controller, that an operation performed in a first convolution layer is a transpose convolution operation, dividing, by the controller, a kernel used for the transpose convolution operation into a plurality of sub-kernels, and performing, by at least one processing element, a convolution operation between an input feature map and each of the plurality of sub-kernels in the first convolution layer.

Abstract: A computer-implemented method for determining a classification and/or regression result based on a provided input signal. The method includes: providing a first part, configured to denoise the provided input signal based on the input signal and a randomly drawn first value; randomly drawing a plurality of first values; determining, by the first part, a plurality of denoised signals, wherein denoised signals are each determined based on the provided input signal and a first value from the plurality of first values; determining, by a model, a plurality of predicted values based on the denoised values, wherein each predicted value characterizes a classification of a denoised signal or a regression results based on a denoised signal; providing an aggregated signal characterizing an aggregation of the predicted values, wherein the aggregated signal characterizes the classification and/or regression result determined by the method.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium. In one aspect, a method includes the actions of receiving a request to perform computations for a neural network on a hardware circuit having a matrix computation unit, the request specifying a transpose operation to be performed on a first neural network matrix; and generating instructions that when executed by the hardware circuit cause the hardware circuit to transpose the first neural network matrix by performing first operations, wherein the first operations include repeatedly performing the following second operations: for a current subdivision of the first neural network matrix that divides the first neural network matrix into one or more current submatrices, updating the first neural network matrix by swapping an upper right quadrant and a lower left quadrant of each current submatrix, and subdividing each current submatrix into respective new submatrices to update the current subdivision.

Abstract: Provided are an integrated circuit chip apparatus and a related product, the integrated circuit chip apparatus being used for executing a multiplication operation, a convolution operation or a training operation of a neural network. The present technical solution has the advantages of a small amount of calculation and low power consumption.

Abstract: A neuromorphic computing circuit includes a plurality of memristors that function as synapses. The neuromorphic computing circuit also includes a superconducting quantum interference device (SQUID) coupled to the plurality of memristors. The SQUID functions as a neuron such that the plurality of memristors and the SQUID form a neural unit of the neuromorphic computing circuit.

Abstract: Some embodiments provide a three-dimensional (3D) circuit structure that has two or more vertically stacked bonded layers with a machine-trained network on at least one bonded layer. As described above, each bonded layer can be an IC die or an IC wafer in some embodiments with different embodiments encompassing different combinations of wafers and dies for the different bonded layers. The machine-trained network in some embodiments includes several stages of machine-trained processing nodes with routing fabric that supplies the outputs of earlier stage nodes to drive the inputs of later stage nodes. In some embodiments, the machine-trained network is a neural network and the processing nodes are neurons of the neural network. In some embodiments, one or more parameters associated with each processing node (e.g., each neuron) is defined through machine-trained processes that define the values of these parameters in order to allow the machine-trained network (e.g.

Abstract: Described are methods, devices and applications for learning noise distribution on information from any data processing method. In an embodiment of the described technology, a method includes determining an amount of shredding used in a shredding operation by which source data is converted into shredded data, and transferring the shredded data over an external network to a remote server for a data processing task. The shredding reduces the information content and incurs a limited degradation to an accuracy of the data processing task due to the shredding operation.

Abstract: A method comprises receiving event-based data, extracting one or more attributes from the event-based data, and analyzing the one or more attributes to classify whether the one or more attributes comprise personally identifiable information. The analyzing is performed using one or more machine learning models. The event-based data corresponds to one or more events where the one or more attributes are added to at least one of a database and an application.

Abstract: A processing system including at least one processor may obtain at least a first objective associated with a demeanor of at least a first participant for a conversation and may activate at least one machine learning model associated with the at least the first objective. The processing system may then apply a conversation content of the at least the first participant as at least a first input to the at least one machine learning model and perform at least one action in accordance with an output of the at least one machine learning model.

Abstract: In an example embodiment, a neural network is trained to classify three-dimensional spatial-channel images in a manner that allows the training data to include two-dimensional images. Specifically, rather than redesign the neural network completely to accept three-dimensional images as input, two-dimensional slices of three-dimensional spatial-channel images are input in groupings that match the groupings that a two-dimensional image would be grouped as in the neural network. For example, if the neural network is designed to accept RGB images, it therefore is designed to accept images in groupings of three (a red component image, a green component image, and a blue component image). In such a case, the two-dimensional slices of the three-dimensional spatial-channel images will also be grouped in grouping of three so the neural network can accept them. Thus, a neural network originally designed to classify two-dimensional color images can be modified to classify three-dimensional spatial-channel images.

Abstract: Embodiments of the present disclosure provide a method of training a neural network model for controlling an operation of a system represented by partial differential equations (PDEs). The method comprises collecting digital representation of time series data indicative of measurements of the operation of the system at different instances of time. The method further comprises training the neural network model having an autoencoder architecture including an encoder to encode the digital representation into a latent space, a linear predictor to propagate the digital representation into the latent space, and a decoder to decode the digital representation to minimize a loss function including a prediction error between outputs of the neural network model decoding measurements of the operation at an instant of time and measurements of the operation collected at a subsequent instance of time, and a residual factor of the PDE having eigenvalues dependent on parameters of the linear predictor.

Abstract: There is disclosed a computer-implemented method for training a neural network-based system. The method comprises receiving a training data item and target data associated with the training data item. The training data item is processed using an encoder to generate an encoding of the training data item. A subset of neural networks is selected from a plurality of neural networks stored in a memory based upon the encoding; wherein the plurality of neural networks are configured to process the encoding to generate output data indicative of a classification of an aspect of the training data item. The encoding is processed using the selected subset of neural networks to generate the output data. An update to the parameters of the selected subset of neural networks is determined based upon a loss function comprising a relationship between the generated output data and the target data associated with the training data item.

Abstract: An industrial machine (123) may not have a sensor for a particular parameter, so that a computer uses a neural network (473) to virtualize the missing sensor. The computer trains the neural network (373) to provide a parameter indicator (Z?) of a further process parameter (173, z) for the industrial machine (123) with steps that comprise receiving measurement time-series with historical measurement data from reference machines, obtaining transformation rules by processing the time-series to feature series that are invariant to domain differences of the reference machines, transforming time-series by using the transformation rules, receiving a uni-variate time-series of the further process parameter (z), and training the neural network with features series at the input, and with the uni-variate time-series at the output.

Abstract: The present invention is a scalable VR game for generating synthetic data. The synthetic data includes RGB and XYZ images from digital twins of entire cities at up to 01-meter accuracy and are updated by Google Maps, synthetic sensor, pilot data (including both performance and haptics such as haptics, response time, button presses, heartbeat, skin conductance, pupil dilation, etc.), and weather data, to be used for SLAM (Simultaneous Location and Mapping) and regulatory approval of EVTOL (Electric Vertical Take-off and Landing Aircraft), autonomous cars and robots. The synthetic data is generated from collected RGB images during the game. The VR-based game is interfaced with a full-stack neuromorphic backend platform for generating a VR-based anatomically accurate NeuroSLAM algorithm.

Abstract: A user equipment (UE) may receive, from a network node one or more reference signals and perform one or more measurements using the one or more reference signals. The UE may compress the one or more measurements into one or more measurement results and transmit the one or more measurement results to a server. The UE may request, from the server, at least one of one or more identifiers (IDs) or one or more models associated with the one or more IDs. The UE may receive, from the server, the at least one of the one or more IDs or one or more models, wherein the one or more IDs are provided based on the one or more measurement results. The UE may transmit, to the network node, an ordered list of the one or more IDs, receive a response from the network node indicating selection of an ID of the one or more IDs, and communicate with the network node using the ID.

Abstract: Disclosed is a processor implemented method that includes calculating a quantization error for each channel of a neural network using activation data output from a first layer of the neural network and a quantization scale of a second layer connected to the first layer, calculating a final loss using a regularization loss term determined based on the quantization error for each channel, and updating a batch norm parameter of the first layer in a direction to decrease the final loss.

Abstract: Systems and methods for generating embeddings for nodes of a corpus graph are presented. More particularly, operations for generation of an aggregated embedding vector for a target node is efficiently divided among operations on a central processing unit and operations on a graphic processing unit. With regard to a target node within a corpus graph, processing by one or more central processing units (CPUs) is conducted to identify the target node's relevant neighborhood (of nodes) within the corpus graph. This information is prepared and passed to one or more graphic processing units (GPUs) that determines the aggregated embedding vector for the target node according to data of the relevant neighborhood of the target node.

Abstract: An apparatus comprising: means for providing a first secret and data as inputs to a trained neural network to produce an output by inference; means for sending the output from the trained neural network to a remote server; means for receiving in reply from the server, an encoded label; means for using a second secret to decode the encoded label to obtain a label for the data.

Abstract: In one embodiment, a method of training an autoencoder neural network includes determining autoencoder design parameters for the autoencoder neural network, including an input image size for an input image, a compression ratio for compression of the input image into a latent vector, and a latent vector size for the latent vector. The input image size is determined based on a resolution of training images and a size of target features to be detected. The compression ratio is determined based on entropy of the training images. The latent vector size is determined based on the compression ratio. The method further includes training the autoencoder neural network based on the autoencoder design parameters and the training dataset, and then saving the trained autoencoder neural network on a storage device.

Abstract: Methods for image processing are described. Embodiments of the present disclosure identifies an image generation network that includes an encoder and a decoder; prunes channels of a block of the encoder; prunes channels of a block of the decoder that is connected to the block of the encoder by a skip connection, wherein the channels of the block of the decoder are pruned based on the pruned channels of the block of the encoder; and generates an image using the image generation network based on the pruned channels of the block of the encoder and the pruned channels of the block of the decoder.

Abstract: Apparatus and methods for training a neural network accelerator using quantized precision data formats are disclosed, and in particular for storing activation values from a neural network in a compressed format having lossy or non-uniform mantissas for use during forward and backward propagation training of the neural network. In certain examples of the disclosed technology, a computing system includes processors, memory, and a compressor in communication with the memory. The computing system is configured to perform forward propagation for a layer of a neural network to produced first activation values in a first block floating-point format. In some examples, activation values generated by forward propagation are converted by the compressor to a second block floating-point format having a non-uniform and/or lossy mantissa. The compressed activation values are stored in the memory, where they can be retrieved for use during back propagation.

Abstract: A computerized method for determining an operability status of a ball-throwing machine is disclosed. The method includes operations of obtaining sensor data collected by a sensor during deployment of the ball-throwing machine to perform operations according to a training program, wherein the sensor is disposed on the ball-throwing machine, deploying a trained machine learning model on the sensor data resulting in a classification of a current operability status of the ball-throwing machine or a prediction of a future operability status of the ball-throwing machine, and generating an alert when the classification indicates that the ball-throwing machine is operating abnormally or the prediction indicates the ball-throwing machine is predicted to operate abnormally.

Abstract: A method of training and applying contrastive learning model. The method includes obtaining a sample set and label information for training contrastive learning model, the sample set including a plurality of first samples of a first modality and a plurality of second samples of a second modality, the label information indicating a correlation between samples of the plurality of first samples and samples of the plurality of second samples; determining whether sample mixing is to be performed on the first modality or the second modality; in accordance with a determination that sample mixing is to be performed on the first modality, generating at least one first mixed sample of the first modality by mixing at least one pair of first samples among the plurality of first samples; and training the contrastive learning model at least based on the at least one first mixed sample and first mixed label information.

Abstract: Artificial intelligence is an increasingly important sector of the computer industry. However, artificial intelligence is extremely computationally intensive field such that it can be expensive, time consuming, and energy consuming field. Fortunately, many of the calculations required for artificial intelligence can be performed in parallel such that specialized processors can great increase computational performance for AI applications. Specifically, artificial intelligence generally requires large numbers of matrix operations such that specialized matrix processor circuits can greatly improve performance. To efficiently execute all these matrix operations, the matrix processor circuits must be quickly and efficiently supplied with a stream of data and instructions to process or else the matrix processor circuits end up idle. Thus, this document discloses packet architecture for efficiently creating and supplying neural network processors with work packets to process.

Abstract: Disclosed herein are a multi-target analysis apparatus and method. The multi-target analysis apparatus includes: an input/output interface configured to receive data and output the results of computation of the data; storage configured to store a program for performing a multi-target analysis method; and a controller provided with at least one process, and configured to analyze multiple targets received through the input/output interface by executing the program. The controller is further configured to: collect instruction-target pairs in each of which an instruction and state information for a target are matched with each other so that the target is specified through the instruction, and generate an instruction-target set having a plurality of instruction-target pairs; and train a reinforcement learning-based learning model configured to receive the instruction for the target and the state information for the target and output action information by referring to the instruction-target set.

Abstract: The proposed model is a Variational Autoencoder having a learnable prior that is parametrized with a Tensor Train (VAE-TTLP). The VAE-TTLP can be used to generate new objects, such as molecules, that have specific properties and that can have specific biological activity (when a molecule). The VAE-TTLP can be trained in a way with the Tensor Train so that the provided data may omit one or more properties of the object, and still result in an object with a desired property.

Abstract: A method and electronic device with adversarial data augmentation are provided. The electronic device includes a processor configured to execute instructions; and a memory storing the instructions, where the execution of the instructions by the processor configures the processor to, based on a biased feature within original data, train a biased model to generate biased prediction information using biased training data related to the original data; train a debiased model to generate debiased prediction information using debiased training data, less biased with respect to the biased feature than the biased training data, related to the original data and first adversarial data; and retrain the debiased model using second adversarial data generated based on the biased model and the debiased model.

Abstract: Disclosed is a prediction model generation method for predicting training time and resource consumption required for distributed deep learning training and a prediction method using the prediction model. The prediction model generation method is performed by a computing device including at least one processor and includes constructing a training dataset; and generating a prediction model by training a graph neural network (GNN). The training dataset includes input data and result data, the construction of the training dataset includes converting a distributed deep learning training code (distributed training (DT) code) to a graph; and extracting an adjacency matrix and a feature matrix from the graph.

Abstract: A model training method and a related apparatus to help improve a convergence speed of model training and improve end-to-end communication quality. The method includes: a first communication apparatus sends first data to a second communication apparatus through a channel, where the first data is an output result of the first machine learning model. The second communication apparatus receives second data through a channel, inputs the second data into a second machine learning model to obtain third data; determines a first loss function based on the third data and the first training data; and sends the first loss function to the first communication apparatus through a feedback channel.

Abstract: Systems and methods for training a visual question answer model include training a teacher model by performing image conditional visual question generation on a visual language model (VLM) and a targeted visual question answer dataset using images to generate question and answer pairs. Unlabeled images are pseudolabeled using the teacher model to decode synthetic question and answer pairs for the unlabeled images. The synthetic question and answer pairs for the unlabeled images are merged with real data from the targeted visual question answer dataset to generate a self-augmented training set. A student model is trained using the VLM and the self-augmented training set to return visual answers to text queries.

Abstract: There are provided measures for enabling/realizing efficient model training, including model collection and/or aggregation, for federated learning, including hierarchical federated learning, in a wireless communication system. Such measures exemplarily comprise that a federated-learning training host configured for local model training decides on how to perform the local model training depending on availability of a cluster head and computation and communication costs for a federated-learning training task, and either locally performs the local model training or delegates at least part of a federated-learning training task to the cluster head. Also, such measures exemplarily comprise that a federated-learning training host configured for local model training computes a similarity metric between a locally computed set of local model parameters and each the received sets of local model parameters, and decides on whether to operate as a temporary cluster head for one or more federated-learning training hosts.

Abstract: Systems and methods for automatic forecasting are described. Embodiments of the present disclosure receive a time-series dataset; compute a time-series meta-feature vector based on the time-series dataset; generate a performance score for a forecasting model using a meta-learner machine learning model that takes the time-series meta-feature vector as input; select the forecasting model from a plurality of forecasting models based on the performance score; and generate predicted time-series data based on the time-series dataset using the selected forecasting model.

Abstract: This application relates to the artificial intelligence field, and discloses a neural network search method and a related apparatus. The neural network search method includes: constructing attention heads in transformer layers by sampling a plurality of candidate operators during model search, to construct a plurality of candidate neural networks, and comparing performance of the plurality of candidate neural networks to select a target neural network with higher performance. In this application, a transformer model is constructed with reference to model search, so that a new attention structure with better performance than an original self-attention mechanism can be generated, and effect in a wide range of downstream tasks is significantly improved.

Abstract: Tabular data machine-learning model techniques and systems are described. In one example, common-sense knowledge is infused into training data through use of a knowledge graph to provide external knowledge to supplement a tabular data corpus. In another example, a dual-path architecture is employed to configure an adapter module. In an implementation, the adapter module is added as part of a pre-trained machine-learning model for general purpose tabular models. Specifically, dual-path adapters are trained using the knowledge graphs and semantically augmented trained data. A path-wise attention layer is applied to fuse a cross-modality representation of the two paths for a final result.

Abstract: A non-transitory computer-readable recording medium storing a prediction program that uses knowledge graph embedding, for causing a computer to execute processing including: determining whether or not graph data to be predicted is data that includes a node link that indicates a relationship between nodes not included in training data used for training of the knowledge graph embedding; specifying, in a case where it is determined that the graph data to be predicted is the data that includes the node link not included in the training data, graph data similar to the graph data to be predicted from the training data based on a result of embedding prediction for a label of a node included in the graph data to be predicted; and determining a prediction result for the graph data to be predicted based on the specified similar graph data.

Abstract: Device and computer-implemented method for determining a state of a technical system, in particular of an infrastructure element or of a road user. A first node represents a first object, in particular the technical system, a second node represents a second object, in particular a further infrastructure element or a further road user. An edge between the first node and the second node represents a relationship between the objects. A prediction is determined which characterizes a behavior of one of the objects. The determination is in accordance with information about the objects and in accordance with a representation of a knowledge graph which includes the first node, the second node, and the edge. The state is determined in accordance with the prediction.

Abstract: Disclosed herein are methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for modeling agents in multi-agent systems as reinforcement learning (RL) agents and training control policies that cause the agents to cooperate towards a common goal. A method can include generating, for each of a group of simulated local agents in an agent network in which the simulated local agents share resources, information, or both, experience tuples having a state for the simulated local agent, an action taken by the simulated local agent, and a local result for the action taken, updating each local policy of each simulated local agent according to the respective local result, providing, to each of the simulated local agents, information representing a global state of the agent network, and updating each local policy of each simulated local agent according to the global state of the agent network.

Abstract: The disclosure includes a system and methods of generating a forecast index based on customer digital intent data. The methods include generating a plurality of prediction models using a unique machine learning algorithm for each prediction model, determining an optimal prediction model based on an overall model score assigned to each model of the plurality of prediction models, and generating a forecast index based on the optimal prediction model. The forecast index can be used to adjusts the supply of one or more products based on predicted customer demand.

Abstract: Techniques for the implantation of time-dependent features (e.g., slope features) in existing data analysis models are disclosed. Time-dependent features are applied in machine learning algorithms to provide deeper analysis of temporally spaced data. Temporally spaced data is time-based or time-dependent data where data is populated at different points in time over some period of time. Implementing the time-dependent features enables application of first derivatives that define slopes over time (e.g., performance) windows within the period of time of the data. Application of the first derivatives provides analysis of the trend of the data over time. Additional features and/or higher order derivatives may also be applied to the first derivatives to provide further refinement to analysis of the temporally spaced data.

Abstract: An apparatus for training a driving path prediction model of a vehicle and a method therefor are disclosed. The apparatus includes a sensor that obtains a time series of training images in a real environment and a controller that trains a path prediction model based on dynamic objects in the time series of training images. The controller is configured to train the path prediction model to recognize a dynamic object disappearing from the time series of training images and a dynamic object appearing in the time series of training images.

Abstract: A model generation apparatus includes: an acquisition unit for obtaining learning data including a first index value indicating a condition of a sample patient at a first time, and a second index value indicating a condition of the sample patient at a second time that is after the first time; and a learning unit for learning a prediction mode for predicting a condition of a target patient at the second time on the basis of a condition of the target patient at the first time, wherein the learning unit is configured to learn the prediction model by updating the prediction model on the basis of a third index value indicating a condition of the sample patient at the second time that is predicted by the prediction model on the basis of the first index value, and a change information indicating a change tendency of the condition over time.

Abstract: Disclosed herein are an apparatus and method for sharing an augmented intelligence model of a containerized artificial intelligence (AI) model. The apparatus includes memory in which at least one program is recorded and a processor for executing the program. The program may perform downloading an AI module of a Docker image included in a robot application from container storage by referring to previously stored application configuration information, retrieving model information about an intelligence model used by the AI module from model information storage and downloading the corresponding intelligence model from model storage based on the retrieved model information, and executing the AI module in a container by mounting the intelligence model stored in a local file system as a volume in the container.

Abstract: Methods, systems, and techniques for density ratio estimation of data that includes a covariate variable (W) and a treatment variable (T). The density ratio estimation may be performed using a transformer-based architecture, and the density ratio may be used to control confounding bias in the data. An electronic representation of the data is obtained. At first and second self-attention layers, respectively, covariate variable embeddings based on the data representing the covariate variable and treatment variable embeddings based on the data representing the treatment variable are determined. Cross-attention embeddings based on the covariate and treatment variable embeddings are then determined at a cross-attention layer. At a linear layer and based on the cross-attention embeddings, a density ratio is estimated. The self-attention layers, cross-attention layer, and linear layer are trained using a loss function that determines a loss between an output of the linear layer and the density ratio.

Abstract: The example embodiments are directed to a host system that can convert human-readable rules (e.g., statutes, regulations, laws, etc.) into a semantic model. The host system can then apply the semantic model to a set of circumstances to determine whether and how the rule applies to the circumstances. In one example, the method may include storing a knowledge graph with a semantic model of a rule embodied therein with nodes representing entities within the rule, edges between the nodes representing relationships between the entities, and identifiers of an input data set used by the rule, receiving input data corresponding to the rule, generating a determination from the rule via execution of the semantic model embodied within the knowledge graph on the received input data, and displaying a notification of the determination via a user interface.