Abstract: A method of generating a circuit model used to simulate an integrated circuit may include generating first feature element data and second feature element data by classifying feature data of a target semiconductor device according to measurement conditions, generating first target data and second target data by preprocessing the first feature element data and the second feature element data, respectively, generating a first machine learning model using the first target data and extracting a second machine learning model using the second target data, and generating the circuit model used to simulate the integrated circuit using the first machine learning model and the second machine learning model.

Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: obtaining communication data streams, extracting data relevant to a point of view of a user, and generating a point of view record in a knowledge base that may be utilized by another user communicating with the user.

Abstract: Methods and devices are provided for performing federated learning. A global model is distributed from a server to a plurality of client devices. At each of the plurality of client devices: model inversion is performed on the global model to generate synthetic data; the global model is on an augmented dataset of collected data and the synthetic data to generate a respective client model; and the respective client model is transmitted to the server. At the server: client models are received from the plurality of client devices, where each client model is received from a respective client device of the plurality of client devices; model inversion is performed on each client model to generate a synthetic dataset; the client models are averaged to generate an averaged model; and the averaged model is trained using the synthetic dataset to generate an updated model.

Abstract: Systems, methods, and computer readable media for vehicular search recommendations using machine learning. A computing model may receive a query for a vehicular recommendation. The model may be trained based on historical queries, webpages visited, and attributes of vehicles. The model may generate a decision tree comprising a plurality of paths for processing the query, the plurality of paths comprising a subset of a plurality of available paths for processing the query, each path associated with a plurality of search phases. The model may select, based on the decision tree, a first path of the plurality of paths for processing the query. The model may select, based on the first path, a first search phase of the plurality of search phases as corresponding to the query. The model may then return a search result corresponding to the first search phase as responsive to the query.

Abstract: Methods, apparatus, and processor-readable storage media for automated subscription management for remote infrastructure are provided herein.

Abstract: Methods and systems for managing artificial intelligence (AI) models are disclosed. To manage AI models, poisoned training data introduced into an instance of the AI models may be identified and the impact of the poisoned training data on the AI models may be efficiently mitigated. To do so, a first poisoned AI model instance may be obtained. Rather than re-training an un-poisoned AI model instance to remove the impact of poisoned training data, the first poisoned AI model instance may be selectively un-trained whenever poisoned training data is found in the training dataset. Subsequently, weights of the first poisoned AI model instance may be adjusted to account for future training data. As poisoned training data may occur infrequently, selectively un-training the AI model may conserve computing resources and minimize AI model downtime when compared to a full or partial re-training process of an un-poisoned AI model instance.

Abstract: The technology disclosed proposes using a combination of computationally cheap, less-accurate bag of words (BoW) model and computationally expensive, more-accurate long short-term memory (LSTM) model to perform natural processing tasks such as sentiment analysis. The use of cheap, less-accurate BoW model is referred to herein as “skimming”. The use of expensive, more-accurate LSTM model is referred to herein as “reading”. The technology disclosed presents a probability-based guider (PBG). PBG combines the use of BoW model and the LSTM model. PBG uses a probability thresholding strategy to determine, based on the results of the BoW model, whether to invoke the LSTM model for reliably classifying a sentence as positive or negative. The technology disclosed also presents a deep neural network-based decision network (DDN) that is trained to learn the relationship between the BoW model and the LSTM model and to invoke only one of the two models.

Abstract: A processor may initiate metadata discovery. The processor may identify an asset category and asset count. The processor may determine whether one or more assets can be imported. The processor may determine whether an import was successful. The processor may terminate the import.

Abstract: A computing system can include one or more machine-learned models configured to receive context data that describes one or more entities to be named. In response to receipt of the context data, the machine-learned model(s) can generate output data that describes one or more names for the entity or entities described by the context data. The computing system can be configured to perform operations including inputting the context data into the machine-learned model(s). The operations can include receiving, as an output of the machine-learned model(s), the output data that describes the name(s) for the entity or entities described by the context data. The operations can include storing at least one name described by the output data.

Abstract: The present disclosure generally relates to systems and methods for searching media content. In some implementation examples, a search system receives an input query, generates a query embedding of the input query, and generates a bias mitigation transformation associated with a sensitive attribute. Based on the query embedding and the bias mitigation transformation, the search system generates a transformed query embedding that suppresses at least a portion of the query embedding related to the sensitive attribute. Using the transformed query embedding, the search system executes a similarity search in a media embedding model to identify one or more media embeddings that are similar to the transformed query embedding and transmits the one or more media embeddings.

Abstract: A system and method for generating a decision tree having a plurality of nodes, arranged hierarchically as parent nodes and child nodes, comprising: generating a node including: receiving i) training data including data instances, each data instance having a plurality of attributes and a corresponding label, ii) instance weightings, iii) a valid domain for each attribute generated, and iv) an accumulated weighted sum of predictions for a branch of the decision tree; and associating one of a plurality of binary prediction of an attribute with each node including selecting the one of the plurality of binary predictions having a least amount of error; in accordance with a determination that the node includes child nodes, repeat the generating the node step for the child nodes; and in accordance with a determination that the node is a terminal node, associating the terminal node with an outcome classifier; and displaying the decision tree including the plurality of nodes arranged hierarchically.

Abstract: According to an example aspect of the present invention, there is provided an input device and corresponding method which digitizes and transforms minute hand movements and gestures into user interface commands without interfering with the normal use of one's hands. The device and method may, for example, determine user actions based on detected user action characteristics from a plurality of sensors, where the sensors are preferably of different types.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support cooperative training of machine learning (ML) models that preserves privacy in untrusted environments. For example, a server (or cloud-based computing device(s)) may be configured to “split” an initial ML model into various partial ML models, some of which are provided to client devices for training based on client-specific data. Output data generated during the training at the client devices may be provided to the server for use in training corresponding server-side partial ML models. After training of the partial ML models is complete, the server may aggregate the trained partial ML models to construct an aggregate ML model for deployment to the client devices. Because the client data is not shared with other entities, privacy is maintained, and the splitting of the ML models enables offloading of computing resource-intensive training from client devices to the server.

Abstract: A method, computer system, and a computer program product for automatically generated question suggestions is provided. The present invention may include generating an intent-entity sequence representative of a current chat conversation with a user. The present invention may also include predicting, using a trained intent-entity prediction model, a next intent-entity sequence based on the generated intent-entity sequence representative of the current chat conversation. The present invention may further include converting the predicted next intent-entity sequence into at least one likely question suggestion to return back to the user.

Abstract: Disclosed herein are an apparatus and method for developing a neural network application. The apparatus includes one or more processors and executable memory for storing at least one program executed by the one or more processors. The at least one program receives a target specification and an application specification including user requirements, searches for a neural network model corresponding to the target specification and the application specification in a database, builds an inference engine for performing a neural network operation used by the neural network model, and generates a target image for executing the neural network model to be suitable for a target device using the inference engine.

Abstract: An approach is provided for augmenting text of a small class for text classification. An imbalanced dataset is received. A small class is identified. The small class includes initial text records in the imbalanced dataset. A balanced dataset is generated from the imbalanced dataset by augmenting the initial text records by using weighted word scores indicating respective measures of importance of words in classes in the imbalanced dataset. The balanced dataset is sent to a supervised machine learning model. The supervised machine learning model is trained on the balanced dataset. Using the supervised machine learning model which employs the augmented initial text records, a text classification of a new dataset is performed. The domain of the new dataset matches the domain of the imbalanced dataset.

Abstract: Various embodiments may be generally directed to the use of an adversarial learning framework for persona-based dialogue modeling. In some embodiments, automated multi-turn dialogue response generation may be performed using a persona-based hierarchical recurrent encoder-decoder-based generative adversarial network (phredGAN). Such a phredGAN may feature a persona-based hierarchical recurrent encoder-decoder (PHRED) generator and a conditional discriminator. In some embodiments, the conditional discriminator may include an adversarial discriminator that is provided with attribute representations as inputs. In some other embodiments, the conditional discriminator may include an attribute discriminator, and attribute representations may be handled as targets of the attribute discriminator. The embodiments are not limited in this context.

Abstract: Embodiments of a method and system for determining delivery estimates for shipping a parcel include: retrieving historical delivery data from a plurality of shipping carriers; generating cross-carrier delivery features based on normalizing the historical delivery data; generating a cross-carrier delivery prediction model based on the cross-carrier delivery features; retrieving parcel data for the parcel based on a tracking number S140; generating parcel features based on normalizing the parcel data S150; determining a delivery estimate for the parcel based on processing the parcel features with the cross-carrier delivery prediction model S160; and responding to the delivery estimate S170.

Abstract: Systems and methods for generating a content item based on a difference between a user confidence score and a confidence score are disclosed. For example, a system generates for output a first content item. While the first content item is being outputted, the system receives user data via sensors of a device. The system determines a user confidence score based on the user data and metadata of the first content item. The user confidence score indicates a user's perceived probability of an event occurring in the future. The system calculates a prediction score which estimates the likelihood of the event occurring in the future. In response to determining that the difference between the user confidence score and the prediction score exceeds a threshold, the system selects a second content item related to the event and generates for output a recommendation comprising an identifier of the second content item.

Abstract: There is disclosed a system and method of detecting security threats for an enterprise, including: filtering a first set of endpoint metadata records to identify a subset of metadata records, wherein filtering includes identifying endpoint security metadata records that are uncommon in context of the enterprise; and designating the subset of metadata records as indicating a potential security threat including designating the subset of metadata records for human analysis.

Abstract: This disclosure provides systems, methods and apparatuses for classifying traffic flow using a plurality of learning machines arranged in multiple hierarchical levels. A first learning machine may classify a first portion of the input stream as malicious based on a match with first classification rules, and a second learning machine may classify at least part of the first portion of the input stream as malicious based on a match with second classification rules. The at least part of the first portion of the input stream may be classified as malicious based on the matches in the first and second learning machines.

Abstract: A first dataset that includes an indication of a plurality of network events associated with a time-period is received. For each time sub-period from a plurality of time sub-periods that together span the time-period and to generate a second dataset, a value for each network event from the plurality of network events that occur within that time sub-period is summed. A discrete Fourier transform is performed based on the second dataset to generate a third dataset that includes an indication of a plurality of frequency ranges and a plurality of magnitude values for the plurality of frequency ranges. Each frequency from the plurality of frequencies ranges is associated with a magnitude value from the plurality of magnitude values. A set of candidate frequencies from the plurality of frequencies determined to potentially cause periodic behavior is identified based on the plurality of frequency ranges and the plurality of magnitude values.

Abstract: Systems and methods include processors for receiving training data for a user activity; receiving bias criteria; determining a set of model parameters for a machine learning model including: (1) applying the machine learning model to the training data; (2) generating model prediction errors; (3) generating a data selection vector to identify non-outlier target variables based on the model prediction errors; (4) utilizing the data selection vector to generate a non-outlier data set; (5) determining updated model parameters based on the non-outlier data set; and (6) repeating steps (1)-(5) until a censoring performance termination criterion is satisfied; training classifier model parameters for an outlier classifier machine learning model; applying the outlier classifier machine learning model to activity-related data to determine non-outlier activity-related data; and applying the machine learning model to the non-outlier activity-related data to predict future activity-related attributes for the user activity

Abstract: A method, apparatus, and computer-readable medium for efficiently optimizing a phenotype with a combination of a generative and a predictive model, training a phenotype prediction model based on experiential genotype vectors, training a genotype generation model based on sample genotype vectors, generating new genotype vectors, applying the phenotype prediction model to the new genotype vectors to generate scores, determining result genotypes based on a ranking of the available genotypes according to the scores, and generating a result based on the result genotypes, the result indicating one or more genetic constructs for testing.

Abstract: According to one or more embodiments of the disclosure, a networking device receives a policy for an endpoint in a network. The policy specifies one or more component tags and one or more activity tags that were assigned to the endpoint based on deep packet inspection of traffic associated with the endpoint. The networking device identifies a set of tags for a particular traffic flow in the network associated with the endpoint. The set of tags comprises one or more component tags or activity tags associated with the particular traffic flow. The networking device makes a determination that the particular traffic flow violates the policy based on the set of tags comprising a tag that is not in the policy. The networking device initiates, based on the determination that the particular traffic flow violates the policy, a corrective measure with respect to the particular traffic flow.

Abstract: An embodiment for improved linking of tabular columns to column types in an ontology unseen during training. The embodiment may for a target table, encode a target tabular query column, table headers, and target types independently to generate permutation invariant representations of tabular data associated with the target table. The embodiment may, for each of the target types, extract and further encode auxiliary information. The embodiment may process the encoded tabular data to obtain a first vector and a second vector. The embodiment may concatenate the first vector and the second vector to generate a final query vector. The embodiment may process the encoded target types through a third transformer to obtain a third vector. The embodiment may calculate a score to model interactions between the target tabular query column of the target table and the target types.

Abstract: An apparatus for determining system model comparison is disclosed. The apparatus includes a processor and a memory communicatively linked to the processor. The memory instructs the processor to receive a first plurality of system data, wherein the first plurality of system data represents a first state of a system, and a second plurality of system data representing a second of the system. The memory instructs the processor to generate a first and second model of the system using the system data. The memory instructs the processor to output a first model output using the first model of the system and the second plurality of system data. The memory instructs the processor to modify the second plurality of system data using perturbation function. The memory instructs the processor to output a second model output using the second model of the system and the modified second plurality of system data.

Abstract: A computerized method is disclosed for automated handling of data ingestion anomalies. The method features operations of detecting a data ingestion anomaly and determining a cause for the data ingestion anomaly. The causal determination may be conducted by at least (i) determining features of an anomalous data ingestion volume, (ii) training a second data model, after a first data model being used to detect the data ingestion anomaly, with data sets consistent with the determined features, (iii) applying the second data model to predict whether a data ingestion sub-volume is anomalous, (iv) obtaining system state information during ingestion of the anomalous data ingestion sub-volume, and (v) determining the cause of the anomalous data ingestion volume based on the system state information.

Abstract: Implementations disclosed herein relate to utilizing at least one existing manually engineered policy, for a robotic task, in training an RL policy model that can be used to at least selectively replace a portion of the engineered policy. The RL policy model can be trained for replacing a portion of a robotic task and can be trained based on data from episodes of attempting performance of the robotic task, including episodes in which the portion is performed based on the engineered policy and/or other portion(s) are performed based on the engineered policy. Once trained, the RL policy model can be used, at least selectively and in lieu of utilization of the engineered policy, to perform the portion of robotic task, while other portion(s) of the robotic task are performed utilizing the engineered policy and/or other similarly trained (but distinct) RL policy model(s).

Abstract: Embodiments of various systems, methods, and devices are disclosed for generating artificial intelligence or machine learning models for predicting denials of medical claims, predicting approvals of resubmitted medical claims, as well as automatic workflow clustering processes for automatically assigning medical claims to workflow queues using predictive segmentation and smart resource allocation.

Abstract: A method for controlling a robot. The method includes providing demonstrations for performing each of a plurality of skills; training from the demonstrations, a robot trajectory model for each skill, each trajectory model is a hidden semi-Markov model having one or more initial states and one or more final states; training, from the demonstrations, a precondition model for each skill comprising, for each initial state, a probability distribution of robot configurations before executing the skill, and a final condition model for each skill comprising, for each final state, a probability distribution of robot configurations after executing the skill; receiving a description of a task, the task includes performing the skills of the plurality of skills in sequence and/or branches; generating a composed robot trajectory model; and controlling the robot according to the composed robot trajectory model to execute the task.

Abstract: The present teaching relates to method, system, medium, and implementations for machine learning. Machine learning is performed based on training data via a dual loop learning process that includes a first loop for data decoding learning and a second loop for label decoding learning. In the first loop, first parameters associated with decoding are updated to generate updated first parameters based on a first label, estimated via the decoding using the first parameters, and a second label, predicted via the label decoding using second parameters. In the second loop, the second parameters associated with the label decoding are updated to generate updated second parameters based on a third label, obtained via the decoding using the updated first parameters, and a ground truth label.

Abstract: The present application discloses a processor video memory optimization method and apparatus for deep learning training tasks, and relates to the technical field of artificial intelligence. In the method, by determining an optimal path for transferring a computing result, the computing result of a first computing unit is transferred to a second computing unit by using the optimal path. Thus, occupying the video memory is avoided, and meanwhile, a problem of low utilization rate of the computing unit of a GPU caused by video memory swaps is avoided, so that training speed of most tasks is hardly reduced.

Abstract: Systems, methods, tangible non-transitory computer-readable media, and devices for detecting objects are provided. For example, the disclosed technology can obtain a representation of sensor data associated with an environment surrounding a vehicle. Further, the sensor data can include sensor data points. A point classification and point property estimation can be determined for each of the sensor data points and a portion of the sensor data points can be clustered into an object instance based on the point classification and point property estimation for each of the sensor data points. A collection of point classifications and point property estimations can be determined for the portion of the sensor data points clustered into the object instance. Furthermore, object instance property estimations for the object instance can be determined based on the collection of point classifications and point property estimations for the portion of the sensor data points clustered into the object instance.

Abstract: Systems and methods for low bias negative sampling of classes according to the sampled softmax method are described herein. The systems and methods can include training a machine-learned model for classifying inputs into one or more classes of a plurality of classes, each of the plurality of classes having an associated class embedding in a plurality of class embeddings. The systems and methods can include selecting, by the one or more computing devices, one or more negative classes from the plurality of classes based at least in part on a probability distribution approximating a softmax distribution, wherein the probability distribution is determined based at least in part on a Random Fourier Features map.

Abstract: A method, system and product including obtaining offline user information at an end device, wherein the offline user information is obtained from offline sensors of the end device; based on the offline user information, generating a user profile indicating that a user of the end device matches at least one micro-segment, wherein the at least one micro-segment comprises at least one detailed population category; based on the at least one micro-segment, selecting a campaign from a set of one or more campaigns retained at a server, wherein the campaign comprises one or more rules for displaying at least one content item; monitoring the offline sensors of the end device to identify real time user activities; and upon identifying, based on the real time user activities, that a rule of the one or more rules for displaying a content item is complied with, displaying the content item in the end device.

Abstract: The present disclosure provides systems and methods for training probabilistic object motion prediction models using non-differentiable representations of prior knowledge. As one example, object motion prediction models can be used by autonomous vehicles to probabilistically predict the future location(s) of observed objects (e.g., other vehicles, bicyclists, pedestrians, etc.). For example, such models can output a probability distribution that provides a distribution of probabilities for the future location(s) of each object at one or more future times. Aspects of the present disclosure enable these models to be trained using non-differentiable prior knowledge about motion of objects within the autonomous vehicle's environment such as, for example, prior knowledge about lane or road geometry or topology and/or traffic information such as current traffic control states (e.g., traffic light status).

Abstract: Systems and methods of generating rating indicators for a portfolio of financial assets are described. A machine learning model is trained using a training data set that includes one or more qualitative features and one or more first quantitative features to generate an output predictor of the performance of the portfolio. The qualitative features are converted into quantitative features before being used as input to the machine learning model. Input features are generated for a new portfolio of financial assets whose rating indicator is to be generated, and fed into the trained machine learning model to generate a new output predictor for the new portfolio of financial assets. The rating indicator for the new portfolio of financial assets is determined based at least on the generated new output predictor.

Abstract: Described herein are systems, methods, and instrumentalities associated with image segmentation such as tubular structure segmentation. An artificial neural network is trained to segment tubular structures of interest in a medical scan image based on annotated images of a different type of tubular structures that may have a different contrast and/or appearance from the tubular structures of interest. The training may be conducted in multiple stages during which a segmentation model learned from the annotated images during a first stage may be modified to fit the tubular structures of interest in a second stage. In examples, the tubular structures of interest may include coronary arteries, catheters, guide wires, etc., and the annotated images used for training the artificial neural network may include blood vessels such as retina blood vessels.

Abstract: A system for privacy-preserving distributed training of a global model on distributed datasets has a plurality of data providers being communicatively coupled. Each data provider has a local model and a local training dataset for training the local model using an iterative training algorithm. Further it has a portion of a cryptographic distributed secret key and a corresponding collective cryptographic public key of a multiparty fully homomorphic encryption scheme. All models are encrypted with the collective public key. Each data provider trains its local model using the respective local training dataset, and combines the local model with the current global model into a current local model. A data provider homomorphically combines current local models into a combined model, and updates the current global model based on the combined model. The updated global model is provided to at least a subset of the other data providers.

Abstract: Embodiments of the present invention set forth a technique for predicting fraud based on multiple inputs including user behavior biometric data along with one or more other parameters associated with the user. The technique includes receiving cursor movement data generated via a client device. The technique further includes generating a image based on the cursor movement data. The technique further includes receiving client parameters generated via the client device. The technique further includes analyzing the image and the client parameters based on a model to generate a prediction result, where the model is generated based on second cursor movement data and a second set of client parameters associated with a first group of one or more users. The technique further includes determining, based on the prediction result, that a user of the client device is not a member of the first group.

Abstract: A suite of fluidless predictive machine learning models includes a fluidless mortality module, smoking propensity model, and prescription fills model. The fluidless machine learning models are trained against a corpus of historical underwriting applications of a sponsoring enterprise, including clinical data of historical applicants. A data appended procedure supplements historical applications data with public records and credit risks. Various features of this data are engineered for improved predictive characteristics. Fluidless models are trained by application of a random forest ensemble including survival, regression and classification models. The trained models produce high-resolution, individual mortality scores. A fluidless underwriting protocol runs these predictive models to assess mortality risk and other risk attributes of a fluidless application that excludes clinical data to determine whether to present an accelerated underwriting offer.

Abstract: There is provided a method and system for training an embedding model to perform relation predictions in a knowledge hypergraph to output a trained embedding model. A training dataset comprising tuples representing relations between entities in the knowledge hypergraph are received. The embedding model is trained to perform relation predictions for each given tuple from a subset of tuples in the training dataset by generating a respective entity vector for each entity and a respective relation matrix representing relations between the entities. The entity vectors and relation matrix are split into a plurality of windows, and interaction values between elements in each window are calculated. A relation score indicative of the relation in the given tuple being true is calculated. Parameters of the embedding model are updated based on the relation scores for the subset of tuples. The trained embedding model is then output.

Abstract: There is provided a method of monitoring performance of a machine learning model externally to the machine learning model, comprising: monitoring data elements being fed into a machine learning model trained on a training dataset of historical training data elements, wherein the data elements are each associated with a respective time after the time associated with the training dataset, analyzing the data elements for identifying shift(s) between at least two subsets of the data elements, computing according to the shift(s), measurement(s) denoting an expected effect on output of the model, and detecting a misclassification event by the model when the measurement(s) exceeds a threshold of the model, wherein the monitoring, the analyzing, the computing, and the detecting are performed externally to the model, without accessing at least one of: data stored within the machine learning model, an implementation of the model, and data structures of the model.

Abstract: Disclosed are techniques for using machine learning models to more reliably predict likelihoods of application failure. A model is trained to identify and display events that may cause high severity application failures. Logistic regression may be used to fit the model such that application features are mapped to a high severity event flag. Significant features of applications that relate to the high severity flag may be selected using stepwise regression. The identified applications may be displayed on a graphical user interface for review and reprioritization. Information may be ranked and displayed according to multiple different ranking criteria, such as one ranking generated by a first model, and another determined by one or more users. The multiple ranking criteria may be used to inform steps taken, and/or to retrain or tune the parameters of the model for subsequent predictions or classifications.

Abstract: A system described herein may train an explanation model based on a set of images and a set of explanation labels. The system may receive input data, and may provide the input data to the explanation model and a second model. The second model may provide a set of output labels, which may include performing unknown or “black box” processing on the input data. The explanation model may generate one or more images based on the input data, compare the images to the set of images based on which the explanation model was trained, and accordingly identify one or more explanation labels with bounding boxes associated with the generated one or more images. The system may output, in response to the input data, the set of output labels provided by the second model as well as the identified explanation labels.

Abstract: A system and method for connected vehicle risk detection are presented. The includes computing risk scores for a plurality of behaviors detected with respect to a connected vehicle, wherein each of the detected plurality of behaviors is associated with the connected vehicle based on a contextual vehicle state of the connected vehicle; aggregating the computed risk scores to determine an aggregated risk score; determining a risk level for the connected vehicle based on the aggregated risk score; and causing execution of at least one mitigation action based on the determined risk level.

Abstract: Disclosed is an operation method of a storage device supporting a multi-stream, which includes receiving an input/output request from an external host, generating a plurality of stream identifier candidates by performing machine learning on the input/output request based on a plurality of machine learning models that are based on different machine learning algorithms, generating a model ratio based on a characteristic of the input/output request, applying the model ratio to the plurality of stream identifier candidates to allocate a final stream identifier for the input/output request, and storing write data corresponding to the input/output request in a nonvolatile memory device of the storage device based on the final stream identifier.

Abstract: A visualization tool for machine learning models obtains metadata from a first training node at which a multi-layer machine learning model is being trained. The metadata includes a parameter of an internal layer of the model. The tool determines a plurality of metrics from the metadata, including respective loss function values corresponding to several training iterations of the model. The tool indicates the loss function values and the internal layer parameter values via a graphical interface.

Abstract: Systems and methods of generating an interface including elements related to a next best state prediction are disclosed. A request for an interface including a user identifier is received. A next state prediction engine receives a sequence unit set including at least one sequence unit associated with the user identifier and a set of features associated with the at least one sequence unit and generates at least one next state prediction using a trained sequential prediction model. The trained sequential prediction model is configured to receive the sequence unit set and the set of features for the at least one sequence unit and output at least one predicted next state for the sequence unit set. An interface generation engine generates an interface including at least one element related to the at least one predicted next state and transmits the interface to a user device associated with the user identifier.