Abstract: The present invention relates to systems and methods for automated interpretable and generalizable biological morphological profiling. The method for identifying single-cell morphological profiling based on deep learning includes collecting and pre-processing at least one single-cell image data; training Variational Autoencoder (VAE) by defining an arbitrary dimension size of a latent space; distilling a learnt latent space from the VAE to Generative Adversarial Network (GAN) and training a generator-discriminator combination within the GAN; generating a realistic image aligned with the learnt latent space; and interpreting data by incorporating statistical variance analysis and hierarchical clustering.

Abstract: Systems and methods of online active learning with corrective and iterative feedback for streaming data classification are provided. In one exemplary embodiment, a method includes initializing trainable parameters of a detection model based on verified samples including a representation of a target event and receiving unverified samples from a client network node device. The method further includes iteratively ranking the unverified samples based on an informativeness criterion, selecting a query subset based on an adaptive query budget, obtaining HITL verification responses for the query subset, and updating the detection model based on the HITL verification responses.

Abstract: The disclosed subject matter provides systems and methods for processing light. An example system can include one or more substrates, and a plurality of meta-units, which are patterned on each of the substrates and configured to modify a phase, an amplitude, or a polarization of the light with a subwavelength resolution. The system can be in a form of a diffractive neural network and be configured to perform target recognition.

Abstract: Disclosed is a technology of independently connecting a plurality of neuromorphic elements three-dimensionally laminated on a Complementary Metal-Oxide Semiconductor (CMOS) wafer through an interconnection layer and selectively driving and testing the plural neuromorphic elements with pulses generated and transmitted from the CMOS wafer.

Abstract: A method of training a machine learning model for evaluating flavor profiles of an alcoholic beverage includes a spectral sensor collecting optical properties of the beverage or flavoring agent suspended in the beverage within a container. Temperature and pressure sensors collect environment data of the container during a flavoring session. A control apparatus displays a user interface requesting a taste tester to select labels for the session. A control apparatus memory stores selected labels responsive to the taste tester's selections. A feature extraction module extracts features from the optical properties and environment data based on predetermined criteria. A machine learning engine trains a machine learning model to continuously learn flavor characteristics under a plurality of flavoring sessions using the extracted features and selected labels.

Abstract: Some embodiments of this specification disclose graphical model training methods, apparatuses, and devices. In an embodiment, a method performed by a terminal device includes: acquiring node information of a first graph to be constructed, node information of a second graph and node connection information of the second graph, acquiring a latent vector of a first node that has training label information in the first graph, acquiring latent vectors of the second node and the node in the second graph. performing, based on the first sample data and the second sample data, model training on a graphical model sent by a server, acquiring gradient information corresponding to the trained graphical model, and sending the gradient information to the server for the server to update model parameters in the graphical model based on gradient information provided by different terminal devices to obtain an updated graphical model.

Abstract: A communication device (100; 100a) has a processing device (110), a user interface (120), and a short-range wireless communication transceiver (130). The short-range wireless communication transceiver (130) is configured for receiving a short-range wireless communication signal(S) from another communication device (100b). The processing device (110) is configured for reading a communication identifier (CI) of the received short-range wireless communication signal (S), determining-based on the communication identifier (CI)—a particular audio/visual/audiovisual media content (140a) among a plurality of possible media contents (140), controlling the user interface (120) to present the particular media content (140a) to a user (1a) of the communication device (100; 100a), and controlling the user interface (120) to receive an input (INP) from the user (1a).

Abstract: Provided are a method for evaluating performance and a system thereof. The method according to some embodiments may include obtaining a first model trained using a labeled dataset, obtaining a second model built by performing unsupervised domain adaptation on the first model, generating pseudo labels for an evaluation dataset using the second model, wherein the evaluation dataset is an unlabeled dataset, and evaluating performance of the first model using the pseudo labels.

Abstract: A system and method of predicting behavior of at least one electric machine is provided, wherein the method includes: generating a simulated-dataset including simulated design results, (e.g., individually), for electromagnetic properties, structural properties, and acoustic properties of the electric machine, wherein the simulated-dataset is generated by simulating at least one operating condition of the electric machine on parametric models generated from design parameters of the electric machine; training artificial neural network models using the design parameters and the simulated design results output from the parametric models in response to at least one operating condition of the electric machine; and predicting behavior of the electric machine by orchestrating execution of the artificial neural network models for custom design parameters.

Abstract: Aspects of the disclosure are directed to implementing a projection-based stochastic gradient descent technique that maintains label differential privacy when training one or more machine learning models. The technique includes denoising gradients by exploiting projections when training the machine learning models to improve performance of the trained machine learning models while maintaining label differential privacy. For instance, the projection-based stochastic gradient descent technique can improve performance of machine learning models in higher-privacy regimes, such as digital content management.

Abstract: The disclosure relates to systems and methods of detecting anomalies using a plurality of machine learning models. Each of the machine learning models may be trained to detect a respective behavior of historical data values for a given metric. Thus, a system may perform anomaly detection based on different behaviors of the same metric of data, reducing instances of false positive anomaly detection while also reducing instances of false negative reporting. The plurality of machine learning models may be trained to detect anomalies across multiple different types of metrics as well, providing robust multi-metric anomaly detection across a range of behaviors of historical data values. The system may implement a pluggable architecture for the plurality of machine learning models in which models may be added or removed from pluggable architecture. In this way, the system may detect anomalies using a configurable set of machine learning models.

Abstract: Flaws in a codebase for an organization are triaged with a naïve Bayes classifier that determines likelihoods of triage decisions corresponding to actions (e.g., remediating via code change, deferring due to network mitigation, labeling as false positive) given the context of the flaw, application, and organization. The naïve Bayes classifier is trained on the triage outcomes of previously detected flaw instances in the codebase and provides interpretable results including feature-level likelihood scores of each triage approach. In addition to recommending the highest likelihood triage outcome provided by the naïve Bayes model, a flaw similarity model identifies previously triaged flaw instances from the organization to recommend more granular triage instructions that have been documented alongside the previous flaw instances.

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: A system including a processor and a memory storing program instructions and a machine learning module. The machine learning module is configured for outputting one or more suggested data element change requests in response to receiving an initial data element change request. Execution of the program instructions causes the processor to receive the initial data element change request, receive the one or more suggested data element change requests in response to inputting the initial data element change request into the machine learning module, receive one or more subsequent data element change requests, detect if the one or more subsequent data element change requests contain the one or more suggested data element change requests, and provide an alert signal if the one or more suggested data element change requests are not detected within the one or more subsequent data element change requests.

Abstract: A distributed reinforcement learning system includes one or more actor devices configured to acquire experience data, the experience data being used for reinforcement learning and corresponding to an action determined based on a model to be trained, a plurality of replay buffers configured to store the experience data acquired from the one or more actor devices, and one or more learner devices configured to train the model in the reinforcement learning, the reinforcement learning using the experience data stored in the plurality of replay buffers. The plurality of replay buffers are distributed and arranged in a plurality of nodes.

Abstract: Systems, methods, and devices for generating constrained outputs for large language models are disclosed. The system receives an input query relating to a request from a user via a user interface associated with an entity. The system retrieves knowledge elements from a proprietary knowledge corpus, including frameworks, methodologies, or constraints specific to the entity. For each knowledge element, the system inputs the query and knowledge element into a large language model to generate a preliminary output. The system applies user-specific rules to generate a constrained output by parsing the preliminary output, evaluating response components against user-specific constraints as logical predicates, and modifying components that violate rules. The system generates citations linking outputs to source identifiers, aggregates the constrained outputs and citations to generate a composite response, and transmits it to the user interface.

Abstract: Embodiments relate to providing explainable classifications with abstention using client agnostic machine learning models. A technique includes classifying, by a processor, a record with a label using a machine learning model, the machine learning model abstaining from classifying a given record in response to the given record being outside of a scope of an information technology (IT) domain. The processor generates an explanation of a decision by the machine learning model to classify the record with the label and displays the explanation in a human readable form.

Abstract: A citation network graph representation learning system and method based on multi-view contrastive learning is provided. The citation network graph representation learning system involved in the present application comprises: a sample construction module, which is configured to construct a corresponding negative sample based on an original graph; a graph enhancement module, which is configured to obtain a positive sample graph and a negative sample graph; a fusion module, which is configured to obtain a consensus representation of the positive sample graph and the negative sample graph by means of a cross view concentration fusion layer; a mutual information estimation module, which is configured to compare learning representations of positive sample pairs and negative sample pairs by means of a discriminator; and a hard sample mining module, which is configured to represent the consistency between the negative sample pairs according to a pre-calculated affinity vector, and select and reserve nodes.

Abstract: This disclosure relates to methods and systems for high resolution and scalable crop yield forecasting by first developing a first crop yield forecasting model to generate coarse resolution yield maps and further dynamically selecting a set of pixels from the coarse resolution yield maps. The coarse resolution yield maps, satellite, weather and soil related data are fed as input to a second crop yield forecasting to generate high resolution crop yield forecasting maps. Further, domain knowledge about crop growth stages, economically important crop growth stages and weather based triggers are identified to quantify extent of change in crop yield. This helps in crop yield forecasting during real time adverse weather conditions. Finally, an adjusted crop yield model is obtained after adjusting losses incurred due to the real time adverse weather conditions to obtain accurate high resolution crop yield forecasting maps. The method of present disclosure is inexpensive, light-weight, and scalable.

Abstract: Collaborative training with compressed transmissions is performed by partitioning a plurality of layers of a neural network model into a device partition and a server partition, combining a plurality of encoding layers of an auto-encoder neural network with the device partition, wherein a largest encoding layer among the plurality of encoding layers is adjacent a layer of the device partition bordering the server partition, combining a plurality of decoding layers of the auto-encoder neural network with the server partition, wherein a largest decoding layer among the plurality of decoding layers is adjacent a layer of the server partition bordering the device partition, transmitting, to a computation device, the device partition combined with the plurality of encoding layers, and training, collaboratively with the computation device through a network, the neural network model.