Abstract: Systems and methods for providing insights about a machine learning model are provided. The method includes, using training data to train the machine learning model to learn patterns to determine whether data associated with an event provides an indication that the event belongs to a certain class from among a plurality of classes, evaluating one or more features of the machine learning model to produce a data set pairing observed scores S and a set of predictive input variables Vi, and constructing at least one data-driven estimator based on an explanatory statistic, the estimator being represented in a computationally efficient form and packaged with the machine learning model and utilized to provide a definition of explainability for a score generated by the machine learning model.

Abstract: A system and method that guides a user to record initial and final walkthrough videos, along with audio, of their rental property is disclosed. The embodiments also provide a system that can scan a lease for the rental property, as well as audio from a user's narration of the video, and extract and tag particular features of the video for efficient querying at a later time. The system can interpret portions of the lease that are relevant to particular items or components in the apartment (for example, major appliances) and guide the user to perform a comprehensive inspection of the items mentioned in the lease.

Abstract: A diagnostic system for model governance is presented. The diagnostic system includes an auto-encoder to monitor model suitability for both supervised and unsupervised models. When applied to unsupervised models, the diagnostic system can provide a reliable indication on model degradation and recommendation on model rebuild. When applied to supervised models, the diagnostic system can determine the most appropriate model for the client based on a reconstruction error of a trained auto-encoder for each associated model. An auto-encoder can determine outliers among subpopulations of consumers, as well as support model go-live inspections.

Abstract: A stream of continuous data processing commands or workloads may be tracked for completeness. A stream of signal is received that includes a plurality of commands to be executed. At least one timeslice is generated in the received stream which corresponds to a period of time and includes the plurality of commands received in the period of time. The plurality of commands are categorized into at least one page and a unique identifier is assigned to the at least one page. The at least one page, including the unique identifier, is transmitted to a command executing agent and in response to receiving an indication that the at least one page is completed, each of the plurality of commands included in the at least one page is marked as completed.

Abstract: An electric pump system and method of operating the same involves pumping a fluid through a fluid passageway defined in a mechanical pump from a pump inlet to a hollow shaft of a motor, through the hollow shaft to an internal motor cavity defined by a housing of the motor, and through another fluid passageway defined in the motor housing and mechanical pump that leads to a pump outlet. The system and method further involve pumping the fluid through another fluid passageway defined in the mechanical pump from yet another pump inlet to the pump outlet. The temperature of fluid exiting the hollow shaft can be assessed and used by an electronic control unit (ECU) of the electric pump system to control the same. The electric pump system can be part of a cooling and lubrication system for an electric vehicle transmission, gearbox, differential or transfer case, for example.

Abstract: Systems, methods and computer program products for improving security of artificial intelligence systems. The system comprising processors for monitoring one or more transactions received by a machine learning decision model to determine a first score associated with a first transaction. The first transaction may be identified as likely adversarial, in response to the first score being lower than a certain score threshold and the first transaction having a low occurrence likelihood. A second score may be generated in association with the first transaction based on one or more adversarial latent features associated with the first transaction. At least one adversarial latent feature may be detected as being exploited by the first transaction, in response to determining that the second score falls above the certain score threshold. Accordingly, an abnormal volume of activations of adversarial latent features spanning across a plurality of transactions scored may be detected and blocked.

Abstract: The technology disclosed provides systems and methods related to preventing exfiltration of training data by feature reconstruction attacks on model instances trained on the training data during a training job. The system comprises a privacy interface that presents a plurality of modulators for a plurality of training parameters. The modulators are configured to respond to selection commands via the privacy interface to trigger procedural calls. The procedural calls modify corresponding training parameters in the plurality of training parameters for respective training cycles in the training job. The system comprises a trainer configured to execute the training cycles in dependence on the modified training parameters. The trainer can determine a performance accuracy of the model instances for each of the executed training cycles.

Abstract: A method, a system, and a computer program product for detecting a diverse set of rare behavior. A time-series data representing one or more actions executed by an entity is received from a plurality of time-series data sources and is processed. A data structure corresponding to the entity, identifying the entity, and including one or more representations of processed time-series data identifying the actions is generated. A current action executed by the entity is detected. Current time-series data corresponding to the current action is received and associated with the data structure. First features are extracted from the generated data structure based on current time-series data and compared to second features extracted for at least another entity to determine difference parameters between first and second features. One or more models are trained using difference parameters, and a score for each action executed by the entity is determined.

Abstract: An electric drive unit and method of assembling the same is disclosed. The electric drive unit includes a rotor having a rotor shaft, and gear shaft, where the rotor shaft is inserted into the gear shaft. The gear shaft is supported by two bearings, while the rotor shaft supported directly at one end by a bearing and at the other by the gear shaft. A wave spring is also disclosed that provides an axial loading to the rotor shaft. Also disclosed is a balancing ring secured to an end of the rotor via a locknut. The balancing ring can be machined in order to balance the rotor. The rotor shaft can be connected to the gear shaft via a spline connection. The rotor shaft can bear against the gear shaft via a pilot journal and pilot bore defined on the rotor shaft and gear shaft respectively.

Abstract: Disclosed herein are anti-Seizure Related 6 Homolog Like 2 (SEZ6L2) antibodies and antibody drug conjugates (ADCs), including compositions and methods of using said antibodies and ADCs.

Abstract: In accordance with the present approach, a collaborative chat system is provided that enables collaborative communication between a user preferring a native chat platform and a different user who communicates using a non-native or third-party chat platform. The user may particularly select various other users to be added to a collaborative chat, regardless of which chat platform the other users have set as their main or preferred communication channel. Indeed, when displaying the various users available to be added to the collaborative chat, the collaborative chat system may display a visual indication or icon that indicates which chat platform the various users are available through. The collaborative chat system or platform may utilize bi-directional communications between the native chat platform and any communicatively coupled third-party chat platforms to link the user with the selected other users within a collaborative chat instance.

Abstract: System, methods, and other embodiments described herein relate to providing visual cues on a steering apparatus of a vehicle. In one embodiment, a method includes determining a difference between a current angle of a steering apparatus of a vehicle and a recommended angle of the steering apparatus. The method further includes illuminating a region of a grip of the steering apparatus based upon the difference and a position of a hand of an operator of the vehicle on the grip. The method additionally includes unilluminating the region as the steering apparatus is rotated from the current angle to the recommended angle.

Abstract: Protocol detection circuits for detecting the field device protocol of a field device attached to a channel having a data line are sequentially connected to the data line by a controller that communicates to the field device through the protocol detection circuits. The protocol detection circuits are disconnected from the field device after protocol detection is completed. The channel may be one of a number of channels; the controller may also connect and disconnect the protocol detection circuits from the one channel whereby the protocol detection circuits can be used with any channel. After the protocol of the field device is identified a protocol adapter can be inserted into the channel to enable data conversion between the field device protocol and an another protocol used by the channel for data transmission.

Abstract: An access door may include a core, stiles on opposites sides of the core, an elongated panel extending transversely between the stiles, wherein a major face of the door between the stiles defines one or more recessed pockets along a top edge of the elongated panel between and the stiles and the elongated panel, and a notch formed into the top edge of the elongated panel and extending along at least one of the recessed pockets, the notch extending downwardly into the top edge of the elongated panel so as to form a lip in the top edge of the elongated panel with the notch positioned between the lip and the one or more recessed pockets, the lip and the notch of the elongated panel together forming an accessory rail, wherein the accessory rail is configured to receive and support a complementarily configured hanger base of an accessory hanger.

Abstract: An electric motor includes a case, a stator having stator windings, a rotor having a rotor shaft, and at least one main rotor bearing coupling the rotor shaft to the case. Common mode charge builds up on the rotor due to imbalances of the stator. The at least one main rotor bearing has a first electrical resistance between the rotor shaft and the case. The electric motor further includes a rotor discharge bearing coupling the rotor shaft to the case. The rotor discharge bearing has a second electrical resistance between the rotor shaft and the case that is less than the first electrical resistance, causing the common mode charge to discharge through the rotor discharge bearing. In a differing embodiment, a rotor discharge brush coupled between the case and the rotor shaft along the longitudinal axis of the rotor shaft discharges the common mode charge.

Abstract: The technology disclosed relates to authenticating users using a plurality of non-deterministic registration biometric inputs. During registration, a plurality of non-deterministic biometric inputs are given as input to a trained machine learning model to generate sets of feature vectors. The non-deterministic biometric inputs can include a plurality of face images and a plurality of voice samples of a user. A characteristic identity vector for the user can be determined by averaging feature vectors. During authentication, a plurality of non-deterministic biometric inputs are given as input to a trained machine learning model to generate a set of authentication feature vectors. The sets of feature vectors are projected onto a surface of a hyper-sphere. The system can authenticate the user when a cosine distance between the authentication feature vector and a characteristic identity vector for the user is less than a pre-determined threshold.

Abstract: Systems and methods for generating concise explanations of scored observations that strike good, and computationally efficient, trade-offs between rank-ordering performance and explainability of scored observations are disclosed. The systems and methods described herein for explaining scored observations are based on a framework of partial dependence functions (PDFs), multi-layered neural networks (MNNs), and Latent Explanations Neural Network Scoring (LENNS).

Abstract: An exemplary controller may include a single clock source configured to generate a single clock signal used to drive one or more components within a plurality of magnetometers and a plurality of differential signal measurement circuits configured to measure current output by a photodetector of each of the plurality of magnetometers.

Abstract: A system and method for constructing an improved computing model that preserves use rights for data utilized by the model. A first dataset is accessed to build a computing model. The first data set is subject to terminable usage rights provisions. A portion of the first dataset is sampled to generate a second dataset. Vectors present in the first dataset and the second dataset are discretized. In response to determine that the usage rights associated with the primary dataset have been terminated, a coverage depletion for the second dataset is computed based on the usage rights termination associated with the first dataset. An estimated mean time to coverage failure for the first model based on the depletion coverage is determined for the second dataset. One or more data points are removed from the first dataset due to the termination of usage rights.

Abstract: Techniques for generating an interactive and dynamically updated content stream are described herein. Data assets that correspond to video segments provided by a plurality of content providers may be maintained. A video segment may correspond to a first portion of a full video segment for a piece of content from the plurality of content providers. A content stream may be generated that includes video segments provided by the plurality of content providers based at least in part on the data assets. The content stream may be presented via an application. Second input may be received via the application during presentation of the content stream that corresponds to a navigation command for displaying different content. The content stream may be updated to at least one of removing certain video segments or adding new video segments to the content stream based at least in part on the second input.