Abstract: Systems and methods for training models to improve fairness.

Abstract: Systems and methods for training models to improve fairness.

Abstract: A method of reconfiguring a modular wall is provided. The modular wall comprises a panel receiver structure with a panel receiving channel; a vertical panel received in the channel of the panel receiver structure; and a first and second fascia trim members. Each of the first and second fascia trim members is configured to be removably attached to opposing first and second sides of the panel receiver structure. The fascia trim members are configured to conceal the panel receiver structure. The method comprises removing at least the second fascia trim member; positioning a second panel receiver structure on the second side of the panel receiver structure, the second panel receiver structure having a second panel receiving channel; positioning a second vertical panel in the second panel receiving channel to be parallel to the vertical panel; and re-installing the removed second fascia trim member to the second panel receiver structure.

Abstract: A system and connector interconnects modular wall vertical frame members meeting at a junction. A first vertical frame member and a second vertical frame member each have a central section for mounting a wall panel. Each vertical frame member has a pair of lateral legs at the lateral edges thereof and a pair of inner legs extending inwardly. The inner legs are spaced laterally away from the lateral edges. Each vertical frame member also has a pair of interlocking members positioned inwardly of the central section. The interlocking members each having a laterally extending leg defining a laterally extending latch receiving space. The connector member has a head or heads for supporting parts and a latch or latches for securing parts, including trim members. The junctions may be corner or end-to-end.

Abstract: A pre-assembled internal frame for a modular wall comprises a pair of vertical frame members. Each first flange of the vertical frame member has a pair of series of vertically arranged mounting openings, and each second flange has a single series. The vertical frame members are oriented to enable the frame to be arranged in an end to end junction with another frame with adjacent vertical frame members thereof abutting. This provides three of the series of vertically arranged openings on each side of the modular wall at the end to end junction. A fastener is positioned in a fastener receiving portion for fastening the horizontal frame member to the vertical frame members. The fastener receiving portions are recessed laterally inwardly from the first and second web sections for avoiding interference between adjacent fasteners when the frame is arranged in the end to end junction with another such frame.

Abstract: A fault tolerant computer system and method are disclosed. The system may include a plurality of CPU nodes, each including: a processor and a memory; at least two IO domains, wherein at least one of the IO domains is designated an active IO domain performing communication functions for the active CPU nodes; and a switching fabric connecting each CPU node to each IO domain. One CPU node is designated a standby CPU node and the remainder are designated as active CPU nodes. If a failure, a beginning of a failure, or a predicted failure occurs in an active node, the state and memory of the active CPU node are transferred to the standby CPU node which becomes the new active CPU node. If a failure occurs in an active IO domain, the communication functions performed by the failing active IO domain are transferred to the other IO domain.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for machine learning model fairness and explainability. In some implementations, a method includes obtaining data relating to a plurality of potential borrowers; providing the data to the trained machine learning model; obtaining, by the trained machine learning model’s processing of the provided data, the one or more outputs of the trained machine learning model; and automatically generating a report that explains the one or more outputs of the trained machine learning model with respect to one or more fairness metrics and one or more accuracy metrics; and providing the automatically generated report for display on a user device.

Abstract: A fault tolerant computer system and method are disclosed. The system may include a plurality of CPU nodes, each including: a processor and a memory; at least two IO domains, wherein at least one of the IO domains is designated an active IO domain performing communication functions for the active CPU nodes; and a switching fabric connecting each CPU node to each IO domain. One CPU node is designated a standby CPU node and the remainder are designated as active CPU nodes. If a failure, a beginning of a failure, or a predicted failure occurs in an active node, the state and memory of the active CPU node are transferred to the standby CPU node which becomes the new active CPU node. If a failure occurs in an active IO domain, the communication functions performed by the failing active IO domain are transferred to the other IO domain.

Abstract: The present invention relates to methods for determining affinity, binding kinetics and/or concentration of an antibody or of an antibody mixture specific for a virus using virus-like particles (VLPs) and/or live viruses or inactivated viruses attached to biosensors. Further, the present invention relates to the VLPs and live or inactivated viruses attached to biosensors and methods for producing them.

Abstract: Methods, systems, and computer readable media for utilizing ultrasound multiple scattering to assess pulmonary fibrosis are disclosed. One method may comprise transmitting ultrasound energy into the lungs of a subject, detecting portions of the ultrasound energy backscattered by structures in the lungs of the subject, calculating a transport parameter from the detected portions of the ultrasound energy backscattered by the structures in the lungs of the subject, inputting the transport parameter into a model that correlates values of the transport parameter with quantitative indications of severity of pulmonary fibrosis, and receiving, from the model, a quantitative indication of the severity of pulmonary fibrosis for the subject that correlates with the transport parameter for the subject.

Abstract: Grinders, analyzers, and related technologies are described herein. The grinders can hold foodstuff that is periodically ground. The analyzers can analyze the foodstuff to determine information about the state of the foodstuff. Algorithms can be used to determine how to process the foodstuff, how to use the foodstuff, and/or when to discard the foodstuff. The grinder can be a portable, rechargeable electric coffee grinder configured to monitor the freshness of the coffee beans. When coffee beans become stale, they can be discarded and the coffee grinder can be refilled with fresh coffee beans.

Abstract: Systems and methods for training models to improve fairness.

Abstract: This disclosure describes techniques implemented partly by a cloud-based service provider for providing a sentiment-analysis service to determine whether sentiments, or attitudes, of users have significantly decreased. The sentiment-analysis service is composed of various services provided by the cloud-based service in order to determine significant negative deviations in sentiment of users towards an entity. Further, the sentiment-analysis service utilizes various techniques, such as encryption, to privatize the process such that private user information is protected. In some examples, an entity may not know the content of communications or the reasons the user's sentiment has dropped, but may be alerted of a general negative change in sentiment towards the entity by the user.

Abstract: Systems and methods for training models to improve fairness.

Abstract: A device for reinforcing a femoral neck bone, the femoral neck bone including a cortical bone region and a cancellous bone region, the device including a support structure having an outer surface and including a portion having one or more bores, the one or more bores adapted to allow filler material to be released therefrom to enter directly into the cancellous bone.

Abstract: Systems and methods for training models to improve fairness.

Abstract: This disclosure describes techniques implemented partly by a cloud-based service provider for providing a sentiment-analysis service to determine whether sentiments, or attitudes, of users have significantly decreased. The sentiment-analysis service is composed of various services provided by the cloud-based service in order to determine significant negative deviations in sentiment of users towards an entity. Further, the sentiment-analysis service utilizes various techniques, such as encryption, to privatize the process such that private user information is protected. In some examples, an entity may not know the content of communications or the reasons the user's sentiment has dropped, but may be alerted of a general negative change in sentiment towards the entity by the user.

Abstract: A device for reinforcing a femoral neck bone, the femoral neck bone including a cortical bone region and a cancellous bone region, the device including a support structure having an outer surface and including a portion having one or more bores, the one or more bores adapted to allow filler material to be released therefrom to enter directly into the cancellous bone.

Abstract: A computer-implemented method includes generating a computer user interface to accept physical counts of items associated with events recorded by a collection of electronic devices. A first user interface screen includes a list of a first subset of electronic devices from the collection of electronic devices and for each electronic device in the first subset, a field indicating a physical count of items associated with the electronic device. A second user interface screen includes input fields for receiving physical counts of the items and a menu of a second subset of the collection of electronic devices, wherein none of the electronic devices in the second subset are in the first subset such that it is impossible for a user to select an electronic device displayed on the first screen.

Abstract: A fault tolerant computer system and method are disclosed. The system may include a plurality of CPU nodes, each including: a processor and a memory; at least two IO domains, wherein at least one of the IO domains is designated an active IO domain performing communication functions for the active CPU nodes; and a switching fabric connecting each CPU node to each IO domain. One CPU node is designated a standby CPU node and the remainder are designated as active CPU nodes. If a failure, a beginning of a failure, or a predicted failure occurs in an active node, the state and memory of the active CPU node are transferred to the standby CPU node which becomes the new active CPU node. If a failure occurs in an active IO domain, the communication functions performed by the failing active IO domain are transferred to the other IO domain.