Abstract: A door lock and a refrigerator equipped with the door lock. The door lock includes a lock hook and a lock body; the lock body includes a lock housing and a partition arranged inside the lock housing, the partition divides the internal space of the lock housing into a lock core area and a lock hook connection area, the partition is provided with a through hole connecting the lock core area and the lock hook connection area; the lock housing is provided with a lock opening connected to the lock hook connection area; a latch bolt and a driving element are installed in the lock core area, the driving element drives the latch bolt to perform linear reciprocating motion between a locked position and an unlocked position.

Abstract: Techniques for machine learning (ML) model adaptation via segment replacement and student-teacher training are described. A model optimizer determines that a ML model sought to be deployed to an edge device includes a source segment that is not supported by the edge device. The model optimizer identifies a replacement segment that is equivalent to the source segment, constructs an adapted ML model by swapping in the replacement segment for the source segment, and trains the adapted ML model, optionally from the source ML model using a student-teacher training procedure. The trained adapted ML model is then deployed to the edge device, where it can be successfully run with minimal (if any) degradation of performance compared to the original source ML model.

Abstract: A volume liquid fixed-ratio multi-batch mixing valve includes a valve body, including a first inlet liquid chamber, a second inlet liquid chamber and a mixing chamber; a first valve core, controlling liquid inflow into the first inlet liquid chamber; a second valve core, controlling liquid inflow into the second inlet liquid chamber; a liquid drainage valve core, controlling liquid drainage; a rotary shaft includes a first impeller and a second impeller. The first impeller and the second impeller alternately control flow to the first inlet liquid chamber and the second inlet liquid chamber and to the second inlet liquid chamber and the mixing chamber. The first impeller and the second impeller are symmetrically about the rotary shaft; when the first inlet liquid chamber and the second inlet liquid chamber are communicated, the second valve core closes the second inlet liquid chamber and the mixing chamber.

Abstract: A volume liquid fixed-ratio multi-batch mixing valve includes a valve body, including a first inlet liquid chamber, a second inlet liquid chamber and a mixing chamber; a first valve core, controlling liquid inflow into the first inlet liquid chamber; a second valve core, controlling liquid inflow into the second inlet liquid chamber; a liquid drainage valve core, controlling liquid drainage; a rotary shaft includes a first impeller and a second impeller. The first impeller and the second impeller alternately control flow to the first inlet liquid chamber and the second inlet liquid chamber and to the second inlet liquid chamber and the mixing chamber. The first impeller and the second impeller are symmetrically about the rotary shaft; when the first inlet liquid chamber and the second inlet liquid chamber are communicated, the second valve core closes the second inlet liquid chamber and the mixing chamber.

Abstract: Techniques for training a machine learning model based on captured images are described. A method described include filtering a first set of collected images using one or more machine learning models; labeling the first set of filtered, collected images using a data labeling service using a service of the provider network; training a machine learning model from a machine learning algorithm using the first set of filtered, collected images using a service of the provider network; and causing deployment of the trained machine learning model onto a device.

Abstract: Techniques for high-performance machine learning (ML) inference in heterogenous edge devices are described. A ML model trained using a variety of different frameworks is translated into a common format that is runnable by inferences engines of edge devices. The translated model is optimized in hardware-agnostic and/or hardware-specific ways to improve inference performance, and the optimized model is sent to the edge devices. The inference engine for any edge device can be accessed by a customer application using a same defined API, regardless of the hardware characteristics of the edge device or the original format of the ML model.

Abstract: Embodiments of the present disclosure disclose a method and apparatus for automatically discovering a gas station point of interest (POI), a storage medium and a device, wherein, the method includes: collecting information uploaded by an on-board diagnostic (OBD) device, and determining gasoline quantity surge points based on the information; clustering the gasoline quantity surge points based on location information by using a clustering algorithm; and determining a location region having the gas station POI based on a result of clustering. The technical schemes provided by the embodiments of the present disclosure are capable of automatically determining the location region having the gas station POI, thereby reducing collection cost of gas station POI information and shortening an updating period for gas station POIs.

Abstract: Techniques for high-performance machine learning (ML) inference in heterogenous edge devices are described. A ML model trained using a variety of different frameworks is translated into a common format that is runnable by inferences engines of edge devices. The translated model is optimized in hardware-agnostic and/or hardware-specific ways to improve inference performance, and the optimized model is sent to the edge devices. The inference engine for any edge device can be accessed by a customer application using a same defined API, regardless of the hardware characteristics of the edge device or the original format of the ML model.

Abstract: Techniques for training a machine learning model based on captured images are described. A method described include filtering a first set of collected images using one or more machine learning models; labeling the first set of filtered, collected images using a data labeling service using a service of the provider network; training a machine learning model from a machine learning algorithm using the first set of filtered, collected images using a service of the provider network; and causing deployment of the trained machine learning model onto a device.

Abstract: Techniques for machine learning (ML) model knowledge distillation and automatic retraining are described. A model adaptation controller obtains samples generated by an edge device and inference values generated based on the samples by a deployed ML model of the edge device. The model adaptation controller runs inference on the samples using a different ML model to generate inferences that can be used to determine whether the performance of the deployed ML model is lacking. If so, the model adaptation controller can retrain the deployed ML model using samples with ground truth values generated by the different ML model, resulting in a light-weight retrained model that can be provisioned to the edge device. This retraining process may be performed iteratively to automatically improve and adapt the ML model running at the edge device.

Abstract: A portable intrinsically-safe device for acquiring vibration signal comprises a vibration signal acquisition unit, a signal processing unit, a signal storage unit and an intrinsically-safe power source, wherein the vibration signal acquisition unit is connected to the signal storage unit through the signal processing unit; the intrinsically-safe power source supplies power for all the units; the signal processing unit and the signal storage unit are integrally packaged in an intrinsically safe manner. The device solves the vibration signal acquisition from a large portable electromechanical equipment, providing a powerful data information for equipment status monitoring and failure analysis. Particularly, intrinsically safe design satisfies the test requirements when the portable equipment is under a complex and harsh environment of a coal mine well. The device is easy to mount, low in power consumption, easy to control, flexible to use and wide in range of application.

Abstract: Embodiments of the present disclosure disclose a method and apparatus for automatically discovering a gas station point of interest (POI), a storage medium and a device, wherein, the method includes: collecting information uploaded by an on-board diagnostic (OBD) device, and determining gasoline quantity surge points based on the information; clustering the gasoline quantity surge points based on location information by using a clustering algorithm; and determining a location region having the gas station POI based on a result of clustering. The technical schemes provided by the embodiments of the present disclosure are capable of automatically determining the location region having the gas station POI, thereby reducing collection cost of gas station POI information and shortening an updating period for gas station POIs.

Abstract: A communications system includes a macro base station, a plurality of UEs (user equipment), a plurality of small cells, and a network through which the macro base station, the UEs, and the small cells communicate with each other, the small cells within a macro coverage area of the macro base station. The macro base station comprises a processor, a memory, and a small cell on/off module which is operable, for each small cell of the plurality of small cells, to: determine an interference metric for the small cell; if the determined interference metric meets a preset condition for the small cell, then determine a loss in signal strength to the UEs associated with the small cell caused by switching off the small cell; and judge whether to switch off the small cell based on at least one of the determined interference metric or the determined loss in signal strength.

Abstract: A communications system includes a macro base station, a plurality of UEs (user equipment), a plurality of small cells, and a network through which the macro base station, the UEs, and the small cells communicate with each other, the small cells within a macro coverage area of the macro base station. The macro base station comprises a processor, a memory, and a small cell on/off module which is operable, for each small cell of the plurality of small cells, to: determine an interference metric for the small cell; if the determined interference metric meets a preset condition for the small cell, then determine a loss in signal strength to the UEs associated with the small cell caused by switching off the small cell; and judge whether to switch off the small cell based on at least one of the determined interference metric or the determined loss in signal strength.

Abstract: Example implementations are directed to systems and methods based on physical broadcast channel (PBCH) muting are utilized to avoid frequent cell selection/reselection and handover in a LTE-advanced heterogeneous network. In the example implementations, a pico eNB that is fully covered by a macro eNB or other pico eNBs transmits blank PBCH such that it is inaccessible to the UEs who perform cell selection/reselection. Furthermore, a macro eNB may handover a UE to the inaccessible pico eNB by signalling the necessary information to the UE to detect the inaccessible pico eNB. Frequent cell selection/reselection and handover may therefore be avoided in a dense deployment situation.

Abstract: Embodiments described herein are directed to a power control scheme for Long Term Evolution Advanced (LTE-A) heterogeneous networks to reduce the interference from macro base stations (BS) to pico user equipment (UE). The embodiments described herein may be used to develop LTE-A heterogeneous networks to balance the achievable throughput between macro and pico UEs and may thereby improve the overall system performance.