Abstract: A machine learning (ML)-based technique for user behavior analysis that detects when users deviate from expected behavior. A ML model is trained using training data derived from activity data from a first set of users. The model is refined in a computationally-efficient manner by identifying a second set of users that constitute a “watch list.” At a given time, a differential data ingestion operation is then performed to incorporate data for the second set of users into the training data, while also pruning at least a portion of the data set corresponding to data associated with any user included in the first set but not in the second set. These operations update the training data used for the machine learning. The machine learning model is then refined based on the updated training data that incorporates the activity data ingested from the users identified in the watch list.

Abstract: Embodiments of a method and system for inversion of high-resolution aquifer storage coefficient based on gravity satellite data are disclosed. The method comprises: acquiring gravity satellite data, soil moisture storage data, surface water storage data and other non-groundwater component data, based on a preset region. The method also comprises: obtaining monthly time series of terrestrial water storage anomalies, by inversing the gravity satellite data; obtaining monthly time series of soil moisture storage anomalies, surface water storage anomalies, and other non-groundwater component anomalies; obtaining monthly time series of groundwater storage anomalies; and obtaining monthly time series of groundwater level anomalies in different aquifers.

Abstract: The application provides an air purifier comprising a shell assembly having air inlets and air outlets, and a filter assembly, an air inflow guide assembly, a fan and a diffusion air outflow assembly sequentially arranged from back to front. The filter assembly can purify air entering the filter assembly via the air inlets and flowing through the filter assembly. The air inflow guide assembly defines a guide channel between the filter assembly and the fan. The fan generates an air flow that flows into a first receiving cavity via the air inlets and sequentially flows through each assembly. The diffusion air outflow assembly can rotate to guide the air flow to the air outlets, enable the air flow to diffuse outwards and decrease the flow velocity of the air flow. The air purifier expands the purification range of the air purifier and improves the purification effect of the air purifier.

Abstract: A semantic recognition method and a semantic recognition device are provided. A spectrogram of a speech signal is generated. At least one keyword of the spectrogram is detected by inputting the spectrogram into a neural network model. A semantic category to which each of the at least one keyword belongs is distinguished. A semantic intention of the speech signal is determined according to the at least one keyword and the semantic category of the at least one keyword.

Abstract: A blender comprises a stirring cup and a light-emitting assembly which includes a battery, a lamp panel, and two spaced conductive parts, wherein each conductive part has one end electrically connected to the lamp panel and the other end stretching into the stirring cup, and the battery is electrically connected to the lamp panel. Liquids such as water, fruit juices, and liquor make contact with the two conductive parts after being added to the stirring cup, a conduction circuit is formed by the two conductive parts and the lamp panel through the liquids in the stirring cup, and then, the lamp panel lights up. The blender of the invention can automatically light up when used by users, thus having good user experience and increasing the pleasure in use.

Abstract: An object recognition method and system thereof are provided. A recognition result of a first object of a (i?1)th frame of a video stream is obtained. A ith frame is received, and a second object is detected from the ith frame. Whether the first object and the second object are corresponding to the same target object is determined according to a position of the first object in the (i?1)th frame and a position of the second object in the ith frame. If the first object and the second object are corresponding to the same target object, whether a recognition confidence level is greater than a predetermined threshold is determined so as to perform the object recognition on the second object or assign the recognition result of the first object to the second object.

Abstract: A liquid crystal display device proofed against light leaking from a join between backlight module and display panel also has a camera. The display device defines a display area surrounding a light transmission area which allows external light to enter the camera. The backlight module defines a mounting hole corresponding to the light transmission area, the camera being in the mounting hole. Black foam is arranged at the join of the display panel and the backlight module, the black foam has light absorbing properties.

Abstract: An object recognition method and system thereof are provided. A recognition result of a first object of a (i?1)th frame of a video stream is obtained. A ith frame is received, and a second object is detected from the ith frame. Whether the first object and the second object are corresponding to the same target object is determined according to a position of the first object in the (i?1)th frame and a position of the second object in the ith frame. If the first object and the second object are corresponding to the same target object, whether a recognition confidence level is greater than a predetermined threshold is determined so as to perform the object recognition on the second object or assign the recognition result of the first object to the second object.

Abstract: A tiered machine learning-based infrastructure comprises a first machine learning (ML) tier configured to execute within an enterprise network environment and that learns statistics for a set of use cases locally, and to alert deviations from the learned distributions. Use cases typically are independent from one another. A second machine learning tier executes external to the enterprise network environment and provides further learning support, e.g., by determining a correlation among multiple independent use cases that are running locally in the first tier. Preferably, the second tier executes in a cloud compute environment for scalability and performance.

Abstract: A water pipe includes a pipe body (1), a woven layer (2) sleeved around the pipe body (1), a first joint assembly (3) arranged at one end of the pipe body (1) and secured to the pipe body (1), and a second joint assembly (4) arranged at another end of the pipe body (1) and secured to the pipe body (1); the woven layer (2) includes bamboo charcoal fibers and polypropylene filaments. The woven layer (2) is sleeved around the pipe body (1) and is not connected with the pipe body (1), and the woven layer (2) includes bamboo charcoal fibers and polypropylene filaments. Bamboo charcoal fibers have the functions of resisting bacteria and removing peculiar smell and can effectively prevent the generation of bacteria on the water pipe and the emission of peculiar smell when the water pipe is used in humid outdoor environments.

Abstract: A machine learning (ML)-based technique for user behavior analysis that detects when users deviate from expected behavior. A ML model is trained using training data derived from activity data from a first set of users. The model is refined in a computationally-efficient manner by identifying a second set of users that constitute a “watch list.” At a given time, a differential data ingestion operation is then performed to incorporate data for the second set of users into the training data, while also pruning at least a portion of the data set corresponding to data associated with any user included in the first set but not in the second set. These operations update the training data used for the machine learning. The machine learning model is then refined based on the updated training data that incorporates the activity data ingested from the users identified in the watch list.

Abstract: A backlight module having better heat dissipation, includes a backboard, a frame coupled to a peripheral area of the backboard, at least one light source, and a thermally conductive layer located on the backboard. The backboard is made of a thermally conductive material. The at least one light source is embedded in the thermally conductive layer and in direct contact with the thermally conductive layer which removes heat as it is generated by the light source.

Abstract: A water pipe includes a pipe body (1), a woven layer (2) sleeved around the pipe body (1), a first joint assembly (3) arranged at one end of the pipe body (1) and secured to the pipe body (1), and a second joint assembly (4) arranged at another end of the pipe body (1) and secured to the pipe body (1); the woven layer (2) includes bamboo charcoal fibers and polypropylene filaments. The woven layer (2) is sleeved around the pipe body (1) and is not connected with the pipe body (1), and the woven layer (2) includes bamboo charcoal fibers and polypropylene filaments. Bamboo charcoal fibers have the functions of resisting bacteria and removing peculiar smell and can effectively prevent the generation of bacteria on the water pipe and the emission of peculiar smell when the water pipe is used in humid outdoor environments.

Abstract: A clutch matched with a transmission screw rod comprises a left clutch half tooth, a right clutch half tooth, a clutch screw rod and a clutch base, the left clutch half tooth and the right clutch half tooth can slidably move relative to the clutch base; first threads matched with the transmission screw rod are arranged on the right side face of the left clutch half tooth and the left side face of the right clutch half tooth; the clutch screw rod and the transmission screw rod are located in different planes; a second thread and a third thread in opposite directions are arranged on the clutch screw rod; a fourth thread matched with the second thread and a fifth thread matched with the third thread are separately arranged on the left clutch half tooth and the right clutch half tooth.

Abstract: A display apparatus which is not subject to light leakage from a cut edge of a quantum dot layer includes a display panel and a backlighting module. The backlight module includes light source, light guide plate, and quantum dot enhancement layer. The quantum dot enhancement layer has functional and invalidation portions. The functional portion converts light from the light guide plate into predetermined wavelengths. The invalidation portion is formed in the non-display region, and is unable to convert light from the light guide plate into predetermined wavelengths. A distance between a side surface of the invalidation portion and a side surface of the main display region is larger than or equal to a predetermined distance. The effect of the predetermined distance is to prevent the light leaking or transmitted from the invalidation portion from entering the display region at normal viewing angles.

Abstract: A kettle includes a kettle body, a top cap, a heating base, a filter cup and a filter cup lifting arrangement. The kettle body includes a water reservoir having a receiving cavity. The filter cup is adapted for storing food or beverage items and communicating with the receiving cavity. The filter cup lifting arrangement includes a supporting member supported in the water reservoir, a stopper and a lifting member movably extended between the supporting member and the stopper through the top cap. The lifting member is arranged to drive the filter cup to move between an immersion position and a lifted-up position, wherein in the immersion position, the filter cup is moved to immerse under the hot water stored in the receiving cavity, wherein in the lifted-up position, the filter cup is moved to lift up from the hot water.