119 results about "Gradient function" patented technology

The invention discloses a hollowed-out lattice sandwich layer with gradient functionality. The hollowed-out lattice sandwich layer with gradient functionality comprises two cladding plates, wherein a core sandwich layer is arranged between the cladding plates and comprises a plurality of hollowed-out lattice filling units which are connected with one another; and each hollowed-out lattice filling unit consists of a plurality of tiny dense hollowed-out base cells. The hollowed-out sandwich layer is formed by the hollowed-out lattice filling units, and is light in structure weight and high in specific strength, the maximum weight reduction benefit can reach 85% or above, and meanwhile, the hollowed-out lattice sandwich layer has functional benefits such as heat insulation and sound insulation; comprehensive benefits of structure light weight and functionalization are improved, property mismatching of a sandwich board with the single structure in the application field under various working conditions is avoided, and finally, the gradient function is fulfilled; and meanwhile, a selective laser melting technology is adopted, the risk of reduction of strength of a connecting region between a traditional sandwich layer and an outer cladding plate and complicated multi-process flow can be avoided, and therefore, the use value is quite high.

The apparatus and method invented are operating upon a digital image signal obtained from an image sensor. The sensor is covered with different colored filters and is only able to record the color transmitted through each specific filter into the photosite or pixel. This type of sensor is known as a color filter array or CFA sensor. The different colored filters are arranged in a predefined pattern across the sensor. To obtain a full color image the missing color information is estimated by a set of weighed values obtained by an inverted gradient function. The set of weighted values is found from the neighboring pixels in the four compass directions, north, east, west and south or is found horizontally and vertically. The surrounding pixels are corrected by the chrominance channel to better fit the center pixel in the luminance channel, prior to using the gradient functions. The chrominance channel is interpolated in a similar manner and is also corrected to better fit the center pixel before the inverted gradient functions are applied.

The present invention relates to aluminum electrolysis, and is especially inert ceramet anode. The ceramet material has structure of at least two layers and is rod shaped, plate shaped or cup shaped. The ceramic phase is spinelle type composite oxide; and the metal phase is made of Cu, Ni, Co, Fe, Ag, Pt, Au and other metal or their binary or multiple-element alloy. The composite material with gradient function has excellent comprehensive performance under electrolysis, and the anode of the composite material needs no frequent replacement. Using the composite material of the present invention can alter the electrolytic bath structure, raise production efficiency, lower aluminum ingot producing cost, improve environment, and reduce power consumption.

The invention provides a silicon nitride-based nano composite functionally gradient ceramic cutter material and a preparation method thereof. The ceramic cutter material is provided with a five-layer symmetrical gradient level structure; the component contents of the symmetrical layers relative to a central layer are the same and the thickness of each layer is consistent; the components of each layer are all micron Si3N4, nano Si3N4, nano TiC, nano TiN, Al2O3 and Y2O3. The preparation method thereof is as follows: (1) preparing materials according to the content of each component in each layer; (2) dispersing the nano materials in each layer; (3) mixing the other materials in each layer with the dispersed nano materials to obtain the powder material of a composite ceramic material which disperses excellently; and (4) employing a powder layering paving method and a hot pressing sintering technique to carry out loading and sintering. The invention improves the anti-bending intensity and the fracture toughness property of the materials, and improves the fracture toughness property of the materials; while the gradients of the nano TiC and the nano TiN are arranged in layers, thus leading the mechanical property of the cutter material to be changed in stagewise gradient, and being capable of effectively easing the residual heat stress.

The invention discloses a gradient descent calculation method for protecting privacy data. The method is used in gradient function computation of machine learning including one or more data providers,a decryption service provider, and a computing resource provider,. The method is suitable for carrying out fitting calculation on a sigmoid function by utilizing a polynomial function similar to thesigmoid function contour or carrying out fitting calculation on the sigmoid function by utilizing a piecewise function similar to the sigmoid function contour. The method comprises the steps of homomorphic encryption key generation and distribution, training parameter negotiation, data encryption and summarization and gradient descent. The method is high in precision, and precision loss caused bydata processing in the calculation process is within a controllable range; the security is high, and the input data and the intermediate data can meet the semantic security requirements in the calculation process; flexibility is good, and two or more participants can participate in calculation; and the expansibility is good, and the original gradient descent can be expanded to a Newton method or abatch gradient descent.

A reinforcement learning system (1) of the present invention utilizes a value of a first value gradient function (dV1 / dt) in the learning performed by a second learning device (122), namely in evaluating a second reward (r2(t)). The first value gradient function (dV1 / dt) is a temporal differential of a first value function (V1) which is defined according to a first reward (r1(t)) obtained from an environment and is served as a learning result given by a first learning device (121). An action policy which should be taken by a robot (R) to execute a task is determined based on the second reward (r2(t)).

The invention discloses an automatic focusing system of a digital camera for non-uniform sampling window planning based on golden section. A method comprises the following steps of: carrying out the golden section on a shooting scene for many times in automatic focus window planning; carrying out non-uniform sampling on an area where golden section points are located; and endowing the non-uniform sampling area with different weight values by using an energy gradient function based on a green component as a focusing evaluation function. Therefore, complexity of calculation is reduced on the premise of guaranteeing the focusing accuracy of an imaging body and a background area, and a real-time requirement of the automatic focusing system of the digital camera is met.

The invention discloses an image-based target point position detection method, belongs to the technical field of computers, and is used for solving the problem of low target point position detection performance. The point position detection method disclosed by the embodiment of the invention comprises the following steps: acquiring a plurality of image samples marked with real target point positions; calculating a loss value of the neural network model according to a variable through a loss function of the neural network model, adjusting parameters of the neural network model by taking the minimum loss value as a target, and training the neural network model; wherein the loss function is configured to monotonically increase relative to the absolute value of a variable, the gradient function corresponding to the loss function is a monotonically decreasing function, the value of the gradient function is greater than 1, and the gradient function gradually approaches 1 when the variable approaches infinite; wherein the variable is used for representing a difference value between a true value and a predicted value, the true value is determined based on a sample label, and the predictedvalue is predicted through the neural network model.

A method of preparing a hydroxyapatite nano composite coating with gredient function relates to a method of preparing the hydroxyapatite nano composite coating on the surface of titanium or titanium alloy. The invention provides a method of preparing the hydroxyapatite nano composite coating with gredient function on the surface of titanium or titanium alloy by multi-layer electrophoretic codeposition of the hydroxyapatite nano powder and the metal A1 micro powder, reaction, sintering and heat treatment, so as to improve the bonding strength between the hydroxyapatite nano composite coating and the basal. The titanium or titanium alloy after mechanical finishing, alkaline washing and degreasing as well as acid etch treatment receive at least three times of electrophoretic deposition subsequently in the suspension with different Hap powder and A1 powder content in the order of gradually reduced content of A1; DC current is adopted to deposit the voltage 30 to 40V for 15 to 30 minutes; and then the sample reacts and receives sintering and heat treatment in a tube type resistance furnace to obtain the hydroxyapatite nano composite coating with gredient function, and the coating is not easy to crack during the preparation process.

The invention belongs to the technical field of wireless communication, and particularly relates to a congestion control based wireless sensor network routing optimization method. The method comprises the steps of establishing a link gradient function from each node to a prior neighbor node according to geographical location information of the nodes; constructing a node queue model, and establishing a flow radius function of each link according to the data packet service rate of the nodes; constructing a link flow rate model among the nodes according to the established link gradient function and the flow radius function; establishing a routing forward probability function by the nodes according to difference in each link flow rate so as to complete establishment of routing; and establishing a congestion detection function so as to carry out real-time detection on the congestion degree of a network, and carrying out local routing maintenance once node congestion occurs. According to the invention, the link gradient function is established through using basic conceptions and principles in hydrology for reference, thereby sufficiently realizing effectiveness of node energy consumption, enabling energy consumption of the whole network to be optimized, and having good real-time performance and reliability.

A kind of two-way gradient function ceramics water coal liquid air spout and the manufacturing method it makes the A1203 as the basal body and makes the TiC or (W, Ti)C solid solution as the strengthening phase and the content of TiC or (W, Ti)C presents the two-way gradient variation along the air spout axial and the distribution law of the content decreases gradually from the air spout inlet to the middle and then increases gradually from middle to air spout outlet through allocating multiple kinds of ceramics composite powder and processing by strengthening ball polish, ridding, layer-by layer spreading, preliminary compression, and heat pressing sintering technology preparation. The inlet and outlet of the air spout have high flexibility and the middle if it has high degree of hardness used to release the heat stress of outlet area when the ceramics water coal liquid air spout working and the impact pressure of the inlet area and improve its heat-resistant ablating and abrasion ability. The ceramics water coal liquid air spout's heat-resistant ablating and abrasion ability is higher than the homogeneous ceramics water coal liquid air spout and the one-way gradient function air spout apparently and the use life is prolonged largely.