The present disclosure provides an Industrial Internet of Things system conducive to system function expansion and system adjustment and control method thereof. The system comprises a user platform, a service platform, a management platform, a sensor network platform, and an object platform. The service platform and the sensor network platform adopt independent layout, and the management platform adopts centralized layout. The management platform is configured to store a control program that drives operation of production line equipment. The management platform is configured to call the control parameters in the database through communicating with the service platform and configure the control parameters in the control program to control the operation of production line equipment; and a data interaction mode between the user platform and the service platform is modifying and deleting the control parameters in the service platform through data transmission between the user platform and the service platform.

The current disclosure describes techniques for managing vertical alignment or overlay in semiconductor manufacturing using machine learning. Alignments of interconnection features in a fan-out WLP process are evaluated and managed through the disclosed techniques. Big data and neural networks system are used to correlate the overlay error source factors with overlay metrology categories. The overlay error source factors include tool related overlay source factors, wafer or die related overlay source factors and processing context related overlay error source factors.

A wire electric discharge machine includes a machine learning device that learns an adjustment of an axis feed command of the wire electric discharge machine. The machine learning device determines an adjustment amount of the axis feed command by using data related to a movement state of an axis, and adjusts the axis feed command based on the determined adjustment amount of the axis feed command. Subsequently, the machine learning device performs machine learning of the adjustment of the axis feed command based on the determined adjustment amount of the axis feed command, the data related to the movement state of the axis, and a reward calculated based on the data related to the movement state of the axis.

Implementations utilize deep reinforcement learning to train a policy neural network that parameterizes a policy for determining a robotic action based on a current state. Some of those implementations collect experience data from multiple robots that operate simultaneously. Each robot generates instances of experience data during iterative performance of episodes that are each explorations of performing a task, and that are each guided based on the policy network and the current policy parameters for the policy network during the episode. The collected experience data is generated during the episodes and is used to train the policy network by iteratively updating policy parameters of the policy network based on a batch of collected experience data. Further, prior to performance of each of a plurality of episodes performed by the robots, the current updated policy parameters can be provided (or retrieved) for utilization in performance of the episode.

A method for detecting a production error of an assembly in a manufacturing facility includes (i) providing sensor data having at least two dimensions, wherein a respective dimension of the sensor data comprises measurement data with respect to the assembly, (ii) performing a dimensional reduction of the sensor data, wherein at least one feature is extracted based on the at least two dimensions of the sensor data, (iii) reconstructing the dimension-reduced sensor data based on the at least one extracted feature to provide reconstructed sensor data, (iv) determining a reconstruction error based on a comparison of the sensor data with the reconstructed sensor data, and (v) detecting the production error of the assembly based on the determined reconstruction error. Also disclosed is a computer program, a device, and a storage medium for this purpose.

A material completeness detection method configured to detect whether materials of a target object in a physical production line are complete, includes: inputting an image of the target object in the physical production line into a material completeness detection algorithm to acquire a first detection result; inputting a virtual model of the target object in a virtual production line into the material completeness detection algorithm to acquire a second detection result, where the virtual production line is a DT of the physical production line; and acquiring a material completeness detection result of the target object based on the first detection result and the second detection result. The embodiments of the present disclosure can realize efficient and accurate material completeness detection.

A tactile sensing system of a robot may include: a plurality of piezoelectric elements disposed at an object, and including a transmission (TX) piezoelectric element and a reception (RX) piezoelectric element; and at least one processor configured to: control the TX piezoelectric element to generate an acoustic wave having a chirp spread spectrum (CSS) at every preset time interval, along a surface of the object; receive, via the RX piezoelectric element, an acoustic wave signal corresponding to the generated acoustic wave; select frequency bands from a plurality of frequency bands of the acoustic wave signal; and estimate a location of a touch input on the surface of the object by inputting the acoustic wave signal of the selected frequency bands into a neural network configured to provide a touch prediction score for each of a plurality of predetermined locations on the surface of the object.

A machine management method for performing planned activities on a machine is provided. Workpieces are manufactured by the machine according to a predetermined schedule during a respective planned activity of the machine. Manufactured workpieces are sorted from a removal area of the machine into a deposition area during a further respective planned activity of the machine according to the predetermined schedule. The method includes determining a manufacturing time for the workpieces by a machine management system based on manufacturing process data of the machine, determining a removal time for sorting the workpieces by the machine management system based on removal process data, changing the predetermined schedule to obtain a changed schedule for the planned activities by the machine management system based on a comparison of the manufacturing time with the removal time, and transferring the changed schedule to a controller of the machine.

Embodiments disclosed herein include a method of training a DNN. A processor initializes an element of an A matrix. The element may include a resistive processing unit. A processor determines incremental weight updates by updating the element with activation values and error values from a weight matrix multiplied by a chopper value. A processor reads an update voltage from the element. A processor determines a chopper product by multiplying the update voltage by the chopper value. A processor directs storage of an element of a hidden matrix. The element of the hidden matrix may include a summation of continuous iterations of the chopper product. A processor updates a corresponding element of a weight matrix based on the element of the hidden matrix reaching a threshold state.

Implementations utilize deep reinforcement learning to train a policy neural network that parameterizes a policy for determining a robotic action based on a current state. Some of those implementations collect experience data from multiple robots that operate simultaneously. Each robot generates instances of experience data during iterative performance of episodes that are each explorations of performing a task, and that are each guided based on the policy network and the current policy parameters for the policy network during the episode. The collected experience data is generated during the episodes and is used to train the policy network by iteratively updating policy parameters of the policy network based on a batch of collected experience data. Further, prior to performance of each of a plurality of episodes performed by the robots, the current updated policy parameters can be provided (or retrieved) for utilization in performance of the episode.