Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using skill templates for robotic demonstration learning. One of the methods includes receiving a skill template for a task to be performed by a robot, wherein the skill template defines a state machine having a plurality of subtasks and one or more respective transition conditions between one or more of the subtasks. Local demonstration data for a demonstration subtask of the skill template is received, where the local demonstration data is generated from a user demonstrating how to perform the demonstration subtask with the robot. A machine learning model is refined for the demonstration subtask and the skill template is executed on the robot, causing the robot to transition through the state machine defined by the skill template to perform the task.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for integrating sensor streams for robotic demonstration learning. One of the methods includes selecting, by a learning system for a robot, a base update rate for combining multiple sensor streams into a task state representation. The learning system repeatedly generates the task state representation at the base update rate, including combining, during each time period defined by the update rate, the task state representation from most recently updated sensor data processed by the plurality of neural networks. The learning system repeatedly uses the task state representations to generate commands for the robot at the base update rate.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for distributing skill templates for robotic demonstration learning. One of the methods includes receiving, from the user device by a skill template distribution system, a selection of an available skill template. The skill template distribution system provides a skill template, wherein the skill template comprises information representing a state machine of one or more tasks, and wherein the skill template specifies which of the one or more tasks are demonstration subtasks requiring local demonstration data. The skill template distribution system trains a machine learning model for the demonstration subtask using a local demonstration data to generate learned parameter values.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using simulated local demonstration data for robotic demonstration learning. One of the methods includes receiving perceptual data of a workcell of a robot to be configured to execute a task according to a skill template, wherein the skill template specifies one or more subtasks required to perform the skill, wherein at least one of the subtasks is a demonstration subtask that relies on learning visual characteristics of the workcell. A virtual model is generated of a portion of the workcell. A training system generates simulated local demonstration data from the virtual model of the portion of the workcell and tunes a base control policy for the demonstration subtask using the simulated local demonstration data generated from the virtual model of the portion of the workcell.

Abstract: A semiconductor device is disclosed including one or more semiconductor dies mounted on substrate. Each semiconductor die may be formed with a ferromagnetic layer on a lower, inactive surface of the semiconductor die. The ferromagnetic layer pulls the semiconductor dies down against each other and the substrate during fabrication to prevent warping of the dies. The ferromagnetic layer also balances out a mismatch of coefficients of thermal expansion between layers of the dies, thus further preventing warping of the dies.

Abstract: An alkali metal monohydrogen cyanurate compound of the chemical formula AM(HC3N3O3).nH2O (specifically such as KLi(HC3N3O3).2H2O, RbLi(HC3N3O3).2H2O, RbNa(HC3N3O3).2H2O) and a nonlinear optical crystal thereof are related to optoelectronic functional materials. Measured using a powder frequency doubling test method, and the powder frequency doubling effect of the nonlinear optical crystal is about 2-3 times that of KH2PO4 (KDP). The ultraviolet absorption edge of the nonlinear optical crystal is shorter than 250 nm. The nonlinear optical crystal can achieve the harmonic generator of double, triple, or quadruple frequency for Nd:YAG (?=1.064 ?m). Moreover, the nonlinear optical crystal is of a single crystalline structure, is colorless and transparent, and does not deliquesce in air.

Abstract: The present invention provides methods for the synthesis of ketones involving a Ni/Zr-mediated coupling reaction. The Ni/Zr-mediated ketolization reactions can be used in the synthesis of halichondrins (e.g., halichondrin A, B, C; homohalichondrin A, B, C; norhalichondrin A, B, C), and analogs thereof. Therefore, the present invention also provides synthetic methods useful for the synthesis of halichondrins, and analogs thereof. Also provided herein are compounds (i.e., intermediates) useful in the synthesis of halichondrins, and analogs thereof.

Abstract: Methods and systems for transparent user interface integration between remote (“published”) applications and their local counterparts are described, providing a seamless, unified user experience, and allowing integration of a start menu, dock, taskbar, desktop shortcuts, windows, window and application switching, system tray elements, client-to-host and host-to-client file type association, URL redirection, browser cookie redirection, token redirection, status message interception and redirection, and other elements. These methods and systems further enhance theme-integration between a client and remote desktop or virtual machine by remoting all UI elements to a recipient for generation, including text controls, buttons, progress bars, radio buttons, list boxes, or other elements; presenting them with the receiver's product and OS-specific UI; and returning status back to the sender. This may achieve a more unified and transparent UI integration.

Abstract: Provided are a photoalignment method and a photoalignment device, including: adjusting direction of light ray emitted from light source so that first angle is formed between that and XOY plane, and second angle is formed between projection of that on XOY plane and Y-axis; adjusting angle of polarizing plate so that it is parallel to XOY plane, and its light transmission axis is parallel to projection of that on XOY plane; moving array substrate along negative direction of Y-axis to perform exposure operation on array substrate; and moving color filter substrate, which rotated by 180 degrees, in negative direction of Y-axis to perform exposure operation on color filter substrate, wherein color filter substrate has exposure regions which are completely identical, in position and number, to those of array substrate, and each exposure region is exposed to light in identical direction to that of corresponding exposure region of array substrate.

Abstract: Crystalline NH4Be2BO3F2 or Be2BO3F (abbreviated as BBF) has nonlinear optical effect, is not deliquescent in the air, is chemically stable. They can be used in a variety of nonlinear optical fields and will pioneer the nonlinear optical applications in the deep UV band.

Abstract: A semiconductor device is disclosed including one or more semiconductor dies mounted on substrate. Each semiconductor die may be formed with a ferromagnetic layer on a lower, inactive surface of the semiconductor die. The ferromagnetic layer pulls the semiconductor dies down against each other and the substrate during fabrication to prevent warping of the dies. The ferromagnetic layer also balances out a mismatch of coefficients of thermal expansion between layers of the dies, thus further preventing warping of the dies.

Abstract: Crystalline NH4Be2BO3F2 or Be2BO3F (abbreviated as BBF) has nonlinear optical effect, is not deliquescent in the air, is chemically stable. They can be used in a variety of nonlinear optical fields and will pioneer the nonlinear optical applications in the deep UV band.

Abstract: A substrate semiconductor layer is attached to a carrier substrate through a sacrificial bonding material layer. A plurality of semiconductor dies included within continuous material layers are formed on a front side of the substrate semiconductor layer. Each of the continuous material layers continuously extends over areas of the plurality of semiconductor dies. A plurality of dicing channels is formed between neighboring pairs among the plurality of semiconductor dies by anisotropically etching portions of the continuous material layers located between neighboring pairs of semiconductor dies. The plurality of dicing channels extends to a top surface of the sacrificial bonding material layer. The sacrificial bonding material layer is removed selective to materials of surface portions of the plurality of semiconductor dies using an isotropic etch process. The plurality of semiconductor dies is singulated from one another upon removal of the sacrificial bonding material layer.

Abstract: Methods and systems for transparent user interface integration between remote (“published”) applications and their local counterparts are described, providing a seamless, unified user experience, and allowing integration of a start menu, dock, taskbar, desktop shortcuts, windows, window and application switching, system tray elements, client-to-host and host-to-client file type association, URL redirection, browser cookie redirection, token redirection, status message interception and redirection, and other elements. These methods and systems further enhance theme-integration between a client and remote desktop or virtual machine by remoting all UI elements to a recipient for generation, including text controls, buttons, progress bars, radio buttons, list boxes, or other elements; presenting them with the receiver's product and OS-specific UI; and returning status back to the sender. This may achieve a more unified and transparent UI integration.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for jointly training an image embedding model and a text embedding model. In one aspect, a method comprises: processing data from a historical query log of a search system to generate a candidate set of training examples, wherein each training example comprises: (i) a search query comprising a sequence of one or more words, (ii) an image, and (iii) selection data characterizing how often users selected the image in response to the image being identified by a search result for the search query; selecting a plurality of training examples from the candidate set of training examples; and using the training data to jointly train the image embedding model and the text embedding model.

Abstract: Methods, systems, and computer readable medium for personalizing an analytics user interface. The method includes generating a set of training data from received user interaction data, inputting the set of training data to a machine learning model to train the model, generating a set of user interest scores for the particular user that each indicate a user's interest in accessing information corresponding to a UI element of the application, determining, from the user interest scores, that the user is interested in a particular UI element that was not included in the initial UI and has at least a threshold score, dynamically modifying the initial UI to include the particular UI element, presenting the updated UI, monitoring further user interactions, updating the model based on the further user interactions, and modifying the updated UI based on the updated model.

Abstract: A substrate semiconductor layer is attached to a carrier substrate through a sacrificial bonding material layer. A plurality of semiconductor dies included within continuous material layers are formed on a front side of the substrate semiconductor layer. Each of the continuous material layers continuously extends over areas of the plurality of semiconductor dies. A plurality of dicing channels is formed between neighboring pairs among the plurality of semiconductor dies by anisotropically etching portions of the continuous material layers located between neighboring pairs of semiconductor dies. The plurality of dicing channels extends to a top surface of the sacrificial bonding material layer. The sacrificial bonding material layer is removed selective to materials of surface portions of the plurality of semiconductor dies using an isotropic etch process. The plurality of semiconductor dies is singulated from one another upon removal of the sacrificial bonding material layer.

Abstract: A semiconductor device is disclosed having reinforced supports at corners of the device. The semiconductor device may include solder balls on a lower surface of the device for soldering the device onto a printed circuit board. In one example, the solder balls at the corners of the semiconductor device may be replaced by support billets having more mass and more contact area between the semiconductor device and the PCB. In a further example, screws may be provided at the corners of the device (instead of the corner solder balls or in addition to the corner solder balls). These screws may be placed through the corners of the semiconductor device and into the printed circuit board.

Abstract: A semiconductor device vertically mounted on a medium such as a printed circuit board, and a method of its manufacture, are disclosed. The semiconductor device includes a stack of semiconductor die having contact pads which extend to an active edge of the die aligned on one side of the stack. The active edges of the die are affixed to the PCB and the contact pads at the active edge are electrically coupled to the PCB. This configuration provides an optimal, high density arrangement of semiconductor die in the device, where a large number of semiconductor die can be mounted and electrically coupled directly to the PCT, without a substrate, without staggering the semiconductor die, and without using wire bonds.

Abstract: Methods and systems for transparent user interface integration between remote (“published”) applications and their local counterparts are described, providing a seamless, unified user experience, and allowing integration of a start menu, dock, taskbar, desktop shortcuts, windows, window and application switching, system tray elements, client-to-host and host-to-client file type association, URL redirection, browser cookie redirection, token redirection, status message interception and redirection, and other elements. These methods and systems further enhance theme-integration between a client and remote desktop or virtual machine by remoting all UI elements to a recipient for generation, including text controls, buttons, progress bars, radio buttons, list boxes, or other elements; presenting them with the receiver's product and OS-specific UI; and returning status back to the sender. This may achieve a more unified and transparent UI integration.