Abstract: A multithread processor includes a time counter and a register scoreboard and provides a method for statically dispatching instructions with preset execution times based on a write time of a register in the register scoreboard and the time counter provided to an execution pipeline.

Abstract: Systems and methods are provided for vector-quantized image modeling using vision transformers and improved codebook handling. In particular, the present disclosure provides a Vector-quantized Image Modeling (VIM) approach that involves pretraining a machine learning model (e.g., Transformer model) to predict rasterized image tokens autoregressively. The discrete image tokens can be encoded from a learned Vision-Transformer-based VQGAN (example implementations of which can be referred to as ViT-VQGAN). The present disclosure proposes multiple improvements over vanilla VQGAN from architecture to codebook learning, yielding better efficiency and reconstruction fidelity. The improved ViT-VQGAN further improves vector-quantized image modeling tasks, including unconditional image generation, conditioned image generation (e.g., class-conditioned image generation), and unsupervised representation learning.

Abstract: Implementations relate to managing multimedia content that is obtained by large language model(s) (LLM(s)) and/or generated by other generative model(s). Processor(s) of a system can: receive natural language (NL) based input that requests multimedia content, generate a response that is responsive to the NL based input, and cause the response to be rendered. In some implementations, and in generating the response, the processor(s) can process, using a LLM, LLM input to generate LLM output, and determine, based on the LLM output, at least multimedia content to be included in the response. Further, the processor(s) can evaluate the multimedia content to determine whether it should be included in the response. In response to determining that the multimedia content should not be included in the response, the processor(s) can cause the response, including alternative multimedia content or other textual content, to be rendered.

Abstract: Various embodiments enable a robot, or other autonomous or semi-autonomous device or system, to receive data involving the performance of a task in the physical world. The data can be provided as input to a perception network to infer a set of percepts about the task, which can correspond to relationships between objects observed during the performance. The percepts can be provided as input to a plan generation network, which can infer a set of actions as part of a plan. Each action can correspond to one of the observed relationships. The plan can be reviewed and any corrections made, either manually or through another demonstration of the task. Once the plan is verified as correct, the plan (and any related data) can be provided as input to an execution network that can infer instructions to cause the robot, and/or another robot, to perform the task.

Abstract: The present invention provides a method of dynamical adjustment for manufacturing a thermally treated steel sheet.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a machine learning model using teacher annealing.

Abstract: A hinge module capable of fixing a first housing and a second at various opening angles and a foldable electronic device including the same are provided.

Abstract: A method and a system for integrating renewable power with a natural gas hydrogen production plant are provided. An exemplary method include generating electricity and a reformed hydrogen stream in a solid oxide fuel cell (SOFC) stack, and providing the electricity to an electrolyzer to generate an electrolysis hydrogen stream. A second stream of electricity is generated in a renewable energy facility, when available, and providing the second stream of electricity to the electrolyzer to increase the generation of the electrolysis hydrogen stream.

Abstract: An electronic device includes a processor configured to estimate green values at red and blue pixel locations of an input Bayer frame based on green values at green pixel locations of the input Bayer frame. The processor is also configured to generate red, green, and blue channels of joint demosaiced-warped output RGB pixels from the input Bayer frame based on the green values at the green pixel locations, the estimated green values at the red and blue pixel locations, an alignment vector map, and kernels for red, green, and blue pixels.

Abstract: Systems and methods are provided that train a machine-learned language encoding model through the use of a contrastive learning task. In particular, the present disclosure describes a contrastive learning task where the encoder learns to distinguish input tokens from plausible alternatives. In some implementations, on each training example the proposed method masks out some subset (e.g., 15%) of the original input tokens, replaces the masked tokens with samples from a “generator” (e.g., which may be a small masked language model), and then trains the encoder to predict whether each token comes from the original data or is a replacement produced by the generator.

Abstract: Implementations relate to generating multi-modal response(s) through utilization of large language model(s) (LLM(s)). Processor(s) of a system can: receive natural language (NL) based input, generate a multi-modal response that is responsive to the NL based output, and cause the multi-modal response to be rendered. In some implementations, and in generating the multi-modal response, the processor(s) can process, using a LLM, LLM input (e.g., that includes at least the NL based input) to generate LLM output, and determine, based on the LLM output, textual content for inclusion in the multi-modal response and multimedia content for inclusion in the multi-modal response. In some implementations, the multimedia content can be obtained based on a multimedia content tag that is included in the LLM output and that is indicative of the multimedia content. In various implementations, the multimedia content can be interleaved between segments of the textual content.

Abstract: An electronic device is provided. The electronic device includes a first housing, a second housing coupled to the first housing, a plurality of first adhesive portions arranged between the first housing and the second housing, and spaced apart from each other along a first perimeter portion of the first housing, and a plurality of second adhesive portions arranged between the plurality of first adhesive portions and arranged along a second perimeter portion of the first housing, wherein the first adhesive portions include an adhesive support portion extending in one direction and a first adhesive material arranged on one side or both sides of the adhesive support portion, and the second adhesive portions include a second adhesive material.

Abstract: In at least one embodiment, under the control of a robotic control system, a gripper on a robot is positioned to grasp a 3-dimensional object. In at least one embodiment, the relative position of the object and the gripper is determined, at least in part, by using a camera mounted on the gripper.

Abstract: Methods and systems for generation of shape data for a set of electronic designs include inputting a set of shape data, where the set of shape data represents a set of shapes for a device fabrication process. A convolutional neural network is used on the set of shape data to determine a set of generated shape data, where the convolutional neural network comprises a generator trained with a set of pre-determined discriminators. The set of generated shape data comprises a scanning electron microscope (SEM) image.

Abstract: A method of operation of a compute system includes: generating a skin segmentation including a non-skin region and a skin prediction based on a patient image; generating a body part segmentation based on the patient image; generating a cropped image based on the skin segmentation and the body part segmentation with the cropped image includes the non-skin region based on the skin prediction; generating a psoriasis segmentation based on the cropped image and the skin prediction; generating intermediate scores for erythema, induration, desquamation for the cropped image; and generating a full body PASI score based a psoriasis segmentation and the intermediate scores on the for displaying on a device to assist in diagnosis.

Abstract: The invention proposes False targets caused by reflection position determining system and solution for the SSR which helps determine reflected real target's coordinate, reflector's coordinate and features, coordinate and features of the false target (multipath target) and using these data as the basis to build up the multipath target suppressing system, guaranteeing the detecting ability of the radar. The proposed system contains: Input data generating component; False target position (multipath position) determining component; Reflected signal power comparison calculating component; Displaying and data transferring component; Reflected multipath targets suppressing component.

Abstract: Methods and systems for compressing shape data for a set of electronic designs include inputting a set of shape data, where the shape data comprises mask designs. A convolutional autoencoder encodes the set of shape data, where the encoding compresses the set of shape data to produce a set of encoded shape data. The convolutional autoencoder is tuned for increased accuracy of the set of encoded shape data based on design rules for the set of electronic designs. The convolutional autoencoder comprises a set of parameters comprising weights, and the convolutional autoencoder has been trained to retain important information needed, based on the design rules for the set of electronic designs.

Abstract: A processor includes a loop detection unit to detect a phantom-loop based on the resources reserved for execution of the phantom loop. The processor executes the phantom loop by reading source operand data on a first iteration of the loop and writing back data on the last iteration of the loop while allowing instructions after the loop to be concurrently executed.

Abstract: A processor includes a time counter, a vector coprocessor, and an extended vector register file for executing vector instructions and extending the data width of vector registers. The processor statically dispatches vector instructions with preset execution times based on a write time of a register in a coprocessor register scoreboard and a time counter provided to a vector execution pipeline.

Abstract: A processor includes a loop detection unit to detect an OOO-loop based on the reserved resources for execution of the OOO loop. The processor executes the OOO loop by reading source operand data on first iteration of the loop and write back data on the last iteration of the loop while allowing instructions after the loop to be concurrently executed.