Explore the 6,888 U.S. Patent Applications published on the 39th week of 2025, including 878 applications that subsequently received a Patent Grant.
WEARABLE DEVICE AND METHOD FOR DISPLAYING USER INTERFACE RELATED TO CONTROL OF EXTERNAL ELECTRONIC DEVICE
A wearable device can connect to an external electronic device and show its display. It tracks where the user is looking on its screen. When the user gazes near the external device, it shows a control screen for that device. This screen includes options related to different functions that can be controlled. By interacting with these options, the user can send commands to the external device.
Published: 2025-09-25 Assignee: INTEL CORPORATION.
DYNAMIC PRECISION FOR NEURAL NETWORK COMPUTE OPERATIONS
A new device has been created to improve how computers perform calculations for neural networks. It includes two parts: one that works with high precision and another that works with low precision. The device can choose which part to use based on the instructions it receives. Additionally, it can apply a special control (called a gate) to help carry out these instructions effectively. This design aims to make computing faster and more efficient.
Published: 2025-09-25 Assignee: INTERCOS S.P.A..
TRANSPARENT COLORED COSMETIC INGREDIENT FOR COSMETIC PRODUCTS, POSSIBLY CAPABLE OF AUTOMATICALLY CHANGING COLOR WHEN IRRADIATED BY ULTRAVIOLET LIGHT
A new cosmetic ingredient is made from a special type of clear polymer that lets light pass through. It includes a color-changing molecule that can either stay the same color or change colors quickly when exposed to UV light. This molecule is attached to a cosmetic oil that helps keep it in place. The oil is designed to work well with the color-changing molecule, ensuring it can switch from clear to colored when in sunlight. This innovation allows for unique cosmetic products that can change their appearance based on light exposure.
Published: 2025-09-25 Assignee: PlayAl GmbH.
ARTIFICIALLY INTELLIGENT AVATARS OF REAL-WORLD ATHLETES
An avatar can be created to represent a real-world athlete using artificial intelligence. This process involves first providing an avatar that looks and acts like the athlete. The athlete then gives feedback on how the avatar behaves. Based on this feedback, changes can be made to improve the avatar's actions or appearance. This system allows for a more accurate and responsive representation of the athlete in a virtual setting.
Published: 2025-09-25 Assignee: NVIDIA Corporation.
MAPPING ABSTRACT DATA MOVEMENTS INTO SEQUENTIAL AND PARALLEL DIRECT MEMORY ACCESS (DMA) PROGRAMMING
A new system helps make programming for data transfers easier and less error-prone. It uses high-level descriptions of how data should move, which are then turned into specific settings for direct memory access (DMA) devices. This means that users don't have to manually configure the hardware, saving time and effort. By automating this process, the chance of mistakes is reduced. Overall, it streamlines the way data is managed in computer systems.
Published: 2025-09-25 Assignee: THE BROAD INSTITUTE, INC..
ADENO-ASSOCIATED VIRAL VECTORS AND USES THEREOF
Adeno-associated viral vectors are special tools used to deliver genetic material into cells. They help scientists introduce new genes into cells, which can be useful for research and medical treatments. These vectors can effectively change how cells behave by adding or modifying their genetic information. The methods described show how to use these vectors to improve cell transduction, which is the process of getting genes into cells. Overall, this technology has the potential to advance gene therapy and other applications in medicine.
Published: 2025-09-25 Assignee: BeijingGiluntidePharmaceuticalCo.,Ltd..
TARGETED RADIOPHARMACEUTICAL FOR TUMOR AND ITS USE IN THE IMAGING-GUIDED COMBINATION THERAPY OF TARGETED RADIOTHERAPY AND IMMUNOTHERAPY
A new type of medicine combines a special protein with a radioactive element to help diagnose or treat certain tumors that have a specific marker called integrin αvβ3. This medicine can also include an immune therapy drug and a tiny imaging probe to track the treatment's effects. Using a PD-L1 blocker after the targeted radiation therapy can enhance the overall effectiveness of the treatment. Additionally, a nanoantibody probe can be used to see how the tumor reacts to the radiation therapy by checking for PD-1 or PD-L1 markers. This approach aims to improve both imaging and treatment for tumors.
Published: 2025-09-25 Assignee: BriefCam Ltd..
METHOD AND SYSTEM FOR QUERYING A VIDEO BY INCORPORATING VIDEO ANALYTICS AND LARGE LANGUAGE MODELS
A system allows users to ask questions about specific objects in a video. It starts by analyzing the video to identify important features of those objects. When a user asks a question, the system selects the relevant objects from the video. It then creates a prompt that combines the user's question and the analyzed features. Finally, this prompt is sent to a large language model, which provides an answer based on the information gathered.
Published: 2025-09-25 Assignee: Paul Gallagher.
Wireless dropper post actuator system
A wireless dropper post actuator system for bicycles allows riders to easily adjust the height of their bike seat without messy cables. It features a small and light motor that includes a battery and a computer chip for control. This system uses wireless signals, making installation simpler and reducing the need for maintenance. Riders can charge the system easily and program it to remember their preferred seat heights. It works with both regular and electric bikes, improving how dropper posts operate.
Published: 2025-09-25 Assignee: ECORATINGS SOFTWARE SOLUTIONS PRIVATE LIMITED.
FOUNDATION GENERATIVE ARTIFICIAL INTELLIGENCE (AI) MODEL WITH TRANSFORMER ARCHITECTURE FOR ENVIRONMENTAL, SOCIAL, AND GOVERNANCE (ESG) IMPACT
A new AI model has been created specifically for analyzing environmental, social, and governance (ESG) issues. It uses a Transformer architecture with about 30 billion parameters, allowing it to handle very long texts, like sustainability reports. The model can understand both text and images by using advanced techniques that improve its comprehension of different types of information. It also employs a special learning method that helps improve its responses based on comparisons with other generated options. Overall, this AI model aims to provide better insights and reasoning for ESG-related topics.