Автор: Академик Олег Фиговский
В период с 2014 по 2023 год в Китае было зарегистрировано более 38 тысяч изобретений в области генеративного ИИ – в шесть раз больше, чем в США, занимающих второе место по количеству патентов. В Индии, которая находится на пятом месте, наблюдался самый высокий среднегодовой темп роста показателя среди пяти лидеров – 56 процентов. Из Китая подают наибольшее число заявок на патенты в области генеративного искусственного интеллекта, значительно опережая другие страны в пятерке лидеров – США, Южную Корею, Японию и Индию. Об этом говорится в новом докладе Всемирной организации интеллектуальной собственности (ВОИС). Авторы доклада ВОИС о патентном ландшафте в области генеративного ИИ сообщают о 54 тысячах изобретений, зарегистрированных в период с 2014 по 2023 год, причем более 25 процентов из них были – за последний год. С 2017 года число патентов увеличилось в восемь раз. Генеративный ИИ позволяет пользователям создавать разнообразный контент, включая тексты, изображения, музыку и компьютерный код. На его основе работает ряд сервисов, в том чат-боты ChatGPT, Google Gemini и ERNIE. Как отмечают специалисты ВОИС, генеративный ИИ уже применяется в таких отраслях, как медико-биологические науки, промышленное производство, транспорт, безопасность и телекоммуникации.
В период с 2014 по 2023 год в Китае было зарегистрировано более 38 тысяч изобретений в области генеративного ИИ – в шесть раз больше, чем в США, занимающих второе место по количеству патентов. В Индии, которая находится на пятом месте, наблюдался самый высокий среднегодовой темп роста показателя среди пяти лидеров – 56 процентов. «Генеративный ИИ – это революционная технология, способная изменить то, как мы живем, работаем и развлекаемся. Анализируя данные по патентованию и тенденции в этой области, ВОИС стремится улучшить понимание того, где разрабатываются эти быстро развивающиеся технологии и в каком направлении они движутся», – отметил Генеральный директор ВОИС Дарен Танг. «Это поможет органам, занимающимся разработкой политики, направить развитие генеративного ИИ на всеобщее благо и гарантировать, чтобы в центре инновационных и творческих экосистем по-прежнему находились интересы человечества», – добавил он.
Согласно докладу, среди патентов на технологии в сфере генеративного ИИ преобладают такие области применения, как изображения и видео, за ними следуют текст и звук (речь и музыка). Стремительно увеличивается число патентов на технологии, использующие данные о молекулах, генах и белках: среднегодовые темпы роста за последние пять лет составили 78 процентов. В десятку ведущих компаний, подающих заявки на получение патентов в области генеративного ИИ, входят Tencent (2074 изобретения), Ping An Insurance (1564 изобретения), Baidu (1234), Китайская академия наук (607), IBM (601), Alibaba Group (571), Samsung Electronics (468), Alphabet (443), ByteDance (418) и Microsoft (377).
Ниже приводим новые патенты Китая в этой области и прежде всего и медицинской ткхнике:
A geomagnetic smart parking system (patent TW202331636) includes a geomagnetic equipment, a smart parking equipment, a central server and a management center server. The geomagnetic equipment includes a first communication module, a second communication module and a detection module. The smart parking equipment includes a third wireless communication module. When the detection module detects that a vehicle enters the parking space area within a preset time range, the first communication module or the second communication module transmits the vehicle detection signal to the third communication module, the smart parking equipment enters the start state from the standby state. The central server receives a vehicle detection signal sent from the smart parking device. The management center server generates a parking payment signal according to the vehicle detection signal. The invention adopts the dual-network dual-mode technology of the smart parking system, which is low power consumption and high recognition rate.
The invention CN106462124 provides a method for identifying and controlling household appliances based on intention, and the method comprises the steps: acquiring multi-modal information of a user; identifying user intention according to the multi-modal information; and controlling the household appliances through combining with context record and variable parameters according to the multi-modal information of the user and the user intention. Then the user intention can be identified through one or more of multi-modal information such as voice of the user, expression of the user, movement of user and the like, the household appliances can be controlled through combining with the context record and the variable parameters, so that the automatic adjustment of the household appliances can be more intelligentized, according to the invention, artificial intelligence can be applied to smart home, the household appliances can be more conveniently, fastly and accurately controlled, the daily life of people can be more convenient, the enjoyment and interactivity of life can be improved, the life can be more splendid, the robot can be more anthropomorphic, and the user experience of the artificial intellegence can be improved on smart home aspect
The present application WO2024146486 discloses a blood glucose management method and a related electronic device. According to the method, a meal time of a dietary event of a user can be determined according to a blood glucose value of the user, and when a blood glucose curve drawn according to blood glucose values of the user is displayed, the meal time and/or the blood glucose value associated with the meal time can be displayed on the blood glucose curve. In this way, the dietary event of the user can be associated with the blood glucose of the user, and a blood glucose change condition before and after the user takes a meal is highlighted, so that the user can more visually know the effect of diet on blood glucose, and then the user can manage and control blood glucose starting with the diet.
First page clipping of CN221285740 (U)
An application apparatus (1000) by patent CN221285740 for applying a medical instrument (800) to a host, comprising a housing (10), an auxiliary mechanism (20), and a first driving mechanism (30). The housing (10) comprises a first holding portion (130), and a proximal end close to the host and a distal end away from the host during operation. The auxiliary mechanism (20) comprises a movement main body (200) releasably retained on the first holding portion (130) and configured to be movable relative to the housing (10) when released, a storage part (250) arranged on the movement main body (200) and configured to be combinable with the medical instrument (800), and a puncture component (260) arranged on the movement main body (200). The first driving mechanism (30) is configured to apply an effect to the movement main body (200) in a manner of facing the proximal end, and when the movement main body (200) is released, the first driving mechanism (30) drives the movement main body (200) to move towards the proximal end so as to push the medical instrument (800) combined with the storage part (250) to the body surface of the host and at least partially place the medical instrument under the skin of the host.
A data processing method and apparatus, and a device and a computer-readable storage medium (WO2024140869). In the method, an electronic device can acquire object basic information of an object, wherein the object basic information can be information related to a physiological index of the object; a measurement policy of the physiological index of the object can be determined according to the object basic information; at least one piece of physiological index measurement data of the object can be acquired according to the measurement policy; and a physiological index measurement report of the object can be generated by analyzing the at least one piece of physiological index measurement data of the object. By means of the present application, according to object basic information of different objects, measurement policies of physiological indexes of the objects can be individually determined, which is beneficial to more effectively measuring the physiological indexes of the objects and thus is conducive to improving the accuracy and reliability of physiological index measurement data.
A method by patent WO2024139588 for collecting and quantifying tongue manifestation features. The method comprises: first, automatically obtaining a high-quality tongue manifestation image that meets the requirements of traditional Chinese medicine tongue diagnosis; then, identifying from the image a tongue part of a patient, and acquiring an edge shape of a tongue surface, so as to obtain a tongue surface inspection area; performing color correction on the image by using the tongue surface inspection area, so as to subsequently perform, in the tongue surface inspection area, color and texture recognition on the tongue part; and further quantifying tongue color and tongue coating features. During remote consultation, traditional Chinese medicine physicians and/or traditional Chinese medicine practitioners can still perform tongue diagnosis, and it is possible to avoid misdiagnosis in remote diagnosis and treatment that is caused by differences in photography skills and photography device quality of patients, an insufficient capability for determining, when tongue manifestations are uploaded, whether the tongue manifestations meet the standards of traditional Chinese medicine diagnosis and treatment, etc.
Patent WO2024140073 provided are a multi-channel pulse high-voltage parameter-controllable shock wave lithotripsy balloon imaging system and a catheter thereof. An intravascular imaging catheter (03A) is arranged in a shock wave lithotripsy balloon (02), and a plurality of high-coverage electrode pairs (01A, 01B) capable of releasing shock wave energy are arranged in a working area of the shock wave lithotripsy balloon (02). For a vascular calcified tissue, a shock wave treatment function is integrated into an intravascular tomography system, such that a doctor can perform effective treatment and evaluation ‘before, in and after operation’, and in combination with an AI intelligent calcification evaluation software of the intravascular tomography system, can effectively identify a calcification category and perform accurate shock wave treatment for deep calcification, thereby shortening the treatment time, reducing the learning curve of the doctor, and improving the surgical treatment effect.
First page clipping of WO2024140132 (A1)
Embodiments of the present application WO2024140132 provide a wearable device and a health monitoring method. The wearable device comprises a watch body, an air pump being arranged in the watch body; a watchband assembly connected to the watch body; and an airbag assembly, comprising an airbag and an air guide part. The airbag is arranged on the side of the watchband assembly that is in contact with a user, and both ends of the air guide part are connected to the airbag and the air pump, respectively, wherein the air guide part can stretch out and draw back so as to adjust the position of the airbag relative to the watchband assembly. According to the wearable device provided in the embodiments of the present application, the position of the airbag can be adjusted, such that the wearable device can be suitable for users with different wrist circumferences. Meanwhile, the part where the user needs to be subjected to health monitoring can be covered, the obtained health monitoring result is relatively accurate, and a health monitoring effect can be improved.
First page clipping of WO2024139873 (A1)
A fully flexible electrode, a preparation method therefor, and use thereof by patent WO2024139873. The fully flexible electrode comprises filamentous PEDOT:PSS and a hydrogel layer wrapped around the filamentous PEDOT:PSS. The hydrogel layer comprises any one of acrylamide, k2-type carrageenan, agar, and polyvinyl alcohol. As the fully flexible electrode has a hydrogel layer wrapped around the filamentous PEDOT:PSS, the synergy between the two makes the electrode have good conductivity, high toughness, and tensile strength, which enable the electrode to withstand a relatively great amount of deformation. In addition, the degree of elastic mismatch is relatively small, and the biocompatibility is also better. The electrode can be effectively applied to an implantable detection process to obtain data with a better signal-to-noise ratio, thereby increasing the survival rate and survival time of animals undergoing implantation.
First page clipping of US2024221482 (A1)
A system, method, and smart wearable device by patent US2024221482 for user posture monitoring are provided, which are related to the technical field of data processing. The system comprises a smart wearable device and a monitoring terminal. The smart wearable device is equipped with motion sensor(s) and a wireless communication device. The smart wearable obtains real-time posture data of the user’s head through the motion sensor(s) and transmits the posture data to the monitoring terminal using the wireless communication device. The monitoring terminal receives the posture data, monitors whether the user suffers from a head disease and/or a neck-shoulder disease based on the posture data, and outputs a reminder message when the user is identified as having the head disease and/or the neck-shoulder disease. This application enables the analysis and provision of reminders for the user’s head disease and/or neck-shoulder disease based on monitoring the posture of the user’s head.
First page clipping of WO2024138665 (A1)
A mobile monitoring device and a method for triggering an exception event (WO2024138665). The mobile monitoring device is integrated with a motion sensor (20) for collecting motion data of a target object, and then identifying an activity mode of the target object according to the motion data. When the activity mode belongs to an exception activity mode, an exception event of physiological parameters is automatically recorded, and the exception event can be used as a basis for subsequent health state analysis. Compared with the recording of exception events triggered by a manual active control key, the mobile monitoring device is more suitable for animal scenarios, thus solving the problem that animals cannot actively record exception events in the field of mobile monitoring.
Model training method, signal recognition method, apparatus, computing and processing device, computer program, and computer-readable medium by panent be patent US2024188895. The model training method comprises: acquiring a training sample set, training sample set includes sample electrocardio-signals and abnormal labels of sample electrocardio-signals, and abnormal labels include a target abnormal labels and at least one related abnormal labels; inputting sample electrocardio-signals into multi-task model, training multi-task model based on a multi-task learning mechanism according to an output of multi-task model and the abnormal labels; multi-task model includes a target task model and at least one related task model, a target output of the target task model is target abnormality labels of inputted sample electrocardio-signals, and a target output of related task model is related abnormal labels of inputted sample electrocardio-signals; determining target task model after trained as target-abnormality-recognition model, and target-abnormality-recognition model is configured for recognizing target abnormality in the electrocardio-signals inputted into target-abnormality-recognition model.
Китай будет всесторонне содействовать эффективному функционированию системы открытых патентных лицензий, говорится в циркуляре, опубликованном Государственным управлением по делам интеллектуальной собственности Китая /ГУИС/. В соответствии с документом, также будут расширены способы и каналы применения патентов и подачи заявок. В документе содержится призыв к совместным усилиям по внедрению системы открытых лицензий в патентном праве. Данная мера направлена на то, чтобы патентообладатели стандартизировали подачу деклараций об открытых лицензиях и обоснованно оценивали лицензионные сборы.
Она также требует усиления открытого обмена информацией об открытых лицензионных заявлениях, укрепления связи между спросом и предложением на патенты и обеспечения полной реализации ролей всех сторон в секторе.
Открытая лицензия на патент – это специальная патентная лицензия, разработанная для повышения эффективности коммерциализации и применения патентов. Она будет способствовать превращению научно-технических достижений в производительные силы нового качества, говорится в циркуляре.
В соответствии с системой открытых патентных лицензий патентообладатель может добровольно подать заявление об открытии лицензии на патент и указать размер лицензионного сбора, а орган, управляющий патентами, опубликует эту информацию. Затем любое юридическое или физическое лицо может получить лицензию на использование патента, уведомив патентообладателя в письменной форме и оплатив пошлину, заявили в ГУИС. Согласно статистике ведомства, к концу 2023 года число действующих патентов на изобретения в Китае превысило 4,99 млн.
Иллюстрация: Digital
