查看更多>>摘要：No assignee for this patent application has been made.News editors obtained the following quote from the background information suppli ed by the inventors:““Technical Field“The present disclosure generally relates to computing hardware, and more partic ularly, to acceleratinginference performance on artificial intelligence acceler ators.“Description of the Related Art“Artificial intelligence (AI) commonly uses deep learning models to construct ou tcomes. Deep learningmodels may be represented by a computational graph that in cludes nodes (representing computer operations;for example, convolutions or lon g-short term memory operations) and edges (representing dataflow). Accelerators can be used in a deep learning compiler to accelerate AI inference in operation s.
查看更多>>摘要：No assignee for this patent application has been made.News editors obtained the following quote from the background information suppli ed by the inventors:““Technical Field“The present disclosure generally relates to computing hardware, and more partic ularly, to acceleratinginference performance on artificial intelligence acceler ators.“Description of the Related Art“Artificial intelligence (AI) commonly uses deep learning models to construct ou tcomes. Deep learningmodels may be represented by a computational graph that in cludes nodes (representing computer operations;for example, convolutions or lon g-short term memory operations) and edges (representing dataflow). Accelerators can be used in a deep learning compiler to accelerate AI inference in operation s.
查看更多>>摘要：The assignee for this patent, patent number 12145258, is Berkshire Grey Operatin g Company Inc.(Bedford, Massachusetts, United States).Reporters obtained the following quote from the background information supplied by the inventors: “The invention generally relates to robotics, and relates in p articular to robotic control systems that aredesigned to accommodate a wide var iety of unexpected conditions and loads.“Most industrial robotic systems operate in a top-down manner, generally as foll ows: a controllersamples a variety of sensors, and then logic on that same cont roller computes whether or not to takeaction. The benefit of this logic flow (u sually referred to as “polling”) is that all of the control logic is inthe same place. The disadvantage is that in practical robotic systems, the signals are o ften sampled quiteslowly. Also, all sensors must be wired to the control cabine t leading to long and error-prone cable runs.
查看更多>>摘要：The assignee for this patent, patent number 12145258, is Berkshire Grey Operatin g Company Inc.(Bedford, Massachusetts, United States).Reporters obtained the following quote from the background information supplied by the inventors: “The invention generally relates to robotics, and relates in p articular to robotic control systems that aredesigned to accommodate a wide var iety of unexpected conditions and loads.“Most industrial robotic systems operate in a top-down manner, generally as foll ows: a controllersamples a variety of sensors, and then logic on that same cont roller computes whether or not to takeaction. The benefit of this logic flow (u sually referred to as “polling”) is that all of the control logic is inthe same place. The disadvantage is that in practical robotic systems, the signals are o ften sampled quiteslowly. Also, all sensors must be wired to the control cabine t leading to long and error-prone cable runs.
查看更多>>摘要：This patent application has not been assigned to a company or institution.The following quote was obtained by the news editors from the background informa tion supplied bythe inventors: “Decentralized learning of machine learning (ML) model(s) is an increasingly popular MLtechnique for updating ML model(s) due t o various privacy considerations. In one common implementationof decentralized learning, an on-device ML model is stored locally on a client device of a user, and a globalML model, that is a cloud-based counterpart of the on-device ML mod el, is stored remotely at a remotesystem (e.g., a server or cluster of servers) . During a given round of decentralized learning, the clientdevice can check-in to a population of client devices that will be utilized in the given round of d ecentralizedlearning, download a global ML model or weights thereof from the re mote system (e.g., to be utilized asthe on-device ML model), generate an update for the weight of the global ML model based on processinginstance(s) of client data locally at the client device and using the on-device ML model, and upload theupdate for the weight of the global ML model back to the remote system and w ithout transmitting theinstance(s) of the client device. The remote system can utilize the update received from the client device,and additional updates gener ated in a similar manner at additional client devices and that are receivedfrom the additional client devices, to update the weights of the global ML model.
查看更多>>摘要：This patent application has not been assigned to a company or institution.The following quote was obtained by the news editors from the background informa tion supplied bythe inventors: “Decentralized learning of machine learning (ML) model(s) is an increasingly popular MLtechnique for updating ML model(s) due t o various privacy considerations. In one common implementationof decentralized learning, an on-device ML model is stored locally on a client device of a user, and a globalML model, that is a cloud-based counterpart of the on-device ML mod el, is stored remotely at a remotesystem (e.g., a server or cluster of servers) . During a given round of decentralized learning, the clientdevice can check-in to a population of client devices that will be utilized in the given round of d ecentralizedlearning, download a global ML model or weights thereof from the re mote system (e.g., to be utilized asthe on-device ML model), generate an update for the weight of the global ML model based on processinginstance(s) of client data locally at the client device and using the on-device ML model, and upload theupdate for the weight of the global ML model back to the remote system and w ithout transmitting theinstance(s) of the client device. The remote system can utilize the update received from the client device,and additional updates gener ated in a similar manner at additional client devices and that are receivedfrom the additional client devices, to update the weights of the global ML model.
查看更多>>摘要：No assignee for this patent application has been made.News editors obtained the following quote from the background information suppli ed by the inventors:“The present disclosure relates to fixturing in manufacturi ng, and, more specifically, to digital twin analysisfor fixture design by addit ive manufacturing.
查看更多>>摘要：No assignee for this patent application has been made.News editors obtained the following quote from the background information suppli ed by the inventors:“The present disclosure relates to fixturing in manufacturi ng, and, more specifically, to digital twin analysisfor fixture design by addit ive manufacturing.
查看更多>>摘要：The patent’s assignee is Ocado Innovation Limited (Hatfield, Hertfordshire, Unit ed Kingdom).News editors obtained the following quote from the background information suppli ed by the inventors:“An important aspect when using a robotic manipulator durin g a picking and packing process is beingable to determine the state of an objec t or item being handled. This is done through the use of sensorsthat detect the presence of the object, interaction forces, contact properties, etc. A known ap proach isto apply sensors directly on the finger elements of the robotic manipu lator as this is where the intendedinteraction occurs between the manipulator a nd the object being handled.
查看更多>>摘要：The patent’s assignee is Ocado Innovation Limited (Hatfield, Hertfordshire, Unit ed Kingdom).News editors obtained the following quote from the background information suppli ed by the inventors:“An important aspect when using a robotic manipulator durin g a picking and packing process is beingable to determine the state of an objec t or item being handled. This is done through the use of sensorsthat detect the presence of the object, interaction forces, contact properties, etc. A known ap proach isto apply sensors directly on the finger elements of the robotic manipu lator as this is where the intendedinteraction occurs between the manipulator a nd the object being handled.
NETL
NSTL