Patents by Inventor Fuxi CAI
Fuxi CAI has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11790989Abstract: A method for setting memory elements in a plurality of states includes applying a set signal to a memory element to transition the memory element from a low-current state to a high-current state; applying a partial reset signal to the memory element to transition the memory element from the high-current state to a state between the high-current state and the low-current state; determining whether the state corresponds to a predetermined state; and applying one or more additional partial reset signals to the memory element until the state corresponds to the predetermined current state. The memory element may be coupled in series with a transistor, and a voltage control circuit may apply voltages to the transistor to set and partially reset the memory element.Type: GrantFiled: May 24, 2021Date of Patent: October 17, 2023Assignee: Applied Materials, Inc.Inventors: Deepak Kamalanathan, Siddarth Krishnan, Archana Kumar, Fuxi Cai, Federico Nardi
-
Patent number: 11127458Abstract: A method of setting multi-state memory elements into at least one low-power state may include receiving a command to cause a memory element to transition into one of three or more states; applying a first signal to the memory element to transition the memory element into the one of the three or more states, where the three or more states are evenly spaced in a portion of an operating range of the memory element; receiving a command to cause a memory element to transition into a low-power state; applying a second signal to the memory element to transition the memory element into the low-power state, where the low-power state is outside of the portion of the operating range of the memory element by an amount greater than a space between each of the three or more states.Type: GrantFiled: April 28, 2020Date of Patent: September 21, 2021Assignee: Applied Materials, Inc.Inventors: Deepak Kamalanathan, Siddarth Krishnan, Fuxi Cai, Christophe J Chevallier
-
Publication number: 20210280247Abstract: A method for setting memory elements in a plurality of states includes applying a set signal to a memory element to transition the memory element from a low-current state to a high-current state; applying a partial reset signal to the memory element to transition the memory element from the high-current state to a state between the high-current state and the low-current state; determining whether the state corresponds to a predetermined state; and applying one or more additional partial reset signals to the memory element until the state corresponds to the predetermined current state. The memory element may be coupled in series with a transistor, and a voltage control circuit may apply voltages to the transistor to set and partially reset the memory element.Type: ApplicationFiled: May 24, 2021Publication date: September 9, 2021Applicant: Applied Materials, Inc.Inventors: Deepak Kamalanathan, Siddarth Krishnan, Archana Kumar, Fuxi Cai, Federico Nardi
-
Patent number: 11017856Abstract: A method for setting memory elements in a plurality of states includes applying a set signal to a memory element to transition the memory element from a low-current state to a high-current state; applying a partial reset signal to the memory element to transition the memory element from the high-current state to a state between the high-current state and the low-current state; determining whether the state corresponds to a predetermined state; and applying one or more additional partial reset signals to the memory element until the state corresponds to the predetermined current state. The memory element may be coupled in series with a transistor, and a voltage control circuit may apply voltages to the transistor to set and partially reset the memory element.Type: GrantFiled: February 18, 2020Date of Patent: May 25, 2021Assignee: Applied Materials, Inc.Inventors: Deepak Kamalanathan, Siddarth Krishnan, Archana Kumar, Fuxi Cai, Federico Nardi
-
Patent number: 10812083Abstract: Sparse representation of information performs powerful feature extraction on high-dimensional data and is of interest for applications in signal processing, machine vision, object recognition, and neurobiology. Sparse coding is a mechanism by which biological neural systems can efficiently process complex sensory data while consuming very little power. Sparse coding algorithms in a bio-inspired approach can be implemented in a crossbar array of memristors (resistive memory devices). This network enables efficient implementation of pattern matching and lateral neuron inhibition, allowing input data to be sparsely encoded using neuron activities and stored dictionary elements. The reconstructed input can be obtained by performing a backward pass through the same crossbar matrix using the neuron activity vector as input. Different dictionary sets can be trained and stored in the same system, depending on the nature of the input signals.Type: GrantFiled: October 30, 2019Date of Patent: October 20, 2020Assignee: The Regents of the University of MichiganInventors: Wei Lu, Fuxi Cai, Patrick Sheridan, Chao Du
-
Publication number: 20200067512Abstract: Sparse representation of information performs powerful feature extraction on high-dimensional data and is of interest for applications in signal processing, machine vision, object recognition, and neurobiology. Sparse coding is a mechanism by which biological neural systems can efficiently process complex sensory data while consuming very little power. Sparse coding algorithms in a bio-inspired approach can be implemented in a crossbar array of memristors (resistive memory devices). This network enables efficient implementation of pattern matching and lateral neuron inhibition, allowing input data to be sparsely encoded using neuron activities and stored dictionary elements. The reconstructed input can be obtained by performing a backward pass through the same crossbar matrix using the neuron activity vector as input. Different dictionary sets can be trained and stored in the same system, depending on the nature of the input signals.Type: ApplicationFiled: October 30, 2019Publication date: February 27, 2020Inventors: Wei LU, Fuxi CAI, Patrick SHERIDAN, Chao DU
-
Patent number: 10498341Abstract: Sparse representation of information performs powerful feature extraction on high-dimensional data and is of interest for applications in signal processing, machine vision, object recognition, and neurobiology. Sparse coding is a mechanism by which biological neural systems can efficiently process complex sensory data while consuming very little power. Sparse coding algorithms in a bio-inspired approach can be implemented in a crossbar array of memristors (resistive memory devices). This network enables efficient implementation of pattern matching and lateral neuron inhibition, allowing input data to be sparsely encoded using neuron activities and stored dictionary elements. The reconstructed input can be obtained by performing a backward pass through the same crossbar matrix using the neuron activity vector as input. Different dictionary sets can be trained and stored in the same system, depending on the nature of the input signals.Type: GrantFiled: December 28, 2018Date of Patent: December 3, 2019Assignee: The Regents of the University of MichiganInventors: Wei Lu, Fuxi Cai, Patrick Sheridan, Chao Du
-
Publication number: 20190158097Abstract: Sparse representation of information performs powerful feature extraction on high-dimensional data and is of interest for applications in signal processing, machine vision, object recognition, and neurobiology. Sparse coding is a mechanism by which biological neural systems can efficiently process complex sensory data while consuming very little power. Sparse coding algorithms in a bio-inspired approach can be implemented in a crossbar array of memristors (resistive memory devices). This network enables efficient implementation of pattern matching and lateral neuron inhibition, allowing input data to be sparsely encoded using neuron activities and stored dictionary elements. The reconstructed input can be obtained by performing a backward pass through the same crossbar matrix using the neuron activity vector as input. Different dictionary sets can be trained and stored in the same system, depending on the nature of the input signals.Type: ApplicationFiled: December 28, 2018Publication date: May 23, 2019Inventors: Wei LU, Fuxi CAI, Patrick SHERIDAN, Chao DU
-
Patent number: 10171084Abstract: Sparse representation of information performs powerful feature extraction on high-dimensional data and is of interest for applications in signal processing, machine vision, object recognition, and neurobiology. Sparse coding is a mechanism by which biological neural systems can efficiently process complex sensory data while consuming very little power. Sparse coding algorithms in a bio-inspired approach can be implemented in a crossbar array of memristors (resistive memory devices). This network enables efficient implementation of pattern matching and lateral neuron inhibition, allowing input data to be sparsely encoded using neuron activities and stored dictionary elements. The reconstructed input can be obtained by performing a backward pass through the same crossbar matrix using the neuron activity vector as input. Different dictionary sets can be trained and stored in the same system, depending on the nature of the input signals.Type: GrantFiled: April 24, 2018Date of Patent: January 1, 2019Assignee: The Regents of The University of MichiganInventors: Wei Lu, Fuxi Cai, Patrick Sheridan, Chao Du
-
Publication number: 20180309451Abstract: Sparse representation of information performs powerful feature extraction on high-dimensional data and is of interest for applications in signal processing, machine vision, object recognition, and neurobiology. Sparse coding is a mechanism by which biological neural systems can efficiently process complex sensory data while consuming very little power. Sparse coding algorithms in a bio-inspired approach can be implemented in a crossbar array of memristors (resistive memory devices). This network enables efficient implementation of pattern matching and lateral neuron inhibition, allowing input data to be sparsely encoded using neuron activities and stored dictionary elements. The reconstructed input can be obtained by performing a backward pass through the same crossbar matrix using the neuron activity vector as input. Different dictionary sets can be trained and stored in the same system, depending on the nature of the input signals.Type: ApplicationFiled: April 24, 2018Publication date: October 25, 2018Inventors: Wei LU, Fuxi CAI, Patrick SHERIDAN, Chao DU