Patents by Inventor Azalia Mirhoseini
Azalia Mirhoseini 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).
-
Publication number: 20220108058Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip placement. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip placement, comprising placing a respective macro node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the macro node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: ApplicationFiled: December 17, 2021Publication date: April 7, 2022Inventors: Anna Darling Goldie, Azalia Mirhoseini, Ebrahim Songhori, Wenjie Jiang, Shen Wang, Roger David Carpenter, Young-Joon Lee, Mustafa Nazim Yazgan, Chian-min Richard Ho, Quoc V. Le, James Laudon, Jeffrey Adgate Dean, Kavya Srinivasa Setty, Omkar Pathak
-
Publication number: 20220043951Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip floorplan. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip floorplan, comprising placing a respective node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: ApplicationFiled: August 23, 2021Publication date: February 10, 2022Inventors: Chian-min Richard Ho, William Hang, Mustafa Nazim Yazgan, Anna Darling Goldie, Jeffrey Adgate Dean, Azalia Mirhoseini, Emre Tuncer, Ya Wang, Anand Babu
-
Patent number: 11216609Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip placement. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip placement, comprising placing a respective macro node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the macro node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: GrantFiled: April 22, 2021Date of Patent: January 4, 2022Assignee: Google LLCInventors: Anna Darling Goldie, Azalia Mirhoseini, Ebrahim Songhori, Wenjie Jiang, Shen Wang, Roger David Carpenter, Young-Joon Lee, Mustafa Nazim Yazgan, Chian-Min Richard Ho, Quoc V. Le, James Laudon, Jeffrey Adgate Dean, Kavya Srinivasa Setty, Omkar Pathak
-
Publication number: 20210334445Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip placement. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip placement, comprising placing a respective macro node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the macro node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: ApplicationFiled: April 22, 2021Publication date: October 28, 2021Inventors: Anna Darling Goldie, Azalia Mirhoseini, Ebrahim Songhori, Wenjie Jiang, Shen Wang, Roger David Carpenter, Young-Joon Lee, Mustafa Nazim Yazgan, Chian-min Richard Ho, Quoc V. Le, James Laudon, Jeffrey Adgate Dean, Kavya Srinivasa Setty, Omkar Pathak
-
Publication number: 20210295166Abstract: A system may include a processor and a memory. The memory may include program code that provides operations when executed by the processor. The operations may include: partitioning, based at least on a resource constraint of a platform, a global machine learning model into a plurality of local machine learning models; transforming training data to at least conform to the resource constraint of the platform; and training the global machine learning model by at least processing, at the platform, the transformed training data with a first of the plurality of local machine learning models.Type: ApplicationFiled: February 6, 2017Publication date: September 23, 2021Applicant: WILLIAM MARSH RICE UNIVERSITYInventors: Bita Darvish Rouhani, Azalia Mirhoseini, Farinaz Koushanfar
-
Patent number: 11100266Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip floorplan. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip floorplan, comprising placing a respective node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: GrantFiled: June 1, 2020Date of Patent: August 24, 2021Assignee: Google LLCInventors: Chian-min Richard Ho, William Hang, Mustafa Nazim Yazgan, Anna Darling Goldie, Jeffrey Adgate Dean, Azalia Mirhoseini, Emre Tuncer, Ya Wang, Anand Babu
-
Publication number: 20210248445Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for optimizing the execution of the operations of a neural network. One of the methods includes obtaining data representing a graph characterizing a plurality of operations of a neural network, wherein each node of the graph characterizes an operation of the neural network and each edge of the graph characterizes data dependency between the operations; processing the data representing the graph using a graph embedding neural network to generate an embedding of the graph; and processing the embedding of the graph using a policy neural network to generate a task output, wherein the task output comprises, for each of the plurality of operations of the neural network, a respective decision for a particular optimization task.Type: ApplicationFiled: April 3, 2020Publication date: August 12, 2021Inventors: Yanqi Zhou, Sudip Roy, Amirali Abdolrashidi, Daniel Lin-Kit Wong, Chao Ma, Qiumin Xu, Azalia Mirhoseini
-
Publication number: 20200364389Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip floorplan. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip floorplan, comprising placing a respective node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: ApplicationFiled: June 1, 2020Publication date: November 19, 2020Inventors: Chian-min Richard Ho, William Hang, Mustafa Nazim Yazgan, Anna Darling Goldie, Jeffrey Adgate Dean, Azalia Mirhoseini, Emre Tuncer, Ya Wang, Anand Babu
-
Publication number: 20200279150Abstract: A system includes a neural network that includes a Mixture of Experts (MoE) subnetwork between a first neural network layer and a second neural network layer. The MoE subnetwork includes multiple expert neural networks. Each expert neural network is configured to process a first layer output generated by the first neural network layer to generate a respective expert output. The MoE subnetwork further includes a gating subsystem that selects, based on the first layer output, one or more of the expert neural networks and determine a respective weight for each selected expert neural network, provides the first layer output as input to each of the selected expert neural networks, combines the expert outputs generated by the selected expert neural networks in accordance with the weights for the selected expert neural networks to generate an MoE output, and provides the MoE output as input to the second neural network layer.Type: ApplicationFiled: May 20, 2020Publication date: September 3, 2020Inventors: Noam M. Shazeer, Azalia Mirhoseini, Krzysztof Stanislaw Maziarz
-
Publication number: 20200279163Abstract: A method for determining a placement for machine learning model operations across multiple hardware devices is described.Type: ApplicationFiled: May 20, 2020Publication date: September 3, 2020Inventors: Samuel Bengio, Mohammad Norouzi, Benoit Steiner, Jeffrey Adgate Dean, Hieu Hy Pham, Azalia Mirhoseini, Quoc V. Le, Naveen Kumar, Yuefeng Zhou, Rasmus Munk Larsen
-
Patent number: 10719761Abstract: A system includes a neural network that includes a Mixture of Experts (MoE) subnetwork between a first neural network layer and a second neural network layer. The MoE subnetwork includes multiple expert neural networks. Each expert neural network is configured to process a first layer output generated by the first neural network layer to generate a respective expert output. The MoE subnetwork further includes a gating subsystem that selects, based on the first layer output, one or more of the expert neural networks and determine a respective weight for each selected expert neural network, provides the first layer output as input to each of the selected expert neural networks, combines the expert outputs generated by the selected expert neural networks in accordance with the weights for the selected expert neural networks to generate an MoE output, and provides the MoE output as input to the second neural network layer.Type: GrantFiled: April 24, 2019Date of Patent: July 21, 2020Assignee: Google LLCInventors: Noam M. Shazeer, Azalia Mirhoseini, Krzysztof Stanislaw Maziarz
-
Patent number: 10699043Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip floorplan. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip floorplan, comprising placing a respective node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: GrantFiled: December 4, 2019Date of Patent: June 30, 2020Assignee: Google LLCInventors: Chian-min Richard Ho, William Hang, Mustafa Nazim Yazgan, Anna Darling Goldie, Jeffrey Adgate Dean, Azalia Mirhoseini, Emre Tuncer, Ya Wang, Anand Babu
-
Patent number: 10692003Abstract: A method for determining a placement for machine learning model operations across multiple hardware devices is described.Type: GrantFiled: June 19, 2019Date of Patent: June 23, 2020Assignee: Google LLCInventors: Samuel Bengio, Mohammad Norouzi, Benoit Steiner, Jeffrey Adgate Dean, Hieu Hy Pham, Azalia Mirhoseini, Quoc V. Le, Naveen Kumar, Yuefeng Zhou, Rasmus Munk Larsen
-
Publication number: 20200175216Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a computer chip floorplan. One of the methods includes obtaining netlist data for a computer chip; and generating a computer chip floorplan, comprising placing a respective node at each time step in a sequence comprising a plurality of time steps, the placing comprising, for each time step: generating an input representation for the time step; processing the input representation using a node placement neural network having a plurality of network parameters, wherein the node placement neural network is configured to process the input representation in accordance with current values of the network parameters to generate a score distribution over a plurality of positions on the surface of the computer chip; and assigning the node to be placed at the time step to a position from the plurality of positions using the score distribution.Type: ApplicationFiled: December 4, 2019Publication date: June 4, 2020Inventors: Chian-min Richard Ho, William Hang, Mustafa Nazim Yazgan, Anna Darling Goldie, Jeffrey Adgate Dean, Azalia Mirhoseini, Emre Tuncer, Ya Wang, Anand Babu
-
Publication number: 20190392294Abstract: A method for determining a placement for machine learning model operations across multiple hardware devices includes receiving data specifying machine learning operations, and determining a placement that assigns each of the operations specified by the data to a respective device from the multiple hardware devices.Type: ApplicationFiled: August 28, 2019Publication date: December 26, 2019Inventors: Benoit Steiner, Anna Darling Goldie, Jeffrey Adgate Dean, Hieu Hy Pham, Azalia Mirhoseini, Quoc V. Le
-
Patent number: 10438113Abstract: A method for determining a placement for machine learning model operations across multiple hardware devices includes receiving data specifying machine learning operations, and determining a placement that assigns each of the operations specified by the data to a respective device from the multiple hardware devices.Type: GrantFiled: July 19, 2018Date of Patent: October 8, 2019Assignee: Google LLCInventors: Benoit Steiner, Anna Darling Goldie, Jeffrey Adgate Dean, Hieu Hy Pham, Azalia Mirhoseini, Quoc V. Le
-
Publication number: 20190303761Abstract: A method for determining a placement for machine learning model operations across multiple hardware devices is described.Type: ApplicationFiled: June 19, 2019Publication date: October 3, 2019Inventors: Samy Bengio, Mohammad Edward Norouzi, Benoit Steiner, Jeffrey Adgate Dean, Hieu Hy Pham, Azalia Mirhoseini, Quoc V. Le, Naveen Kumar, Yuefeng Zhou, Rasmus Munk Larsen
-
Publication number: 20190251423Abstract: A system includes a neural network that includes a Mixture of Experts (MoE) subnetwork between a first neural network layer and a second neural network layer. The MoE subnetwork includes multiple expert neural networks. Each expert neural network is configured to process a first layer output generated by the first neural network layer to generate a respective expert output. The MoE subnetwork further includes a gating subsystem that selects, based on the first layer output, one or more of the expert neural networks and determine a respective weight for each selected expert neural network, provides the first layer output as input to each of the selected expert neural networks, combines the expert outputs generated by the selected expert neural networks in accordance with the weights for the selected expert neural networks to generate an MoE output, and provides the MoE output as input to the second neural network layer.Type: ApplicationFiled: April 24, 2019Publication date: August 15, 2019Inventors: Noam M. Shazeer, Azalia Mirhoseini, Krzysztof Stanislaw Maziarz
-
Publication number: 20190026624Abstract: A method for determining a placement for machine learning model operations across multiple hardware devices includes receiving data specifying machine learning operations, and determining a placement that assigns each of the operations specified by the data to a respective device from the multiple hardware devices.Type: ApplicationFiled: July 19, 2018Publication date: January 24, 2019Inventors: Benoit Steiner, Anna Darling Goldie, Jeffrey Adgate Dean, Hieu Hy Pham, Azalia Mirhoseini, Quoc V. Le
-
Patent number: 9911088Abstract: A “Context-Aware Crowdsourced Task Optimizer” provides various processes to optimize task recommendations for workers in mobile crowdsourcing scenarios by automatically identifying and recommending bundles of tasks compatible with workers' contexts (e.g., worker history, present or expected locations, travel paths, working hours, skill sets, capabilities of worker's mobile computing devices, etc.). The Context-Aware Crowdsourced Task Optimizer bundles tasks to both maximize expected numbers of completed tasks and to dynamically price tasks to maximize the system's utility, which is a function of task values and task completion rates. Advantageously, the resulting task identification and recommendation process incentivizes individual workers to perform more tasks in a shorter time period, thereby helping tasks to complete faster, even with smaller budgets.Type: GrantFiled: May 1, 2014Date of Patent: March 6, 2018Assignee: Microsoft Technology Licensing, LLCInventors: Suman Nath, Michel Goraczko, Jie Liu, Azalia Mirhoseini