公开(公告)号：CA2926649A1

公开(公告)日：2015-05-07

申请号：CA2926649

申请日：2014-10-17

Applicant: QUALCOMM INC

Inventor： CAMPOS MICHAEL ,  WIERZYNSKI CASIMIR MATTHEW ,  BEHABADI BARDIA FALLAH

IPC: G06N3/04 ,  G06N3/08

Abstract: An artificial neural network may be configured to test the impact of certain input parameters. To improve testing efficiency and to avoid test runs that may not alter system performance, the effect of input parameters on neurons or groups of neurons may be determined to classify the neurons into groups based on the impact of certain parameters on those groups. Groups may be ordered serially and/or in parallel based on the interconnected nature of the groups and whether the output of neurons in one group may affect the operation of another. Parameters not affecting group performance may be pruned as inputs to that particular group prior to running system tests, thereby conserving processing resources during testing.

公开(公告)号：ES2912908T3

公开(公告)日：2022-05-30

申请号：ES16816014

申请日：2016-11-03

Applicant: QUALCOMM INC

Inventor： AGHAMOHAMMADI ALIAKBAR ,  BEHABADI BARDIA FALLAH ,  LOTT CHRISTOPHER ,  OMIDSHAFIEI SHAYEGAN ,  SOMASUNDARAM KIRAN ,  GIBSON SARAH PAIGE ,  WIERZYNSKI CASIMIR MATTHEW ,  AGARWAL SAURAV ,  REITMAYR GERHARD ,  DIAZ SPINDOLA SERAFIN

IPC: G05D1/02 ,  G01C21/20 ,  G01C21/32

Abstract: Un procedimiento (800) de planificación simultánea de una ruta y mapeo de un entorno por un robot, que comprende: determinar (802) una media de un nivel de ocupación para una ubicación en un mapa; determinar (804) una función de distribución de probabilidad, PDF, del nivel de ocupación; calcular (806) una función de costo en base al nivel medio de ocupación y una varianza del nivel de ocupación obtenida a partir de la PDF; y simultáneamente planificar (808) la ruta y mapear el entorno en base a la función de costo.

公开(公告)号：CA3004262A1

公开(公告)日：2017-06-22

申请号：CA3004262

申请日：2016-10-27

Applicant: QUALCOMM INC

Inventor： WIERZYNSKI CASIMIR MATTHEW ,  BEHABADI BARDIA FALLAH ,  GIBSON SARAH PAIGE ,  AGHAMOHAMMADI ALIAKBAR ,  AGARWAL SAURAV

IPC: G05D1/02

Abstract: A method of visual navigation for a robot includes integrating a depth map with localization information to generate a three-dimensional (3D) map. The method also includes motion planning based on the 3D map, the localization information, and/or a user input. The motion planning overrides the user input when a trajectory and/or a velocity, received via the user input, is predicted to cause a collision.

公开(公告)号：CA2924468A1

公开(公告)日：2015-04-23

申请号：CA2924468

申请日：2014-08-20

Applicant: QUALCOMM INC

Inventor： WIERZYNSKI CASIMIR MATTHEW ,  LEVIN JEFFREY ALEXANDER

IPC: G06N3/04

Abstract: A method for managing a neural network includes monitoring a congestion indication in a neural network. The method further includes modifying a spike distribution based on the monitored congestion indication.

公开(公告)号：WO2015138466A3

公开(公告)日：2015-12-30

申请号：PCT/US2015019731

申请日：2015-03-10

Applicant: QUALCOMM INC

Inventor： KONERTZ ANNE KATRIN ,  BEHABADI BARDIA FALLAH ,  BERNARTE JOEL SIMBULAN ,  HALL ERIC MARTIN ,  ROMERA JOLLIFF MARIA ELENA ,  WIERZYNSKI CASIMIR MATTHEW ,  SMOOT MICHAEL ELLIOT ,  BERLING JONATHAN NEAL

IPC: G06N3/10

CPC classification number: G06N3/10 ,  G06N3/049

Abstract: A method includes generating contextual feedback in a neuromorphic model. The neuromorphic model includes one or more assets to be monitored during development of the neuromorphic model. The method further includes displaying an interactive context panel to show a representation based on the contextual feedback.

Abstract translation: 一种方法包括在神经形态模型中产生语境反馈。 神经元模型包括在神经元模型开发期间要监测的一个或多个资产。 该方法还包括显示交互式上下文面板以基于上下文反馈来显示表示。

公开(公告)号：WO2015112713A3

公开(公告)日：2015-10-29

申请号：PCT/US2015012454

申请日：2015-01-22

Applicant: QUALCOMM INC

Inventor： CANOY MICHAEL-DAVID NAKAYOSHI ,  LIU YINYIN ,  CHAN VICTOR HOKKIU ,  CAMPOS MICHAEL ,  LEVIN JEFFREY ALEXANDER ,  WIERZYNSKI CASIMIR MATTHEW

IPC: G06N3/04 ,  G06N3/063 ,  G06N99/00

CPC classification number: G06F9/52 ,  G06N3/049 ,  G06N3/063 ,  G06N3/08 ,  G06N99/005

Abstract: Values are synchronized across processing blocks in a neural network by encoding spikes in a first processing block with a value to be shared across the neural network. The spikes may be transmitted to a second processing block in the neural network via an interblock interface. The received spikes are decoded in the second processing block so as to generate a value that is synchronized with the value of the first processing block.

Abstract translation: 通过在第一处理块中对尖峰进行编码以跨神经网络共享的值来在神经网络中的处理块之间同步数值。 尖峰可以经由帧间接口发送到神经网络中的第二处理块。 接收的尖峰在第二处理块中解码，以产生与第一处理块的值同步的值。

公开(公告)号：EP3295382A1

公开(公告)日：2018-03-21

申请号：EP16718942

申请日：2016-04-11

Applicant: QUALCOMM INC

Inventor： LIN DEXU ,  ANNAPUREDDY VENKATA SREEKANTA REDDY ,  JULIAN DAVID JONATHAN ,  WIERZYNSKI CASIMIR MATTHEW

IPC: G06N3/063 ,  G06N3/10

CPC classification number: G06N3/08 ,  G06F17/11 ,  G06N3/063 ,  G06N3/10

Abstract: A method for selecting bit widths for a fixed point machine learning model includes evaluating a sensitivity of model accuracy to bit widths at each computational stage of the model. The method also includes selecting a bit width for parameters, and/or intermediate calculations in the computational stages of the mode. The bit width for the parameters and the bit width for the intermediate calculations may be different. The selected bit width may be determined based on the sensitivity evaluation.

公开(公告)号：WO2017052709A3

公开(公告)日：2017-05-04

申请号：PCT/US2016039661

申请日：2016-06-27

Applicant: QUALCOMM INC

Inventor： WIERZYNSKI CASIMIR MATTHEW

IPC: G06N3/04

CPC classification number: G06N3/08 ,  G06N3/0454

Abstract: A method of transfer learning includes receiving second data and generating, via a first network, second labels for the second data. In one configuration, the first network has been previously trained on first labels for first data. Additionally, the second labels are generated for training a second network.

Abstract translation: 一种转移学习方法包括接收第二数据并且经由第一网络产生第二数据的第二标签。 在一种配置中，第一网络先前已经在第一数据的第一标签上进行了训练。 另外，生成第二个标签用于训练第二个网络。

公开(公告)号：WO2015065738A3

公开(公告)日：2015-07-09

申请号：PCT/US2014061220

申请日：2014-10-17

Applicant: QUALCOMM INC

Inventor： CAMPOS MICHAEL ,  WIERZYNSKI CASIMIR MATTHEW ,  BEHABADI BARDIA FALLAH

IPC: G06N3/04 ,  G06N3/08

CPC classification number: G06N3/082 ,  G06N3/04 ,  G06N3/049 ,  G06N3/10 ,  G06N3/105 ,  G08B17/005

Abstract: An artificial neural network may be configured to test the impact of certain input parameters. To improve testing efficiency and to avoid test runs that may not alter system performance, the effect of input parameters on neurons or groups of neurons may be determined to classify the neurons into groups based on the impact of certain parameters on those groups. Groups may be ordered serially and/or in parallel based on the interconnected nature of the groups and whether the output of neurons in one group may affect the operation of another. Parameters not affecting group performance may be pruned as inputs to that particular group prior to running system tests, thereby conserving processing resources during testing.

Abstract translation: 可以配置人造神经网络来测试某些输入参数的影响。 为了提高测试效率并避免可能不会改变系统性能的测试运行，可以基于某些参数对这些组的影响来确定输入参数对神经元或神经元组的影响，以将神经元分组成组。 可以基于组的相互联系的性质，并且一组中的神经元的输出是否可能影响另一组的操作，可以串联和/或并行排列组。 在运行系统测试之前，不影响组性能的参数可以作为该特定组的输入修剪，从而在测试期间节省处理资源。

公开(公告)号：WO2015119963A2

公开(公告)日：2015-08-13

申请号：PCT/US2015014297

申请日：2015-02-03

Applicant: QUALCOMM INC ,  MARCOS NINA

Inventor： CAREY RYAN MICHAEL ,  CHAN VICTOR HOKKIU ,  WIERZYNSKI CASIMIR MATTHEW ,  HUNZINGER JASON FRANK

CPC classification number: G06N3/063 ,  G06N3/049 ,  G06N3/08 ,  G06N3/088

Abstract: A method for creating and maintaining short term memory using short term plasticity, includes changing a gain of a synapse based on presynaptic spike activity without regard to postsynaptic spike activity. The method also includes calculating the gain based on a continuously updated synaptic state variable associated with the short term plasticity.

Abstract translation: 使用短期可塑性来创建和维持短期记忆的方法包括基于突触前峰值活动来改变突触的增益，而不考虑突触后峰值活动。 该方法还包括基于与短期可塑性相关联的连续更新的突触状态变量来计算增益。