公开(公告)号：EP3329425A1

公开(公告)日：2018-06-06

申请号：EP16745016.2

申请日：2016-07-19

申请人： Qualcomm Incorporated

发明人： TADESSE, Henok Tefera ,  CHAKRABORTY, Avijit ,  JULIAN, David Jonathan ,  STOKMAN, Henricus Meinardus ,  DE ROOIJ, Ork ,  VAN DE SANDE, Koen Erik Adriaan ,  ANNAPUREDDY, Venkata Sreekanta Reddy

IPC分类号： G06K9/62

CPC分类号： G06N3/088 ,  G06K9/6265 ,  G06N3/02 ,  G06N3/04 ,  G06N99/005

摘要： Multi-label classification is improved by determining thresholds and/or scale factors. Selecting thresholds for multi-label classification includes sorting a set of label scores associated with a first label to create an ordered list. Precision and recall values are calculated corresponding to a set of candidate thresholds from score values. The threshold is selected from the candidate thresholds for the first label based on target precision values or recall values. A scale factor is also selected for an activation function for multi-label classification where a metric of scores within a range is calculated. The scale factor is adjusted when the metric of scores are not within the range.

公开(公告)号：EP3289529A1

公开(公告)日：2018-03-07

申请号：EP16722460.9

申请日：2016-04-20

申请人： QUALCOMM Incorporated

发明人： ANNAPUREDDY, Venkata Sreekanta Reddy

IPC分类号： G06N3/08 ,  G06K9/80

CPC分类号： G06T3/4046 ,  G06K9/4623 ,  G06K9/627 ,  G06K9/66 ,  G06N3/082

摘要： A method of reducing image resolution in a deep convolutional network (DCN) includes dynamically selecting a reduction factor to be applied to an input image. The reduction factor can be selected at each layer of the DCN. The method also includes adjusting the DCN based on the reduction factor selected for each layer.

公开(公告)号：EP3170127A1

公开(公告)日：2017-05-24

申请号：EP15748340.5

申请日：2015-07-13

申请人： Qualcomm Incorporated

发明人： ANNAPUREDDY, Venkata Sreekanta Reddy

IPC分类号： G06N3/08

CPC分类号： G06N3/08 ,  G06N3/04 ,  G06N3/049 ,  G06N3/084

摘要： A method of training a neural network includes encouraging one or more filters in the neural network to have a low rank.

公开(公告)号：EP3183695A1

公开(公告)日：2017-06-28

申请号：EP15748396.7

申请日：2015-07-28

申请人： Qualcomm Incorporated

发明人： MAJUMDAR, Somdeb ,  TOWAL, Regan Blythe ,  TALATHI, Sachin Subhash ,  JULIAN, David Jonathan ,  ANNAPUREDDY, Venkata Sreekanta Reddy

IPC分类号： G06N3/04 ,  G06N7/00

CPC分类号： G06N3/08 ,  G06N3/04 ,  G06N3/0427 ,  G06N3/049 ,  G06N7/005

摘要： A method for classifying an object includes applying multiple confidence values to multiple objects. The method also includes determining a metric based on the multiple confidence values. The method further includes determining a classification of a first object from the multiple objects based on a knowledge-graph when the metric is above a threshold.

摘要翻译： 用于分类对象的方法包括将多个置信值应用于多个对象。 该方法还包括基于多个置信度值确定度量。 该方法还包括当度量高于阈值时基于知识图从多个对象中确定第一对象的分类。

公开(公告)号：EP3170126A1

公开(公告)日：2017-05-24

申请号：EP15742477.1

申请日：2015-07-13

申请人： Qualcomm Incorporated

发明人： ANNAPUREDDY, Venkata Sreekanta Reddy

IPC分类号： G06N3/08

CPC分类号： G06N3/08 ,  G06N3/04 ,  G06N3/049 ,  G06N3/084

摘要： A method of training a neural network includes encouraging one or more filters in the neural network to have a low rank.

摘要翻译： 一种训练神经网络的方法包括鼓励神经网络中的一个或多个滤波器具有低秩。

公开(公告)号：EP3123405A2

公开(公告)日：2017-02-01

申请号：EP15719876.3

申请日：2015-03-17

申请人： Qualcomm Incorporated

发明人： ANNAPUREDDY, Venkata Sreekanta Reddy ,  JULIAN, David Jonathan ,  SARAH, Anthony

IPC分类号： G06N3/04 ,  G06N3/08

CPC分类号： G06N3/0454 ,  G06N3/049 ,  G06N3/088

摘要： A method of distributed computation includes computing a first set of results in a first computational chain with a first population of processing nodes and passing the first set of results to a second population of processing nodes. The method also includes entering a first rest state with the first population of processing nodes after passing the first set of results and computing a second set of results in the first computational chain with the second population of processing nodes based on the first set of results. The method further includes passing the second set of results to the first population of processing nodes, entering a second rest state with the second population of processing nodes after passing the second set of results and orchestrating the first computational chain.

摘要翻译： 一种分布式计算的方法包括：在具有第一种处理节点的第一计算链中计算第一组结果，并将第一组结果传递给第二种处理节点。 该方法还包括在通过第一组结果之后，与处理节点的第一群体一起输入第一休息状态，并且基于第一组结果在第一计算链中与第二种处理节点计算第二组结果。 所述方法还包括将所述第二组结果传递到所述第一处理节点群，在通过所述第二组结果并编排所述第一计算链之后，与所述第二处理节点群进入第二休息状态。

公开(公告)号：EP3123404A2

公开(公告)日：2017-02-01

申请号：EP15716612.5

申请日：2015-03-17

申请人： Qualcomm Incorporated

发明人： ANNAPUREDDY, Venkata Sreekanta Reddy ,  JULIAN, David Jonathan ,  TOWAL, Regan Blythe ,  LIU, Yinyin

IPC分类号： G06N3/04

CPC分类号： G06N3/0445 ,  G06N3/0481 ,  G06N3/049

摘要： Differential encoding in a neural network includes predicting an activation value for a neuron in the neural network based on at least one previous activation value for the neuron. The encoding further includes encoding a value based on a difference between the predicted activation value and an actual activation value for the neuron in the neural network.

摘要翻译： 神经网络中的差分编码包括基于神经元的至少一个以前的激活值来预测神经网络中神经元的激活值。 编码还包括基于预测的激活值和神经网络中的神经元的实际激活值之间的差来编码值。

公开(公告)号：EP3937087A1

公开(公告)日：2022-01-12

申请号：EP21183220.9

申请日：2015-07-28

申请人： QUALCOMM Incorporated

发明人： MAJUMDAR, Somdeb ,  TOWAL, Regan Blythe ,  TALATHI, Sachin Subhash ,  JULIAN, David Jonathan ,  ANNAPUREDDY, Venkata Sreekanta Reddy

IPC分类号： G06N3/04 ,  G06N7/00

摘要： A method for classifying an object includes applying multiple confidence values to multiple objects. The method also includes determining a metric based on the multiple confidence values. The method further includes determining a classification of a first object from the multiple objects based on a knowledge-graph when the metric is above a threshold.

公开(公告)号：EP3852016A1

公开(公告)日：2021-07-21

申请号：EP21161624.8

申请日：2016-04-11

申请人： QUALCOMM Incorporated

发明人： LIN, Dexu ,  ANNAPUREDDY, Venkata Sreekanta Reddy ,  JULIAN, David Jonathan ,  WIERZYNSKI, Casimir Matthew

IPC分类号： G06N3/063 ,  G06N3/10

摘要： 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：EP3320486A1

公开(公告)日：2018-05-16

申请号：EP16736704.4

申请日：2016-06-27

申请人： Qualcomm Incorporated

发明人： DIJKMAN, Daniel Hendricus Franciscus ,  TOWAL, Regan Blythe ,  ANNAPUREDDY, Venkata Sreekanta Reddy

IPC分类号： G06K9/32 ,  G06K9/00 ,  G06K9/72 ,  G06K9/46

CPC分类号： G06K9/66 ,  G06K9/00684 ,  G06K9/3241 ,  G06K9/4628 ,  G06K9/6267 ,  G06K9/72 ,  G06N3/02

摘要： Context-based priors are utilized in machine learning networks (e.g., neural networks) for detecting objects in images. The likely locations of objects are estimated based on context labels. A machine learning network identifies a context label of an entire image. Based on the, the network selects a set of likely regions for detecting objects of interest in the image.