公开(公告)号：US09652713B2

公开(公告)日：2017-05-16

申请号：US15090487

申请日：2016-04-04

Applicant: QUALCOMM TECHNOLOGIES INC.

Inventor： Jonathan James Hunt ,  Oleg Sinyavskiy ,  Robert Howard Kimball ,  Eric Martin Hall ,  Jeffrey Alexander Levin ,  Paul Bender ,  Michael-David Nakayoshi Canoy

IPC: G06F15/18 ,  G06N3/00 ,  G06N3/10 ,  G06N3/04

CPC classification number: G06N3/105 ,  G06N3/049

Abstract: Apparatus and methods for developing parallel networks. Parallel network design may comprise a general purpose language (GPC) code portion and a network description (ND) portion. GPL tools may be utilized in designing the network. The GPL tools may be configured to produce network specification language (NSL) engine adapted to generate hardware optimized machine executable code corresponding to the network description. The developer may be enabled to describe a parameter of the network. The GPC portion may be automatically updated consistent with the network parameter value. The GPC byte code may be introspected by the NSL engine to provide the underlying source code that may be automatically reinterpreted to produce the hardware optimized machine code. The optimized machine code may be executed in parallel.

公开(公告)号：US09165245B2

公开(公告)日：2015-10-20

申请号：US14275663

申请日：2014-05-12

Applicant: QUALCOMM TECHNOLOGIES INC.

Inventor： Eugene M. Izhikevich ,  Filip Piekniewski ,  Jayram Moorkanikara Nageswaran ,  Jeffrey Alexander Levin ,  Venkat Rangan ,  Erik Christopher Malone

IPC: G06N3/08 ,  G06N3/04 ,  G06N3/063

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

Abstract: Apparatus and methods for partial evaluation of synaptic updates in neural networks. In one embodiment, a pre-synaptic unit is connected to a several post synaptic units via communication channels. Information related to a plurality of post-synaptic pulses generated by the post-synaptic units is stored by the network in response to a system event. Synaptic channel updates are performed by the network using the time intervals between a pre-synaptic pulse, which is being generated prior to the system event, and at least a portion of the plurality of the post synaptic pulses. The system event enables removal of the information related to the portion of the post-synaptic pulses from the storage device. A shared memory block within the storage device is used to store data related to post-synaptic pulses generated by different post-synaptic nodes. This configuration enables memory use optimization of post-synaptic units with different firing rates.

Abstract translation: 用于部分评估神经网络中突触更新的装置和方法。 在一个实施例中，突触前单元经由通信信道连接到几个后突触单元。 与由突触后单元产生的多个突触后脉冲相关的信息由网络响应于系统事件存储。 突触信道更新由网络使用在系统事件之前产生的预触觉脉冲与多个突触后脉冲的至少一部分之间的时间间隔来执行。 系统事件使得能够从存储设备去除与突触后部分脉冲相关的信息。 存储设备内的共享存储器块用于存储与由不同的突触后节点产生的突触后脉冲相关的数据。 这种配置使得具有不同发射速率的突触后单元的存储器使用优化成为可能。

公开(公告)号：US09177246B2

公开(公告)日：2015-11-03

申请号：US13829919

申请日：2013-03-14

Applicant: QUALCOMM TECHNOLOGIES INC.

Inventor： Marius Buibas ,  Charles Wheeler Sweet, III ,  Mark S. Caskey ,  Jeffrey Alexander Levin

IPC: G06N3/04 ,  G06N3/08 ,  G06N3/02 ,  G06N3/00

CPC classification number: G06N3/04 ,  G06N3/008 ,  G06N3/02 ,  G06N3/049 ,  G06N3/063 ,  G06N3/08

Abstract: Apparatus and methods for an extensible robotic device with artificial intelligence and receptive to training controls. In one implementation, a modular robotic system that allows a user to fully select the architecture and capability set of their robotic device is disclosed. The user may add/remove modules as their respective functions are required/obviated. In addition, the artificial intelligence is based on a neuronal network (e.g., spiking neural network), and a behavioral control structure that allows a user to train a robotic device in manner conceptually similar to the mode in which one goes about training a domesticated animal such as a dog or cat (e.g., a positive/negative feedback training paradigm) is used. The trainable behavior control structure is based on the artificial neural network, which simulates the neural/synaptic activity of the brain of a living organism.

Abstract translation: 具有人工智能和接受训练控制的可扩展机器人装置的装置和方法。 在一个实现中，公开了允许用户完全选择其机器人设备的架构和能力集合的模块化机器人系统。 用户可以添加/删除模块，因为它们各自的功能是必需/免除的。 另外，人工智能是基于神经元网络（例如，刺激神经网络）以及行为控制结构，其允许用户以概念上类似于关于训练驯养动物的模式来训练机器人装置 例如狗或猫（例如，正/负反馈训练范例）。 可训练行为控制结构基于人造神经网络，其模拟活体的脑的神经/突触活动。

公开(公告)号：US09436908B2

公开(公告)日：2016-09-06

申请号：US13774934

申请日：2013-02-22

Applicant: QUALCOMM TECHNOLOGIES INC.

Inventor： Filip Piekniewski ,  Micah Richert ,  Eugene Izhikevich ,  Victor Hokkiu Chan ,  Jeffrey Alexander Levin

IPC: G06N3/08 ,  G06N3/02 ,  G06N99/00 ,  G06N3/04 ,  G06N3/10

CPC classification number: G06N3/08 ,  G06N3/02 ,  G06N3/049 ,  G06N3/10 ,  G06N99/005

Abstract: Apparatus and methods for activity based plasticity in a spiking neuron network adapted to process sensory input. In one approach, the plasticity mechanism of a connection may comprise a causal potentiation portion and an anti-causal portion. The anti-causal portion, corresponding to the input into a neuron occurring after the neuron response, may be configured based on the prior activity of the neuron. When the neuron is in low activity state, the connection, when active, may be potentiated by a base amount. When the neuron activity increases due to another input, the efficacy of the connection, if active, may be reduced proportionally to the neuron activity. Such functionality may enable the network to maintain strong, albeit inactive, connections available for use for extended intervals.

Abstract translation: 适用于处理感觉输入的加标神经元网络中基于活动的可塑性的装置和方法。 在一种方法中，连接的可塑性机构可以包括因果增强部分和反因果部分。 可以基于神经元的先前活动来配置对应于在神经元响应之后发生的神经元的输入的反因果部分。 当神经元处于低活动状态时，当活动时，连接可能被基础量增强。 当神经元活动由于另一个输入而增加时，如果活动，连接的功效可以与神经元活动成比例地降低。 这样的功能可以使得网络能够维持可用于延长的间隔的强的，虽然是不活动的连接。

公开(公告)号：US09460385B2

公开(公告)日：2016-10-04

申请号：US14466917

申请日：2014-08-22

Applicant: QUALCOMM TECHNOLOGIES INC.

Inventor： Filip Piekniewski ,  Micah Richert ,  Eugene Izhikevich ,  Victor Hokkiu Chan ,  Jeffrey Alexander Levin

IPC: G06N3/08 ,  G06N3/02 ,  G06N99/00 ,  G06N3/04 ,  G06N3/10

CPC classification number: G06N3/08 ,  G06N3/02 ,  G06N3/049 ,  G06N3/10 ,  G06N99/005

Abstract: Apparatus and methods for activity based plasticity in a spiking neuron network adapted to process sensory input. In one approach, the plasticity mechanism of a connection may comprise a causal potentiation portion and an anti-causal portion. The anti-causal portion, corresponding to the input into a neuron occurring after the neuron response, may be configured based on the prior activity of the neuron. When the neuron is in low activity state, the connection, when active, may be potentiated by a base amount. When the neuron activity increases due to another input, the efficacy of the connection, if active, may be reduced proportionally to the neuron activity. Such functionality may enable the network to maintain strong, albeit inactive, connections available for use for extended intervals.

公开(公告)号：US09299022B2

公开(公告)日：2016-03-29

申请号：US14468928

申请日：2014-08-26

Applicant: QUALCOMM TECHNOLOGIES INC.

Inventor： Marius Buibas ,  Charles Wheeler Sweet, III ,  Mark S. Caskey ,  Jeffrey Alexander Levin

IPC: G06N3/04

CPC classification number: G06N3/04 ,  G06N3/008 ,  G06N3/02 ,  G06N3/049 ,  G06N3/063 ,  G06N3/08

Abstract: Apparatus and methods for an extensible robotic device with artificial intelligence and receptive to training controls. In one implementation, a modular robotic system that allows a user to fully select the architecture and capability set of their robotic device is disclosed. The user may add/remove modules as their respective functions are required/obviated. In addition, the artificial intelligence is based on a neuronal network (e.g., spiking neural network), and a behavioral control structure that allows a user to train a robotic device in manner conceptually similar to the mode in which one goes about training a domesticated animal such as a dog or cat (e.g., a positive/negative feedback training paradigm) is used. The trainable behavior control structure is based on the artificial neural network, which simulates the neural/synaptic activity of the brain of a living organism.

公开(公告)号：US09177245B2

公开(公告)日：2015-11-03

申请号：US13763005

申请日：2013-02-08

Applicant: QUALCOMM TECHNOLOGIES INC.

Inventor： Micah Richert ,  Filip Piekniewski ,  Eugene Izhikevich ,  Sach Sokol ,  Victor Hokkiu Chan ,  Jeffrey Alexander Levin

IPC: G06N3/08 ,  G06N3/04

CPC classification number: G06K9/00711 ,  G06K9/4671 ,  G06N3/049 ,  G06N3/08 ,  G06T7/246 ,  G06T2207/10016

Abstract: Apparatus and methods for learning in response to temporally-proximate features. In one implementation, an image processing apparatus utilizes bi-modal spike timing dependent plasticity in a spiking neuron network. Based on a response by the neuron to a frame of input, the bi-modal plasticity mechanism is used to depress synaptic connections delivering the present input frame and to potentiate synaptic connections delivering previous and/or subsequent frames of input. The depression of near-contemporaneous input prevents the creation of a positive feedback loop and provides a mechanism for network response normalization.

Abstract translation: 响应于时间上接近的特征学习的装置和方法。 在一个实施方案中，图像处理装置在加标神经元网络中利用双模态尖峰时间相关可塑性。 基于神经元对输入帧的响应，双模式可塑性机制用于抑制递送当前输入帧的突触连接并且加强递送先前和/或后续输入帧的突触连接。 近同期输入的抑制阻止了正反馈回路的创建，并提供了网络响应归一化的机制。