专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明授权

US08359281B2 System and method for parallelizing and accelerating learning machine training and classification using a massively parallel accelerator 有权
标题翻译：使用大规模并行加速器并行和加速学习机训练和分类的系统和方法
公开(公告)号：US08359281B2
公开(公告)日：2013-01-22
申请号：US12478073
申请日：2009-06-04
申请人： Srihari Cadambi , Igor Durdanovic , Venkata Jakkula , Eric Cosatto , Murugan Sankaradass , Hans Peter Graf , Srimat T. Chakradhar
发明人： Srihari Cadambi , Igor Durdanovic , Venkata Jakkula , Eric Cosatto , Murugan Sankaradass , Hans Peter Graf , Srimat T. Chakradhar
IPC分类号： G06F15/18
CPC分类号： G06N99/005
摘要： A method system for training an apparatus to recognize a pattern includes providing the apparatus with a host processor executing steps of a machine learning process; providing the apparatus with an accelerator including at least two processors; inputting training pattern data into the host processor; determining coefficient changes in the machine learning process with the host processor using the training pattern data; transferring the training data to the accelerator; determining kernel dot-products with the at least two processors of the accelerator using the training data; and transferring the dot-products back to the host processor.
摘要翻译：用于训练识别模式的装置的方法系统包括向所述装置提供执行机器学习过程的步骤的主机处理器; 向所述设备提供包括至少两个处理器的加速器; 将训练模式数据输入到主处理器中; 使用所述训练模式数据来确定主处理器在机器学习过程中的系数变化; 将训练数据传送到加速器; 使用所述训练数据确定具有所述加速器的所述至少两个处理器的内核点产品; 并将点产品传送回主机处理器。

2. 发明申请

US20090304268A1 System and Method for Parallelizing and Accelerating Learning Machine Training and Classification Using a Massively Parallel Accelerator 有权
标题翻译：使用大规模并行加速器并行和加速学习机器训练和分类的系统和方法
公开(公告)号：US20090304268A1
公开(公告)日：2009-12-10
申请号：US12478073
申请日：2009-06-04
申请人： Srihari Cadambi , Igor Durdanovic , Venkata Jakkula , Eric Cosatto , Murugan Sankaradass , Hans Peter Graf , Srimat T. Chakradhar
发明人： Srihari Cadambi , Igor Durdanovic , Venkata Jakkula , Eric Cosatto , Murugan Sankaradass , Hans Peter Graf , Srimat T. Chakradhar
IPC分类号： G06K9/62 , G06K9/00
CPC分类号： G06N99/005
摘要： A method system for training an apparatus to recognize a pattern includes providing the apparatus with a host processor executing steps of a machine learning process; providing the apparatus with an accelerator including at least two processors; inputting training pattern data into the host processor; determining coefficient changes in the machine learning process with the host processor using the training pattern data; transferring the training data to the accelerator; determining kernel dot-products with the at least two processors of the accelerator using the training data; and transferring the dot-products back to the host processor.
摘要翻译：用于训练识别模式的装置的方法系统包括向所述装置提供执行机器学习过程的步骤的主机处理器; 向所述设备提供包括至少两个处理器的加速器; 将训练模式数据输入到主处理器中; 使用所述训练模式数据来确定主处理器在机器学习过程中的系数变化; 将训练数据传送到加速器; 使用所述训练数据确定具有所述加速器的所述至少两个处理器的内核点产品; 并将点产品传送回主机处理器。

3. 发明授权

US07406450B2 Spread kernel support vector machine 有权
标题翻译：扩展内核支持向量机
公开(公告)号：US07406450B2
公开(公告)日：2008-07-29
申请号：US11276235
申请日：2006-02-20
申请人： Hans Peter Graf , Igor Durdanovic , Eric Cosatto , Vladimir Vapnik
发明人： Hans Peter Graf , Igor Durdanovic , Eric Cosatto , Vladimir Vapnik
IPC分类号： G06E1/00 , G06E3/00 , G06F15/18 , G06G7/00 , G06N3/02
CPC分类号： G06K9/6269 , G06N99/005
摘要： Disclosed is a parallel support vector machine technique for solving problems with a large set of training data where the kernel computation, as well as the kernel cache and the training data, are spread over a number of distributed machines or processors. A plurality of processing nodes are used to train a support vector machine based on a set of training data. Each of the processing nodes selects a local working set of training data based on data local to the processing node, for example a local subset of gradients. Each node transmits selected data related to the working set (e.g., gradients having a maximum value) and receives an identification of a global working set of training data. The processing node optimizes the global working set of training data and updates a portion of the gradients of the global working set of training data. The updating of a portion of the gradients may include generating a portion of a kernel matrix. These steps are repeated until a convergence condition is met. Each of the local processing nodes may store all, or only a portion of, the training data. While the steps of optimizing the global working set of training data, and updating a portion of the gradients of the global working set, are performed in each of the local processing nodes, the function of generating a global working set of training data is performed in a centralized fashion based on the selected data (e.g., gradients of the local working set) received from the individual processing nodes.
摘要翻译：公开了一种用于解决大量训练数据的问题的并行支持向量机技术，其中内核计算以及内核高速缓存和训练数据分布在多个分布式机器或处理器上。多个处理节点用于基于一组训练数据训练支持向量机。每个处理节点基于处理节点本地的数据，例如梯度的本地子集，选择训练数据的本地工作集。每个节点发送与工作集有关的所选数据（例如，具有最大值的梯度）并且接收训练数据的全局工作集合的标识。处理节点优化训练数据的全局工作集，并更新全局训练数据工作集的一部分梯度。梯度的一部分的更新可以包括生成内核矩阵的一部分。重复这些步骤直到满足收敛条件。每个本地处理节点可以存储训练数据的全部或仅一部分。虽然在每个本地处理节点中执行优化训练数据的全局工作集和更新全局工作集的一部分梯度的步骤，但是在每个本地处理节点中执行生成训练数据的全局工作集的功能，基于从各个处理节点接收的所选数据（例如，本地工作集的梯度）的集中式。

4. 发明申请

US20110119467A1 MASSIVELY PARALLEL, SMART MEMORY BASED ACCELERATOR 有权
标题翻译：基于SMART MEMORY的加速器
公开(公告)号：US20110119467A1
公开(公告)日：2011-05-19
申请号：US12843579
申请日：2010-07-26
申请人： Srihari Cadambi , Abhinandan Majumdar , Michela Becchi , Srimat Chakradhar , Hans Peter Graf
发明人： Srihari Cadambi , Abhinandan Majumdar , Michela Becchi , Srimat Chakradhar , Hans Peter Graf
IPC分类号： G06F9/28 , G06F15/76
CPC分类号： G06F15/7821 , G06N3/063 , G06N99/005
摘要： Systems and methods for massively parallel processing on an accelerator that includes a plurality of processing cores. Each processing core includes multiple processing chains configured to perform parallel computations, each of which includes a plurality of interconnected processing elements. The cores further include multiple of smart memory blocks configured to store and process data, each memory block accepting the output of one of the plurality of processing chains. The cores communicate with at least one off-chip memory bank.
摘要翻译：在包括多个处理核心的加速器上进行大规模并行处理的系统和方法。每个处理核心包括配置成执行并行计算的多个处理链，每个处理链包括多个互连处理元件。核心还包括配置为存储和处理数据的多个智能存储器块，每个存储块接受多个处理链之一的输出。内核与至少一个片外存储体通信。

5. 发明授权

US08583896B2 Massively parallel processing core with plural chains of processing elements and respective smart memory storing select data received from each chain 有权
标题翻译：具有多个处理元件链的大规模并行处理核和存储从每个链接收的选择数据的各自的智能存储器
公开(公告)号：US08583896B2
公开(公告)日：2013-11-12
申请号：US12843579
申请日：2010-07-26
申请人： Srihari Cadambi , Abhinandan Majumdar , Michela Becchi , Srimat Chakradhar , Hans Peter Graf
发明人： Srihari Cadambi , Abhinandan Majumdar , Michela Becchi , Srimat Chakradhar , Hans Peter Graf
IPC分类号： G06F15/80
CPC分类号： G06F15/7821 , G06N3/063 , G06N99/005
摘要： Systems and methods for massively parallel processing on an accelerator that includes a plurality of processing cores. Each processing core includes multiple processing chains configured to perform parallel computations, each of which includes a plurality of interconnected processing elements. The cores further include multiple of smart memory blocks configured to store and process data, each memory block accepting the output of one of the plurality of processing chains. The cores communicate with at least one off-chip memory bank.
摘要翻译：在包括多个处理核心的加速器上进行大规模并行处理的系统和方法。每个处理核心包括配置成执行并行计算的多个处理链，每个处理链包括多个互连处理元件。核心还包括配置为存储和处理数据的多个智能存储器块，每个存储块接受多个处理链之一的输出。内核与至少一个片外存储体通信。

6. 发明授权

US08638846B1 Systems and methods for encoding and decoding video streams 有权
标题翻译：视频流编码和解码的系统和方法
公开(公告)号：US08638846B1
公开(公告)日：2014-01-28
申请号：US10601495
申请日：2003-06-23
申请人： Eric Cosatto , Hans Peter Graf , Joern Ostermann
发明人： Eric Cosatto , Hans Peter Graf , Joern Ostermann
IPC分类号： H04N7/12
CPC分类号： H04N19/573 , G06T9/001 , H04N19/21 , H04N19/23 , H04N19/55
摘要： Systems and methods for encoding/decoding a video stream. Animated talking heads are coded using partial offline encoding, multiple video streams, and multiple reference frames. The content of a face animation video that is known beforehand is encoded offline and the remaining content is encoded online and included in the video stream. To reduce bit rate, a server can stream multiple video sequences to the client and the video sequences are stored in the client's frame store. The server sends instructions to play a particular video sequence instead of streaming the particular video sequence. Multiple video streams can also be streamed to the client. Positional data and blending data are also sent to properly position one video stream relative to another video stream and to blend one video stream into another video stream.
摘要翻译：用于对视频流进行编码/解码的系统和方法。动画通话头使用部分离线编码，多个视频流和多个参考帧进行编码。预先知道的面部动画视频的内容被离线编码，并且剩余的内容被在线编码并包括在视频流中。为了降低比特率，服务器可以将多个视频序列流式传输到客户端，视频序列存储在客户端的帧存储器中。服务器发送播放特定视频序列的指令，而不是流式传输特定的视频序列。多个视频流也可以流式传输到客户端。还发送位置数据和混合数据以相对于另一个视频流适当地定位一个视频流，并将一个视频流混合到另一视频流中。

7. 发明申请

US20120069139A1 DIGITALLY-GENERATED LIGHTING FOR VIDEO CONFERENCING APPLICATIONS 有权
标题翻译：数字照明应用的数字照明
公开(公告)号：US20120069139A1
公开(公告)日：2012-03-22
申请号：US13303325
申请日：2011-11-23
申请人： Andrea Basso , Eric Cosatto , David Crawford Goibbon , Hans Peter Graf , Shan Liu
发明人： Andrea Basso , Eric Cosatto , David Crawford Goibbon , Hans Peter Graf , Shan Liu
IPC分类号： H04N7/14
CPC分类号： G06K9/00281 , G06F3/012 , G06K9/00228 , G06K9/2027 , G06T7/246 , G06T15/50 , G06T17/00 , G06T2200/08 , G06T2207/10152 , G06T2207/30201 , H04N7/147 , H04N7/15
摘要： A method of improving the lighting conditions of a real scene or video sequence. Digitally generated light is added to a scene for video conferencing over telecommunication networks. A virtual illumination equation takes into account light attenuation, lambertian and specular reflection. An image of an object is captured, a virtual light source illuminates the object within the image. In addition, the object can be the head of the user. The position of the head of the user is dynamically tracked so that an three-dimensional model is generated which is representative of the head of the user. Synthetic light is applied to a position on the model to form an illuminated model.
摘要翻译：一种改善真实场景或视频序列的照明条件的方法。数字生成的光被添加到通过电信网络进行视频会议的场景中。虚拟照明方程考虑了光衰减，朗伯和镜面反射。捕获对象的图像，虚拟光源照亮图像内的对象。另外，对象可以是用户的头。动态跟踪用户头部的位置，从而生成代表用户头部的三维模型。将合成光应用于模型上的位置以形成照明模型。

8. 发明授权

US08086066B2 Digitally-generated lighting for video conferencing applications 有权
标题翻译：用于视频会议应用的数字生成照明
公开(公告)号：US08086066B2
公开(公告)日：2011-12-27
申请号：US12877644
申请日：2010-09-08
申请人： Andrea Basso , Eric Cosatto , David Crawford Gibbon , Hans Peter Graf , Shan Liu
发明人： Andrea Basso , Eric Cosatto , David Crawford Gibbon , Hans Peter Graf , Shan Liu
IPC分类号： G06K9/40 , G06K9/00 , H04N7/14
CPC分类号： G06K9/00281 , G06F3/012 , G06K9/00228 , G06K9/2027 , G06T7/246 , G06T15/50 , G06T17/00 , G06T2200/08 , G06T2207/10152 , G06T2207/30201 , H04N7/147 , H04N7/15
摘要： A method of improving the lighting conditions of a real scene or video sequence. Digitally generated light is added to a scene for video conferencing over telecommunication networks. A virtual illumination equation takes into account light attenuation, lambertian and specular reflection. An image of an object is captured, a virtual light source illuminates the object within the image. In addition, the object can be the head of the user. The position of the head of the user is dynamically tracked so that an three-dimensional model is generated which is representative of the head of the user. Synthetic light is applied to a position on the model to form an illuminated model.
摘要翻译：一种改善真实场景或视频序列的照明条件的方法。数字生成的光被添加到通过电信网络进行视频会议的场景中。虚拟照明方程考虑了光衰减，朗伯和镜面反射。捕获对象的图像，虚拟光源照亮图像内的对象。另外，对象可以是用户的头。动态跟踪用户头部的位置，从而生成代表用户头部的三维模型。将合成光应用于模型上的位置以形成照明模型。

9. 发明授权

US07921013B1 System and method for sending multi-media messages using emoticons 有权
标题翻译：使用表情发送多媒体消息的系统和方法
公开(公告)号：US07921013B1
公开(公告)日：2011-04-05
申请号：US11214666
申请日：2005-08-30
申请人： Joern Ostermann , Mehmet Reha Civanlar , Eric Cosatto , Hans Peter Graf , Yann Andre LeCun
发明人： Joern Ostermann , Mehmet Reha Civanlar , Eric Cosatto , Hans Peter Graf , Yann Andre LeCun
IPC分类号： G10L13/00 , G10L13/06 , G10L21/00
CPC分类号： G06Q10/107 , G06F17/241 , G10L13/00
摘要： A system and method of providing sender-customization of multi-media messages through the use of emoticons is disclosed. The sender inserts the emoticons into a text message. As an animated face audibly delivers the text, emoticons associated with the message are started a predetermined period of time or number of words prior to the position of the emoticon in the message text and completed a predetermined length of time or number of words following the location of the emoticon. The sender may insert emoticons through the use of emoticon buttons that are icons available for choosing. Upon sender selections of an emoticon, an icon representing the emoticon is inserted into the text at the position of the cursor. Once an emoticon is chosen, the sender may also choose the amplitude for the emoticon and increased or decreased amplitude will be displayed in the icon inserted into the message text.
摘要翻译：公开了通过使用表情符号来提供多媒体消息的发送者定制的系统和方法。发件人将表情符号插入文本消息。作为动画面部听觉地传送文本，与消息相关联的表情符号在消息文本中表情符号的位置之前的预定时间段或字数开始，并且完成预定时间长度或位置之后的字数的表情符号。发件人可以通过使用可用于选择的图标的表情符号按钮来插入表情符号。在表情符号的发送者选择之后，表示表情符号的图标被插入到光标位置的文本中。一旦选择了表情符号，发送者也可以选择表情符号的幅度，增加或减小的幅度将显示在插入消息文本的图标中。

10. 发明授权

US07353177B2 System and method of providing conversational visual prosody for talking heads 有权
公开(公告)号：US07353177B2
公开(公告)日：2008-04-01
申请号：US11237557
申请日：2005-09-28
申请人： Eric Cosatto , Hans Peter Graf , Thomas M. Isaacson , Volker Franz Storm
发明人： Eric Cosatto , Hans Peter Graf , Thomas M. Isaacson , Volker Franz Storm
IPC分类号： G10L11/00
CPC分类号： G06T13/40 , G10L21/06
摘要： A system and method of controlling the movement of a virtual agent while the agent is listening to a human user during a conversation is disclosed. The method comprises receiving speech data from the user, performing a prosodic analysis of the speech data and controlling the virtual agent movement according to the prosodic analysis.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式