专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

11. 发明申请

US20180316325A1 AMPLIFIER CIRCUIT THAT AMPLIFIES DIFFERENTIAL SIGNAL AND OPTICAL MODULE THAT INCLUDES AMPLIFIER CIRCUIT 审中-公开
公开(公告)号：US20180316325A1
公开(公告)日：2018-11-01
申请号：US15922973
申请日：2018-03-16
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai
IPC分类号： H03F3/45 , H03F1/02 , H04B10/516
CPC分类号： H03F3/45215 , H03F1/0216 , H03F1/0261 , H03F1/086 , H03F3/082 , H03F3/45381 , H03F2200/21 , H03F2200/555 , H03F2203/45022 , H03F2203/45151 , H03F2203/45154 , H03F2203/45178 , H03F2203/45674 , H03F2203/45702 , H04B10/40 , H04B10/516
摘要： An amplifier circuit includes: an amplifier; and a bias circuit that controls an operation point of the amplifier. The amplifier includes: a load resistor; a differential transistor pair electrically coupled to the load resistor; and a tail transistor electrically coupled to the differential transistor pair. The bias circuit includes: a voltage generator circuit that generates a reference voltage corresponding to a sum of a threshold voltage of a transistor in the differential transistor pair and a saturation drain voltage of the tail transistor; and a current generator circuit that generates a reference current that is proportional to a difference between a power supply voltage of the amplifier circuit and the reference voltage by using a reference resistor. The current generator circuit is electrically coupled to the amplifier such that a tail current that flows through the tail transistor is proportional to the reference current.

12. 发明申请

US20170054575A1 DECISION FEEDBACK EQUALIZER AND RECEIVER CIRCUIT 有权
标题翻译：决策反馈均衡器和接收器电路
公开(公告)号：US20170054575A1
公开(公告)日：2017-02-23
申请号：US15200626
申请日：2016-07-01
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai
IPC分类号： H04L25/03 , H04J11/00 , H04W72/04 , H04L12/26
CPC分类号： H04L25/03057 , H04L25/03006
摘要： A plurality of comparator circuits output results of comparing a pulse amplitude modulated input signal having four or more values with a plurality of first thresholds. A selection section decides a value of the input signal at certain timing by selecting one of the comparison results on the basis of a result of deciding a value of the input signal at previous timing. A threshold setting section generates the first thresholds from a plurality of third thresholds obtained by adding an offset value based on magnitude of inter-symbol interference corresponding to each value to one of second thresholds whose number is based on the number of values, on the basis of an average value of third thresholds greater than a second threshold or third thresholds smaller than the second threshold and an adjustment value based on the decision result and sets the first thresholds in the comparator circuits.
摘要翻译：多个比较器电路输出具有四个或更多个值的脉冲幅度调制输入信号与多个第一阈值的结果。选择部根据在上一次的时刻判定输入信号的值的结果，通过选择比较结果之一来决定某一定时的输入信号的值。阈值设定部分根据多个第三阈值生成第一阈值，该多个第三阈值是通过将基于每个值的码元间干扰的大小的偏移值与基于值的数量的第二阈值之一相加而获得的，具有大于第二阈值的第三阈值或小于第二阈值的第三阈值的平均值，以及基于判定结果的调整值，并且设置比较器电路中的第一阈值。

13. 发明授权

US09425777B2 Phase interpolator 有权
标题翻译：相位插值器
公开(公告)号：US09425777B2
公开(公告)日：2016-08-23
申请号：US14682377
申请日：2015-04-09
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai , Toshihiko Mori
IPC分类号： H04L25/02 , H03K5/13 , G06F1/02 , H04L27/18 , H04L7/00 , H04L27/00 , H03K5/00
CPC分类号： H03K5/13 , G06F1/022 , H03K2005/00052 , H04L7/0025 , H04L27/18 , H04L2027/0016
摘要： A phase interpolator has: a mixer configured to synthesize phases of a plurality of input cosine-wave or sine-wave signals whose phases are different from each other; and a bias generator configured to output a bias signal in accordance with a phase control signal to the mixer, and the mixer outputs a signal with a phase in accordance with the phase control signal.
摘要翻译：相位插值器具有：混频器，被配置为合成相位彼此不同的多个输入余弦波或正弦波信号的相位; 以及偏置发生器，被配置为根据相位控制信号向混频器输出偏置信号，并且混频器根据相位控制信号输出具有相位的信号。

14. 发明公开

US20230281440A1 COMPUTER-READABLE RECORDING MEDIUM HAVING STORED THEREIN MACHINE LEARNING PROGRAM, METHOD FOR MACHINE LEARNING, AND INFORMATION PROCESSING APPARATUS 审中-公开
公开(公告)号：US20230281440A1
公开(公告)日：2023-09-07
申请号：US17984285
申请日：2022-11-10
申请人： Fujitsu Limited
发明人： Yasufumi Sakai
IPC分类号： G06N3/08 , G06F17/16
CPC分类号： G06N3/08 , G06F17/16
摘要： A method including: obtaining a reduction ratio of each element of layers in a trained model of a neural network; when the neural network includes a process that outputs a tensor as a result of a given calculation on tensors and when tensors from first layers preceding the process are inputted, inserting a second layer that performs a zero padding between the first layers and the process, the first layers including a preceding layer of the process and including one or more layers preceding the preceding layer and being shortcut-connected to the process; and padding tensors inputted into second layers associated one with each first layer with one or more zero matrices such that a number of elements of each tensor inputted into the process from the first layers after reducing of elements of each first layer in accordance with the reduction ratio comes to be a first number.

15. 发明公开

US20230162036A1 COMPUTER-READABLE RECORDING MEDIUM HAVING STORED THEREIN MACHINE LEARNING PROGRAM, METHOD FOR MACHINE LEARNING, AND INFORMATION PROCESSING APPARATUS 审中-公开
公开(公告)号：US20230162036A1
公开(公告)日：2023-05-25
申请号：US17886499
申请日：2022-08-12
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai
IPC分类号： G06N3/08
CPC分类号： G06N3/082
摘要： A computer-readable recording medium has stored therein a machine learning program for causing a computer to execute a process including: selecting a reduction ratio of each element of a plurality of layers in a trained model of a neural network including the plurality of layers; and adjusting, when the neural network includes a calculating process that outputs a tensor serving as a result of a given calculation on a tensor from a first layer and one or more tensors of one or more second layers preceding the first layer, a first reduction ratio and one or more second reduction ratios based on one or more elements to be reduced in the first layer at the first reduction ratio and one or more elements to be reduced in each of the one or more second layers at the one or more second reduction ratios.

16. 发明申请

US20230130638A1 COMPUTER-READABLE RECORDING MEDIUM HAVING STORED THEREIN MACHINE LEARNING PROGRAM, METHOD FOR MACHINE LEARNING, AND INFORMATION PROCESSING APPARATUS 有权
公开(公告)号：US20230130638A1
公开(公告)日：2023-04-27
申请号：US17863433
申请日：2022-07-13
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai
IPC分类号： G06N3/08
摘要： A computer-readable recording medium has stored therein a program for causing a computer to execute a process including: calculating thresholds of errors in tensors between before and after reduction one for each element of a plurality of layers in a trained model of a neural network including the layers; selecting reduction ratio candidates to be applied one to each of the layers based on the thresholds and errors in tensors between before and after reduction in cases where the elements are reduced by each of reduction ratio candidates in each of the layers; and determining reduction ratios to be applied one to each of the layers based on inference accuracy of the trained model and inference accuracy of a reduced model after machine learning, the reduced model being obtained by reducing each element of the layers in the trained model according to the reduction ratio candidates to be applied.

17. 发明申请

US20220172022A1 STORAGE MEDIUM, QUANTIZATION METHOD, AND QUANTIZATION APPARATUS 有权
公开(公告)号：US20220172022A1
公开(公告)日：2022-06-02
申请号：US17500998
申请日：2021-10-14
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai
IPC分类号： G06N3/04
摘要： A non-transitory computer-readable storage medium storing a quantization program that causes at least one computer to execute a process, the process includes calculating, for all layers of a neural network, differences each between a trust region radius threshold and a quantization error of a first bit width that is narrower by one step than a second bit width; calculating, based on the differences, each scaling coefficient for all the layers; updating a trust region radius by using a smallest value among the scaling coefficients; and quantizing a parameter of the neural network by a third bit width set based on the trust region.

18. 发明申请

US20220027758A1 INFORMATION PROCESSING APPARATUS AND INFORMATION PROCESSING METHOD 有权
公开(公告)号：US20220027758A1
公开(公告)日：2022-01-27
申请号：US17228532
申请日：2021-04-12
申请人： FUJITSU LIMITED
发明人： Enxhi Kreshpa , Shigeto SUZUKI , Yasufumi Sakai , Takashi Shiraishi , Takuji YAMAMOTO
IPC分类号： G06N5/04 , G06F9/48 , G06F1/329 , G06N20/00
摘要： First clustering is performed on a plurality of samples each including time-series measurement values of power consumption to thereby generate a plurality of first clusters. The plurality of first clusters are each classified as a second cluster satisfying a determination condition or a third cluster that does not satisfy the determination condition. The determination condition includes at least one of a first criterion in which the variance of correlation values between samples is less than a first threshold and a second criterion in which the average of the correlation values exceeds a second threshold. Second clustering is performed on samples included in the third cluster to divide the third cluster into a plurality of fourth clusters. Training data for use in generation of a model for predicting power consumption is generated based on the second cluster and at least one of the fourth clusters.

19. 发明申请

US20210081802A1 INFORMATION PROCESSING DEVICE AND METHOD, AND RECORDING MEDIUM 有权
公开(公告)号：US20210081802A1
公开(公告)日：2021-03-18
申请号：US16997989
申请日：2020-08-20
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai
IPC分类号： G06N3/08 , G06N3/063 , G06N5/04
摘要： An information processing device, includes a memory; and a processor coupled to the memory and configured to: calculate a quantization error when a variable to be used in a neural network is quantized, generate a threshold value based on reference information related to a first recognition rate obtained by past learning of the neural network and a second recognition rate that is obtained by calculation of the neural network, determine a variable of data type to be quantized among variables to be used for calculation of the neural network based on the calculated quantization error and the generated threshold value, and execute the calculation of the neural network by using the variable of data type.

20. 发明授权

US10728058B2 Decision feedback equalizer and interconnect circuit 有权
公开(公告)号：US10728058B2
公开(公告)日：2020-07-28
申请号：US16451225
申请日：2019-06-25
申请人： FUJITSU LIMITED
发明人： Yasufumi Sakai
IPC分类号： H04L25/03 , H03K3/037 , H03K5/003 , H03K5/24 , H04B3/04
摘要： A decision feedback equalizer includes: a comparison circuit; a latch circuit configured to latch a result of comparison by the comparison circuit; a setting circuit configured to set a decision threshold of the comparison circuit in accordance with a control signal; and a switch circuit configured to be controlled to be turned on and off by an output signal from the latch circuit, wherein the setting circuit is configured to be connected in parallel with an input stage of the comparison circuit through the switch circuit and operate in synchronization with a clock signal for driving the comparison circuit.

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