Title:
INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD, AND INFORMATION PROCESSING PROGRAM
Document Type and Number:
WIPO Patent Application WO/2022/176521
Kind Code:
A1
Abstract:
This information processing device (10, 10A, 10B) comprises an operation unit (120, 120A, 120B) which integrates input values and weight coefficients. Any one among the input value or the weight coefficient is represented as the addition or subtraction of at most n number of first bit streams. The first bit stream has one bit having a value of 1.
Inventors:
SAKUMA YUIKO (JP)
SUMIHIRO HIROSHI (JP)
SUMIHIRO HIROSHI (JP)
Application Number:
PCT/JP2022/002493
Publication Date:
August 25, 2022
Filing Date:
January 25, 2022
Export Citation:
Assignee:
SONY GROUP CORP (JP)
International Classes:
G06N3/063; G06F7/533
Domestic Patent References:
| WO2018212946A1 | 2018-11-22 | |||
| WO2007034366A2 | 2007-03-29 |
Other References:
VANHOUCKE, V.SENIOR, A.MAO, M. Z, IMPROVING THE SPEED OF NEURAL NETWORKS ON CPUS, 2011
ZHOU, A.YAO, A.GUO, Y.XU, L.CHEN, Y: "Incremental network quantization: Towards lossless cnns with low-precision weights", ARXIV PREPRINT ARXIV, vol. 1702, 2017, pages 03044
LEE, E. H., MIYASHITA, D., CHAI, E., MURMANN, B., & WONG, S. S.: "Lognet: Energy-efficient neural networks using logarithmic computation", IN 2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP, March 2017 (2017-03-01), pages 5900 - 5904, XP033259542, DOI: 10.1109/ICASSP.2017.7953288
VOGEL, S., LIANG, M., GUNTORO, A., STECHELE, W., & ASCHEID, G.: "Efficient hardware acceleration of cnns using logarithmic data representation with arbitrary log-base", IN PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON COMPUTER-AIDED DESIGN, November 2018 (2018-11-01), pages 1 - 8, XP058420986, DOI: 10.1145/3240765.3240803
LI, Y.DONG, X.WANG, W: "Additive Powers-of-Two Quantization: An Efficient Non-uniform Discretization for Neural Networks", IN INTERNATIONAL CONFERENCE ON LEARNING REPRESENTATIONS, September 2019 (2019-09-01)
ZHOU, A.YAO, A.GUO, Y.XU, L.CHEN, Y: "Incremental network quantization: Towards lossless cnns with low-precision weights", ARXIV PREPRINT ARXIV, vol. 1702, 2017, pages 03044
LEE, E. H., MIYASHITA, D., CHAI, E., MURMANN, B., & WONG, S. S.: "Lognet: Energy-efficient neural networks using logarithmic computation", IN 2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP, March 2017 (2017-03-01), pages 5900 - 5904, XP033259542, DOI: 10.1109/ICASSP.2017.7953288
VOGEL, S., LIANG, M., GUNTORO, A., STECHELE, W., & ASCHEID, G.: "Efficient hardware acceleration of cnns using logarithmic data representation with arbitrary log-base", IN PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON COMPUTER-AIDED DESIGN, November 2018 (2018-11-01), pages 1 - 8, XP058420986, DOI: 10.1145/3240765.3240803
LI, Y.DONG, X.WANG, W: "Additive Powers-of-Two Quantization: An Efficient Non-uniform Discretization for Neural Networks", IN INTERNATIONAL CONFERENCE ON LEARNING REPRESENTATIONS, September 2019 (2019-09-01)
Attorney, Agent or Firm:
SAKAI INTERNATIONAL PATENT OFFICE (JP)
Download PDF:
Previous Patent: COMPOSITION FOR FORMING ELECTRODE, SOLAR CELL ELEMENT, AND ALUMINUM/SILVER STACKED ELECTRODE
Next Patent: SENSOR DEVICE
Next Patent: SENSOR DEVICE