John E Arthur

20130073497, Mar 21, 2013

Sep 16, 2011

13/235341

Filipp Akopyan - New Windsor NY, US
John V. Arthur - Mountain View CA, US
Rajit Manohar - Ithaca NY, US
Paul A. Merolla - Palo Alto CA, US
Dharmendra S. Modha - San Jose CA, US
Alyosha Molnar - Ithaca NY, US

Cornell University - Ithaca NY
INTERNATIONAL BUSINESS MACHINES CORPORATION - Armonk NY

G06N 3/04

706 27

An event-driven neural network includes a plurality of interconnected core circuits is provided. Each core circuit includes an electronic synapse array has multiple digital synapses interconnecting a plurality of digital electronic neurons. A synapse interconnects an axon of a pre-synaptic neuron with a dendrite of a post-synaptic neuron. A neuron integrates input spikes and generates a spike event in response to the integrated input spikes exceeding a threshold. Each core circuit also has a scheduler that receives a spike event and delivers the spike event to a selected axon in the synapse array based on a schedule for deterministic event delivery.

57963681, Aug 18, 1998

Apr 23, 1996

8/639914

John E. Arthur - Island Park NY

Lockleed Martin Corporation - Bethesda MD

H01Q 102

343704

A wave reflecting dish is constructed to provide a one piece unit having a hollow, watertight region between a first external surface having a paraboloidal contour and a second external surface which includes a base. Heater elements are positioned in the hollow region on a surface opposite the first external surface. Insulating material cover the heater elements and the hollow region is filled with a foam material. Snow, ice, and water drainage from the first surface is provided through a hole which extends through the unit from the first surface to the second surface. Provisions for mounting wave emitting and receiving apparatus are provided within the first surface. Attachment pads for mounting the dish on an external structure are provided on the base.

52852113, Feb 8, 1994

Sep 2, 1992

7/939260

John C. Herper - Glen Cove NY
Anthony M. Bucceri - Larchmont NY
John E. Arthur - Island Park NY

Unisys Corporation - Blue Bell PA

H01Q 116

343791

An array of collinear coaxial lines utilizes standard size brass tubing with brass rods inserted therein to establish each half-wavelength coaxial section. A coupler positioned between each half-wavelength section has an outer conductor supports and through passages for the inner conductor. Inner conductors of the transmission line elements are soldered to the outer conductors of the respective adjacent transmission line elements so that all elements are excited with the same phase and polarity. Air dielectric between the inner and outer conductors provides a transmission line wavelength equal to that of free space so that the length of each element in the array is a free space half-wavelength long, thereby providing an antenna gain greater than that achievable with transmission lines having solid dielectric supporting the inner conductor.

55024548, Mar 26, 1996

Nov 9, 1994

8/337334

Thomas J. Psinakis - East Meadow NY
John E. Arthur - Island Park NY
Anthony M. Bucceri - Larchmont NY

Unisys Corporation - Blue Bell PA

H01Q 904

343791

A shell structure for a collinear coaxial element array antenna is made of electrical conducting material and includes first and second circular sections separated by a web. The first circular section has an inner diameter that is slightly larger than the diameter of an inner conductor in a coaxial line element of the array antenna, while the second circular section has an inner diameter that is equal to that of an outer conductor of a coaxial transmission line element. The separations of the centers of the first and second circular sections are established by the wed to be equal to the distance between the centers of two adjacent coaxial transmission line elements in the array antenna. The inner conductor of a coaxial transmission line element is inserted in the first circular section and is electrically connected thereto. This electrically connects the inner conductor of one coaxial transmission line element to the second circular section, which serves as the outer conductor of a coaxial transmission line element that is adjacent to the one coaxial transmission line element, thereby electrically connecting the inner conductor of the one coaxial transmission line element to the outer conductor of its adjacent coaxial transmission line element.

20210125040, Apr 29, 2021

Oct 24, 2019

16/662532

- Armonk NY, US
Filipp A. Akopyan - New Windsor NY, US
Rathinakumar Appuswamy - San Jose CA, US
John V. Arthur - Mountain View CA, US
Pallab Datta - San Jose CA, US
Michael V. DeBole - Poughkeepsie NY, US
Steve K. Esser - San Jose CA, US
Myron D. Flickner - San Jose CA, US
Dharmendra S. Modha - San Jose CA, US
Carlos O. Otero - San Jose CA, US
Jun Sawada - Austin TX, US

G06N 3/063
G06N 3/04
G06N 5/04
G11C 11/54
G11C 11/34

Three-dimensional neural inference processing units are provided. A first tier comprises a plurality of neural cores. Each core comprises a neural computation unit. The neural computation unit is adapted to apply a plurality of synaptic weights to a plurality of input activations to produce a plurality of output activations. A second tier comprises a first neural network model memory adapted to store the plurality of synaptic weights. A communication network is operatively coupled to the first neural network model memory and to each of the plurality of neural cores, and adapted to provide the synaptic weights from the first neural network model memory to each of the plurality of neural cores.

20210110245, Apr 15, 2021

Oct 15, 2019

16/653366

- Armonk NY, US
Filipp A. Akopyan - New Windsor NY, US
Rathinakumar Appuswamy - San Jose CA, US
John V. Arthur - Mountain View CA, US
Andrew S. Cassidy - San Jose CA, US
Pallab Datta - San Jose CA, US
Steven K. Esser - San Jose CA, US
Myron D. Flickner - San Jose CA, US
Dharmendra S. Modha - San Jose CA, US
Tapan K. Nayak - San Jose CA, US
Carlos O. Otero - San Jose CA, US

G06N 3/063
G06N 3/04

Neural inference chips for computing neural activations are provided. In various embodiments, the neural inference chip is adapted to: receive an input activation tensor comprising a plurality of input activations; receive a weight tensor comprising a plurality of weights; Booth recode each of the plurality of weights into a plurality of Booth-coded weights, each Booth coded value having an order; multiply the input activation tensor by the Booth coded weights, yielding a plurality of results for each input activation, each of the plurality of results corresponding to the orders of the Booth-coded weights; for each order of the Booth-coded weights, sum the corresponding results, yielding a plurality of partial sums, one for each order; and compute a neural activation from a sum of the plurality of partial sums.

20150262055, Sep 17, 2015

Aug 14, 2012

13/585010

Filipp Akopyan - New Windsor NY, US
John V. Arthur - Mountain View CA, US
Rajit Manohar - Ithaca NY, US
Paul A. Merolla - Palo Alto CA, US
Dharmendra S. Modha - San Jose CA, US
Alyosha Molnar - Ithaca NY, US

Cornell University - Ithaca NY
International Business Machines Corporation - Armonk NY

G06N 3/063
G06N 3/04

An event-driven neural network includes a plurality of interconnected core circuits is provided. Each core circuit includes an electronic synapse array has multiple digital synapses interconnecting a plurality of digital electronic neurons. A synapse interconnects an axon of a pre-synaptic neuron with a dendrite of a post-synaptic neuron. A neuron integrates input spikes and generates a spike event in response to the integrated input spikes exceeding a threshold. Each core circuit also has a scheduler that receives a spike event and delivers the spike event to a selected axon in the synapse array based on a schedule for deterministic event delivery.