IP Library Granted Patent US 12699887
Granted Patent B2
US 12699887 · App. 15/908,780 · Granted Aug 4, 2026

Neuromorphic systems for computing characteristics of a set

Inventors: Arnon Amir (San Jose, CA); Tapan K. Nayak (San Jose, CA)
Assignee: International Business Machines Corporation
G06N3/063G06F7/544G06N3/049
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12699887
App. No.
15/908,780
Granted
Aug 4, 2026
Kind
B2
Abstract

Neurosynaptic systems for computing characteristics of a set are provided. In various embodiments, a plurality of encoders is provided. Each encoder is adapted to receive a population coded input and generate an encoded output. The encoded output comprises a plurality of segments. Each segment corresponds to one or more binary bits. A winner selection component is adapted to receive the encoded outputs from the plurality of encoders and to perform a method comprising: proceeding from highest order to lowest order of the segments of the encoded outputs of the encoders, performing a bitwise OR operation across all segments of equivalent order; disqualifying each encoded output whose bits do not match the result of the bitwise OR operation across all segments of equivalent order; outputting remaining encoded outputs.

Claims (34)

1 . A neurosynaptic circuit comprising:

a subnetwork configured to provide a plurality of encoded values, each encoded value comprising a plurality of segments, each segment corresponding to a plurality of binary bits, wherein each segment comprises a thermometer code;

a winner selection subnetwork, comprising a network layer for each of the plurality of binary bits, configured to receive the encoded values as input spikes and to perform a method comprising:

proceeding from highest order to lowest order of the segments of the encoded values of the subnetwork, performing a bitwise OR operation across all segments of equivalent order, wherein the bitwise OR operation comprises a BusOR operation for each set of binary bits to be ORed, the BusOR operation combining the set of binary bits together on one axon, the one axon being a wire of the neurosynaptic circuit;

disqualifying each encoded output whose bits do not match a result of the bitwise OR operation across all segments of equivalent order;

outputting output spikes corresponding to remaining encoded values.

2 . The neurosynaptic circuit of claim 1 , wherein the subnetwork is adapted to receive and relay the encoded values.

3 . The neurosynaptic circuit of claim 1 , wherein the subnetwork comprises at least one encoder adapted to receive and encode an input.

4 . The neurosynaptic circuit of claim 1 , wherein the subnetwork comprises a plurality of encoders, each encoder adapted to receive a population coded input and generate one of the encoded values therefrom.

5 . The neurosynaptic circuit of claim 1 , comprising at least one neurosynaptic core, neurosynaptic chip, simulated neurosynaptic core, or simulated neurosynaptic chip.

6 . The neurosynaptic circuit of claim 1 , wherein the thermometer code has a length of three.

7 . The neurosynaptic circuit of claim 1 , wherein the thermometer code has a length greater than three.

8 . The neurosynaptic circuit of claim 1 , wherein each segment comprises a binary code.

9 . The neurosynaptic circuit of claim 1 , wherein the method further comprises:

performing the bitwise OR operation of all segments of each encoded value to generate the result; and

disqualifying each encoded output value where the result is zero.

10 . A method of selecting an encoded value using neurosynaptic circuitry comprising:

providing a plurality of encoded values as input spikes, each encoded value comprising a plurality of segments, each segment corresponding to a plurality of binary bits, wherein each segment comprises a thermometer code;

proceeding from highest order to lowest order of the segments of the encoded values of encoders, performing a bitwise OR operation across all segments of equivalent order, wherein the bitwise OR operation comprises a BusOR operation for each set of binary bits to be ORed, the BusOR operation combining the set of binary bits together on one axon, the one axon being a wire of the neurosynaptic circuitry;

disqualifying each encoded output whose bits do not match a result of the bitwise OR operation across all segments of equivalent order;

outputting output spikes corresponding to remaining encoded values.

11 . The method of claim 10 , further comprising:

receiving a plurality of inputs; and

encoding the plurality of inputs to provide the encoded values.

12 . The method of claim 10 , further comprising:

receiving a plurality of population coded inputs; and

encoding the plurality of population coded inputs to provide the encoded values.

13 . The method of claim 10 , wherein the method is performed by at least one neurosynaptic core, neurosynaptic chip, simulated neurosynaptic core, or simulated neurosynaptic chip.

14 . The method of claim 10 , wherein the thermometer code has a length of three.

15 . The method of claim 10 , wherein the thermometer code has a length greater than three.

16 . The method of claim 10 , wherein each segment comprises a binary code.

17 . The method of claim 10 , wherein the method further comprises:

performing the bitwise OR operation of all segments of each encoded output to generate the result; and

disqualifying each encoded output where the result is zero.