US 2019/0095778 A1Published application

Biological organism development system

OwnerCharles L. Buchananinventor

Life of this patent

Estimating the term…

Loading patent…

Abstract

An organism development system includes: at least one neuromorphic device including multiple sets of hardware components that each store one or parameters of neuromorphic configuration data to implement a neuron of a neural network, wherein: the neural network is configured by neuromorphic configuration data to derive a proposed genome or epigenome of a new organism meant to have a sought-for trait, and the neuromorphic configuration data is generated by training the neural network with a usage data set that includes trait data indicative of a trait and biological data indicative of a genome for each of multiple organisms; a genome or epigenome printing device to print genetic/epigenetic material of the new organism based on the proposed genome or epigenome, respectively; and a trait detection device to detect an observed trait of the new organism following its at least its generation for incorporation back into the usage data set.

Claims26

Dependent claims

A processing device comprising:

storage configured to store a usage data set and trained neuromorphic configuration data, wherein:
the usage data set comprises multiple organism entries that each correspond to one of multiple organisms;
each organism entry comprises:
trait data indicative of at least one trait of the corresponding organism;
and
biological data indicative of a genome or an epigenome of the corresponding organism;
and
the trained neuromorphic configuration data comprises multiple trained parameters indicative of training of a neural network with at least a portion of the usage data set;
at least one neuromorphic device comprising multiple sets of hardware components, wherein:
each set of hardware components is configured to store at least one trained parameter of the multiple trained parameters to implement an artificial neuron of multiple artificial neurons of the neural network;
and
the neural network is configured by at least the portion of the trained neuromorphic configuration data to derive a proposed genome or a proposed epigenome of a new organism based on a sought-for trait provided to inputs of the at least one neuromorphic device;
and
a processor coupled to the storage and to the at least one neuromorphic device, wherein the processor is configured to:
train the neural network with at least the portion of the usage data set and generate the trained neuromorphic configuration data, wherein for each organism entry of the usage data set, the processor performs operations comprising:
provide the trait data to the inputs of the at least one neuromorphic device;
and
provide the biological data to outputs of the at least one neuromorphic device;
and
use the neural network to develop the new organism, wherein the processor performs operations comprising:
receive an indication of the sought-for trait that the new organism is meant to have from an input device coupled to the processor;
provide the sought-for trait to the inputs of the at least one neuromorphic device;
retrieve, from the outputs of the at least one neuromorphic device, the proposed genome or the proposed epigenome of the new organism derived by the neural network;
and
transmit the proposed genome or the proposed epigenome derived by the neural network to a printing device to enable generation of genetic or epigenetic material of the new organism.

22 more elements

An organism development system comprising:

at least one neuromorphic device comprising multiple sets of hardware components, wherein:
each set of hardware components is configured to store at least one trained parameter of trained neuromorphic configuration data to implement an artificial neuron of a neural network;
the neural network is configured by the trained neuromorphic configuration data to derive and provide at outputs of the at least one neuromorphic device a proposed genome or a proposed epigenome of a new organism that is meant to have a sought-for trait provided to inputs of the at least one neuromorphic device;
the trained neuromorphic configuration data is generated by the neural network during training of the neural network with at least a portion of a usage data set, wherein:
the usage data set comprises multiple organism entries;
and
each organism entry comprises trait data indicative of at least one trait of a corresponding organism, and biological data indicative of a genome or an epigenome of the corresponding organism;
a genome/epigenome printing device configured to print genetic or epigenetic material of the new organism to enable generation of the new organism based on the proposed genome or epigenome, respectively;
and
a trait detection device configured to detect an observed trait of the new organism following at least the generation of the new organism.

7 more elements

A computer-implemented method comprising:

receiving, at a processor, an indication of a sought-for trait of a new organism from an input device;
providing the sought-for trait to inputs of at least one neuromorphic device coupled to the processor, wherein:
the at least one neuromorphic device comprises multiple sets of hardware components;
each set of hardware components is configured to store at least one trained parameter of multiple trained parameters of trained neuromorphic configuration data to implement an artificial neuron of multiple artificial neurons of a neural network;
and
the neural network is configured by at least a portion of the trained neuromorphic configuration data to derive and provide at outputs of the at least one neuromorphic device a proposed genome or a proposed epigenome of the new organism based on the sought-for trait provided to inputs;
retrieving, from the outputs of the at least one neuromorphic device, the proposed genome or the proposed epigenome of the new organism derived by the neural network;
transmitting the proposed genome or the proposed epigenome derived by the neural network to a printing device to enable generation of genetic or epigenetic material of the new organism;
generating a new organism entry in a usage data set, wherein:
the usage data set comprises multiple organism entries;
and
each organism entry comprises trait data indicative of at least one trait of a corresponding organism, and biological data indicative of a genome or an epigenome of the corresponding organism;
following generation and cultivation of the new organism, operating a trait detection device to detect an observed trait of the new organism or of a derivative of the new organism;
storing an indication of the observed trait as the trait data within the new organism within the new organism entry;
and
using at least the new entry to further train the neural network.

13 more elements

Dependencies are inferred from the claim text.

Facts

Application: 16/202,331Filed 28 Nov 2018
Priority: 27 Dec 2013
Published: 28 Mar 2019
Term ends: Not estimatedThe source records don’t support an estimate.

Classification10 codes

G06N 3/08
Neural networks: learning methods
G06N 3/0499
Architecture, e.g. interconnection topology: feedforward networks
G06N 3/065
Using electronic means: analogue means
G06N 3/09
Learning methods: supervised learning

Family

Application61/921,174Us provisional application · filed 27 Dec 2013
Application62/013,736Us provisional application · filed 18 Jun 2014
Application14/584,154Continuation in part · filed 29 Dec 2014
Application16/038,592Continuation in part · filed 18 Jul 2018
Application62/757,673Us provisional application · filed 8 Nov 2018
Application16/202,331Continuation in part
US 2015/0186597 A1Published application
US 2018/0322241 A1Published application
US 2019/0095778 A1This patent

From the US filing history. Foreign family members aren’t listed.