Systems and methods for using federated learning for training centralized seizure detection and prediction models on decentralized datasets

Patent metadata

Field	Value
Publication number	US-12333443-B2
Application number	US-202117356342-A
Country	US
Kind code	B2
Filing date	Jun 23, 2021
Priority date	Jun 24, 2020
Publication date	Jun 17, 2025
Grant date	Jun 17, 2025

Abstract

A server for updating a current version of a machine learning model resident in implanted medical devices includes an interface, a memory, and a processor. The interface is configured to receive a plurality of updated versions of the machine learning model from a plurality of remote sources remote from the server. The remote source may be, e.g., implanted medical devices and/or subservers. The processor is coupled to the memory and the interface and is configured to aggregate the plurality of updated versions to derive a server-updated version of the machine learning model, and to transmit the server-updated version of the machine learning model to one or more of the plurality of remote sources as a replacement for the current version of the machine learning model.

First claim

What is claimed is:

1. A method of replacing a first machine learning model of a first architecture resident in a plurality of implanted medical devices, wherein the first machine learning model is generated using a first dataset and is configured to detect a neurological event in electrical activity of a brain, the method comprising:
providing, from a server to each of a plurality of remote sources remote from the server, information on a second dataset for generating a second machine learning model of a second architecture different than the first architecture, which second dataset includes at least one type of data that is not included in the first dataset and comprises records of electrical activity collected in response to detections of neurological events by the first machine learning model;
generating, at each of the plurality of remote sources, a version of the second machine learning model based on a corresponding second dataset;
receiving, at a server, a plurality of versions of the second machine learning model from the plurality of remote sources;
aggregating, at the server, the plurality of versions of the second machine learning model to derive a server-generated version of the second machine learning model; and
transmitting, at the server, the server-generated version of the second machine learning model to one or more of the plurality of remote sources as a replacement for the first machine learning model.

2. The method of claim 1 , wherein:
the second machine learning model comprises a neural network architecture having a plurality of nodes, and is characterized by a plurality of biases, each of the plurality of biases being associated with a corresponding node of the plurality of nodes, and
aggregating the plurality of versions of the second machine learning model comprises:
for at least one node of the plurality of nodes included in the plurality of versions of the second machine learning model, calculating an average of the biases associated with the at least one node, and
assigning the average to the at least one node.

3. The method of claim 2 , wherein aggregating the plurality of versions of the second machine learning model further comprises:
prior to calculating an average of the biases, applying a weight factor to each of the biases associated with the at least one node, wherein each weight factor is based on an amount of data included in the second dataset on which the version of the second machine learning model was trained.

4. The method of claim 1 , wherein:
the second machine learning model comprises a neural network architecture having a plurality of nodes and a plurality of interconnections between pair of nodes of the plurality of nodes, and is characterized by a plurality of weights, each weight of the plurality of weights being associated with a corresponding one of the plurality of interconnections, and
aggregating the plurality of versions of the second machine learning model comprises:
for at least one interconnection of the plurality of interconnections included in the plurality of versions of the second machine learning model, calculating an average of the weights associated with the at least one interconnection, and
assigning the average to the at least one interconnection.

5. The method of claim 4 , wherein aggregating the plurality of versions of the second machine learning model further comprises:
prior to calculating an average of the weights, applying a weight factor to each of the weights associated with the at least one interconnection, wherein each weight factor is based on an amount of data included in the second dataset on which the version of the second machine learning model was trained.

6. The method of claim 1 , wherein:
the second machine learning model comprises a neural network architecture having a plurality of nodes, and is characterized by a plurality of biases, each of the plurality of biases being associated with a corresponding node of the plurality of nodes, and
aggregating the plurality of versions of the second machine learning model comprises:
grouping the plurality of nodes into one or more sets of nodes based on one of probabilistic federated neural matching or federated matched averaging;
for at least one of the sets of nodes, calculating an average of the biases associated with the nodes in the at least one set of nodes; and
assigning the average to the nodes included in the at least one set of nodes.

7. The method of claim 1 , wherein:
the first machine learning model comprises a logistic regression having one or more parameters; and
the second machine learning model comprises one of a convolutional neural network (CNN); an autoencoder; and a recurrent neural network (RNN).

8. The method of claim 1 , wherein:
the plurality of remote sources comprises at least two of the plurality of implanted medical devices (IMD), and
the version of the second machine learning model generated by the at least two of the plurality of implanted medical devices is based on a second dataset stored in the respective implanted medical device.

9. The method of claim 8 , wherein generating a version of the second machine learning model comprises:
extracting features from a plurality of physiological records included in the second dataset; and
training the version of the second machine learning model on the extracted features.

10. The method of claim 9 , wherein each of the plurality of physiological records comprises:
electrical activity of a brain, and
at least one of neural tissue motion, heart rate, blood profusion, blood oxygenation, neurotransmitter concentrations, blood glucose, sweat hormones, body motion, and pH level.

11. The method of claim 9 , wherein each of the plurality of physiological records has a same tag, which identifies a common aspect among the plurality of physiological records, the common aspect corresponding to one of:
an occurrence of a neurological event;
absence of a neurological event; and
patient state.

12. The method of claim 8 , further comprising transmitting, at the at least two of the plurality of implanted medical devices, the IMD-generated version of the second machine learning model to the server.

13. The method of claim 8 , further comprising:
transmitting, at the at least two of the plurality of implanted medical devices, the IMD-generated version of the second machine learning model to a subserver remote from the server;
aggregating, at the subserver, the IMD-generated versions to derive a subserver-generated version of the second machine learning model; and
transmitting, at the one or more subservers, the subserver-generated version of the second machine learning model to the server, wherein the subserver-generated version corresponds to one of the plurality of versions of the second machine learning model aggregated at the server.

14. The method of claim 1 , wherein:
the plurality of remote sources comprises one or more subservers remote from the server,
the version of the second machine learning model generated by the one or more subservers is based on a second dataset received by the respective subserver from one or more of the plurality of implanted medical devices; and
and further comprising, transmitting, at the one or more subservers, the subserver-generated version of the second machine learning model to the server, wherein the subserver-generated version corresponds to one of the plurality of versions of the second machine learning model aggregated at the server.

15. The method of claim 14 , wherein generating a subserver-generated version comprises:
pooling,

Assignees

• Neuropace Inc

Inventors

• Arcot Desai Sharanya
• Tcheng Thomas K

Classifications

• G06N3/0895

  Weakly supervised learning, e.g. semi-supervised or self-supervised learning · CPC title

• G06N3/098 Primary

  Distributed learning, e.g. federated learning · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title