Techniques for providing secure federated machine-learning
Embodiments of the invention are directed to systems, methods, and devices for securely performing federated tasks (e.g., the generation and utilizing of machine-learning models). A secure platform computer may operate a secure memory space. Entities participating in a federated project may transmit respective portions of project data defining the federated project. Each entity may provide their respective (encrypted) data sets for the project that in turn can be used to generate a machine-learning model in accordance with the project data. The machine-learning model may be stored in the secure memory space and accessed through an interface provided by the secure platform computer Utilizing the techniques discussed herein, a machine-learning models may be generated and access to these models may be restricted while protect each participant's data set from being exposed to the other project participants.
1 . A computing device, comprising:
one or more processors;
a secure platform computer operating a secure memory space, the secure memory space comprising computer-executable instructions that, when executed by the one or more processors, causes the secure platform computer to:
receive a first portion and a second portion of project data defining a federated project that comprises performing a machine-learning task utilizing data contributed by multiple participating entities, wherein respective data contributed during the federated project by each participating entity is kept private from all other participating entities, the first portion of the project data corresponding to a first participating entity and the second portion of the project data corresponding to a second participating entity;
receive a first data set associated with the first participating entity and a second data set associated with the second participating entity;
maintain, within the secure memory space of the computing device, an immutable ledger for the federated project, the immutable ledger comprising the first data set and the second data set by the secure platform computer, the first data set being inaccessible to the second participating entity while stored, the second data set being inaccessible to the first participating entity while stored;
generate a machine-learning model based at least in part on the first data set corresponding to the first participating entity, the second data set corresponding to the second participating entity, and the project data; and
provide access to the machine-learning model to the first participating entity and the second participating entity, the machine-learning model being stored within the secure memory space.
2 . The computing device of claim 1 , wherein the secure memory space is an enclave managed by a chip set of the one or more processors, wherein the enclave is encrypted by the chip set and accessible only by the chip set.
3 . The computing device of claim 1 , wherein executing the computer-executable instructions further causes the secure platform computer to:
receive, from a requesting entity, a task request requesting the immutable ledger be verified;
compare a first identifier of the requesting entity to a second identifier of the project data;
when the first identifier matches the second identifier, verifying the immutable ledger; and
providing, to the requesting entity, an indication that the immutable ledger was verified.
4 . The computing device of claim 3 , wherein the immutable ledger further comprises an identifier for a provider of the project data, and wherein executing the computer-executable instructions further causes the secure platform computer to verify that the provider of the project data corresponds to the first participating entity or the second participating entity.
5 . The computing device of claim 3 , wherein executing the computer-executable instructions further causes the secure platform computer to generate one or more first keys for retrieving the first data set from the secure memory space and one or more second keys for retrieving the second data set from the secure memory space.
6 . The computing device of claim 5 , wherein information maintained in the immutable ledger comprises first hash values corresponding to the one or more first keys and second hash values corresponding to the one or more second keys.
7 . The computing device of claim 1 , wherein executing the computer-executable instructions further causes the secure platform computer to:
generate an additional machine-learning model based at least in part on the first data set, the second data set, and the project data; and
provide access to the additional machine-learning model for the first participating entity, the additional machine-learning model being stored within the secure memory space.
8 . The computing device of claim 1 , wherein executing the computer-executable instructions further causes the secure platform computer to:
transmit a notification to the second participating entity in response to receiving the first portion of the project data, the notification being transmitted prior to receiving the second portion of the project data, wherein the notification indicates receipt of the first portion of the project data.
9 . The computing device of claim 1 , wherein the first data set is inaccessible by the second participating entity and wherein the second data set is inaccessible by the first participating entity.
10 . The computing device of claim 1 , wherein executing the computer-executable instructions further causes the secure platform computer to transmit a notification to the first participating entity and the second participating entity, wherein the notification indicates generation of the machine-learning model is complete.
11 . The computing device of claim 1 , wherein executing the computer-executable instructions further causes the secure platform computer to:
receive, from a corresponding computing device of the first participating entity, a task request indicative of a request to update the machine-learning model;
update the machine-learning model utilizing a data set comprising the first data set, the second data set, and subsequent data received after receipt of the first data set and the second data set; and
transmit, to the computing device of the first participating entity, a notification indicating the task request was successful or unsuccessful.
12 . The computing device of claim 1 , wherein the immutable ledger further stores the project data, a task request associated with the federated project, and an identifier of an entity that initiated the task request.
13 . A computing device, comprising:
one or more processors; and
a secure platform computer operating a secure memory space, the secure memory space comprising computer-executable instructions that, when executed by the one or more processors, causes the secure platform computer to:
receive a first portion and a second portion of project data defining a federated project that comprises performing a machine-learning task utilizing data contributed by multiple participating entities, wherein respective data contributed during the federated project by each participating entity is kept private from all other participating entities, the first portion of the project data corresponding to a first participating entity and the second portion of the project data corresponding to a second participating entity;
receive a first data set associated with the first participating entity and a second data set associated with the second participating entity;
generate a machine-learning model based at least in part on the first data set corresponding to the first participating entity, the second data set corresponding to the second participating entity, and the project data;
provide access to the machine-learning model to the first participating entity and the second participating entity, the machine-learning model being stored within the secure memory space.
14 . The computing device of claim 13 , wherein executing the computer-executable instructions further causes the secure platform computer to:
maintain, in the secure memory space, an immutable ledger comprising an identifier for a provider of the project data, and
verify that the provider of the project data corresponds to the first participating entity or the second participating entity.
15 . The computing device of claim 13 , wherein executing the computer-executable instructions further causes the secure platform computer to generate one or more first keys for retrieving the first data set from the secure memory space and one or more second keys for retrieving the second data set from the secure memory space.
16 . The computing device of claim 13 , wherein executing the computer-executable instructions further causes the secure platform computer to:
generate an additional machine-learning model based at least in part on the first data set, the second data set, and the project data; and
provide access to the additional machine-learning model for the first participating entity, the additional machine-learning model being stored within the secure memory space.
17 . The computing device of claim 13 , wherein executing the computer-executable instructions further causes the secure platform computer to:
transmit a notification to the second participating entity in response to receiving the first portion of the project data, the notification being transmitted prior to receiving the second portion of the project data, wherein the notification indicates receipt of the first portion of the project data.
18 . The computing device of claim 13 , wherein the first data set is inaccessible by the second participating entity and wherein the second data set is inaccessible by the first participating entity.
19 . The computing device of claim 13 , wherein executing the computer-executable instructions further causes the secure platform computer to:
receive, from a corresponding computing device of the first participating entity, a task request indicative of a request to update the machine-learning model;
update the machine-learning model utilizing a data set comprising the first data set, the second data set, and subsequent data received after receipt of the first data set and the second data set; and
transmit, to the computing device of the first participating entity, a notification indicating the task request was successful or unsuccessful.
20 . The computing device of claim 13 , wherein executing the computer-executable instructions further causes the secure platform computer to maintain, in an immutable ledger of the secure memory space, the project data, a task request associated with the federated project, and an identifier of an entity that initiated the task request.