Tripleblind, Inc.

Triplelind Holdings, Inc.

Tripleblind Holdings, Inc.

Gharib Gharibi

Babak Poorebrahim Gilkalaye

Riddhiman Das

Greg Storm

Ravi Patel

Andrew James Rademacher

Praneeth Vepakomma

David Norman Wagner

Craig Gentry

Steve Penrod

Stephen Scott Penrod

A system and method are disclosed for providing a private multi-modal artificial intelligence platform. The method includes splitting a neural network into a first client-side network, a second client-side network and a server-side network and sending the first client-side network to a first client. The first client-side network processes first data from the first client, the first data having a first type. The method includes sending the second client-side network to a second client. The second client-side network processes second data from the second client, the second data having a second type. The first type and the second type have a common association. Forward and back propagation occurs between the client side networks and disparate data types on the different client side networks and the server-side network to train the neural network.

Systems and methods for providing a private multi-modal artificial intelligence platform

A system and method are disclosed for training a recommendation system. The method includes initiating, at a server device, an item-vector matrix V, wherein the item-vector matrix V includes a value m related to a total number of items across one or more client devices and a value d representing a hidden dimension, transmitting the item-vector matrix V to each client device, wherein each client device trains a local matrix factorization model using a respective user vector U and the item-vector matrix V to generate a respective set of gradients on each respective client device, receiving, via a secure multi-party compute protocol, and from each client device, the respective set of gradients, updating the item-vector matrix V using the respective set of gradients from each client device to generate an updated item-vector matrix V and downloading the updated item-vector matrix V to at least one client device.

Systems and methods for privacy preserving training and inference of decentralized recommendation systems from decentralized data

A system and method are disclosed for providing an averaging of models for federated learning and blind learning systems. The method includes selecting, at a server, a generator g and a number p, transmitting, to at least two n client devices, the generator g and the number p, receiving, from each client device i of the at least two client devices, a respective value k=gmod p and transmitting the set of respective values kto each client device i of the at least two client devices where respective added group of shares are generated on each client device i. The method includes receiving each respective added group of shares from each client device i of the at least two client devices and adding all the respective added group of shares to make a global sum of shares and dividing the global sum of shares by n.

Systems and methods for secure averaging of models for federated learning and blind learning using secure multi-party computation

Disclosed is a process for testing a suspect model to determine whether it was derived from a source model. An example method includes receiving, from a model owner node, a source model and a fingerprint associated with the source model, receiving a suspect model at a service node, based on a request to test the suspect model, applying the fingerprint to the suspect model to generate an output and, when the output has an accuracy that is equal to or greater than a threshold, determining that the suspect model is derived from the source model. Imperceptible noise can be used to generate the fingerprint which can cause predictable outputs from the source model and a potential derivative thereof.

Systems and methods for providing a systemic error in artificial intelligence algorithms

Systems and methods for blind multimodal learning

Disclosed is a method that includes training, at a client, a part of a deep learning network up to a split layer of the client. Based on an output of the split layer, the method includes completing, at a server, training of the deep learning network by forward propagating the output received at a split layer of the server to a last layer of the server. The server calculates a weighted loss function for the client at the last layer and stores the calculated loss function. After each respective client of a plurality of clients has a respective loss function stored, the server averages the plurality of respective weighted client loss functions and back propagates gradients based on the average loss value from the last layer of the server to the split layer of the server and transmits just the server split layer gradients to the respective clients.

Systems and methods for providing a modified loss function in federated-split learning

A system and method for securely computing an inference of two types of tree-based models, namely XGBoost and Random Forest, using secure multi-party computation protocol. The method includes computing a respective comparison result of each respective node of a plurality of nodes in a tree classifier. Each node has a respective threshold value. The respective comparison result is based on respective data associated with a data owner device being applied to a respective node having the respective threshold value. The method includes computing, based on the respective comparison result, a leaf value associated with the tree classifier, generating a share of the leaf value and transmitting, to the data owner device, a share of the leaf value. The data owner device computes, using a secure multi-party computation and between the model owner device and the data owner device, the leaf value for the respective data of the data owner.

Systems and methods for tree-based model inference using multi-party computation

A system and method are disclosed for secure multi-party computations. The system performs operations including establishing an API for coordinating joint operations between a first access point and a second access point related to performing a secure prediction task in which the first access point and the second access point will perform private computation of first data and second data without the parties having access to each other's data. The operations include storing a list of assets representing metadata about the first data and the second data, receiving a selection of the second data for use with the first data, managing an authentication and authorization of communications between the first access point and the second access point and performing the secure prediction task using the second data operating on the first data.

Systems and methods for providing a multi-party computation system for neural networks

A method of providing blind vertical learning includes creating, based on assembled data, a neural network having n bottom portions and a top portion and transmitting each bottom portion of then bottom portions to a client device. The training of the neural network includes accepting a, output from each bottom portion of the neural network, joining the plurality of outputs at a fusion layer, passing the fused outputs to the top portion of the neural network, carrying out a forward propagation step at the top portion of neural network, calculating a loss value after the forward propagation step, calculating a set of gradients of the loss value with respect to server-side model parameters and passing subsets of the set of gradients to a client device. After training, the method includes combining the trained bottom portion from each client device into a combined model.

Systems and methods for blind vertical learning

Disclosed is a method for each party of a group of m parties to be able to learn an Nth smallest value in a combined list. The method includes providing a value Rto a group of members; computing how many numbers are smaller than Rin a respective list of values for each respective member of the group of members; computing, a total number of smaller values (P); identifying a position of Rin a combined list of values comprising each respective list of values; when N=P+1, returning R; when N is greater than P+1, removing all values smaller than Rin their respective list of values and setting N=N&#x2212;(P+1); when N is less than P+1, removing all numbers bigger than Rin their respective list of value; and setting i=i+1.

Systems and methods for finding a value in a combined list of private values

A system and method are disclosed for each party of a group of m parties to be able to learn an Nth smallest value in a combined list of the values in which each party has separate lists of values. A method includes creating, by each party of a group of m parties, m lists of additive shares associated with each party's respective list of data, distributing, from each party to each other party in the group of m parties, m&#x2212;1 of the lists of additive shares to yield a respective combined list of additive shares Wobtained by each party of the m parties, receiving from a trusted party a list of additive shares Vassociated with a hot-code vector V, computing, in a shared space by each party, a respective Rvalue using a secure multiplication protocol and comparing, in the shared space, by each party and using secure multi-party comparison protocol, the respective Rto all elements in the respective combined list of additive shares Wto yield a total number Pof values in Wthat are smaller than R. The value Pis used to either end the method or loop back for further processing with new values of Wand in some cases a new value of N.

A system and method are disclosed for providing a privacy-preserving training approach for split learning methods, including blind learning. A method includes receiving, at a server device, encrypted smashed data from a client device, using a secure enclave on the server device, moving, on the server device, a server model, the encrypted smashed data and computer code for a blind learning operation into the secure enclave, performing, in the secure enclave, forward propagation using the decrypted smashed data to generate predicted values, comparing the predicted values to true labels using a loss function to yield a loss value, repeating the comparing step for all smashed data received at the server device from a plurality of clients to yield a plurality of loss values, averaging the plurality of loss values to yield an average loss value, updating model weights of the server model using the average loss value to yield gradients of the smashed data and transmitting the gradients of the smashed data to the client device.

Growing community of inventors

Lenexa, KS, United States of America

Gharib Gharibi