QUBO Encoding for Cloud Optimization Data Privacy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud-based optimization services face challenges in protecting user data privacy, as conventional optimization solvers can recover sensitive information from inputs submitted via API calls, compromising confidentiality when solving combinatorial optimization problems.
Innovation Solution
A system that generates a QUBO formulation based on user inputs, encodes it with privacy settings, and submits it to a cloud-based optimization solver, ensuring the solver cannot differentiate between encoded and original data, thereby maintaining user data confidentiality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If cloud-based optimization solvers are used to solve combinatorial optimization problems, then computational intractability of NP problems is addressed, but user data privacy is compromised as solvers can recover sensitive information from API inputs
Solution Approach 1:
The patent introduces an encoding layer as an intermediary between the user's optimization problem and the cloud solver. The encoder transforms the original QUBO formulation into an encoded version that preserves the problem structure and solution validity while obscuring sensitive input data. The solver processes the encoded problem without accessing the original confidential information, thus enabling cloud-based computation while protecting user data privacy.
2Productivity
If QUBO formulations are submitted directly to cloud solvers via API calls, then optimization problems can be solved efficiently, but sensitive information in the inputs can be recovered by the solver
Solution Approach 1:
An encoding function serves as a mediator that transforms the original QUBO formulation (with sensitive data) into an encoded QUBO formulation. This encoded version maintains the mathematical structure necessary for the solver to find correct solutions while preventing the solver from recovering the original sensitive inputs. The encoding acts as a protective layer that enables efficient cloud-based solving without compromising confidentiality.
Solution Approach 2:
The patent applies parameter changes by transforming the QUBO formulation through encoding operations that modify the parameters (coefficients and variables) while preserving the problem's essential structure. The encoding transforms the original parameters into encoded parameters that are mathematically equivalent for solving purposes but opaque to anyone attempting to infer the original sensitive data from the solver's perspective.
3Loss of information
If encoding is applied to QUBO formulations before submission, then data privacy is protected, but system complexity increases due to encoding and decoding operations
Solution Approach 1:
The encoding and decoding functions serve as intermediaries that add privacy protection to the system. While these intermediaries do increase system complexity, they enable the use of cloud-based solvers with guaranteed data privacy. The complexity is concentrated in the encoding/decoding layers, while the core optimization solving remains efficient and unchanged.
Data Source
AI summary
According to an aspect of an embodiment, operations may include receiving a first input corresponding to a selection of a combinatorial optimization problem and receiving a set of datapoints as input. The operations may further include generating a first quadratic unconstrained binary optimization (QUBO) formulation based on an objective function for the combinatorial optimization problem and the received set of datapoints. The operations may further include selecting a first privacy setting and encoding the first QUBO formulation based on the selected privacy setting to generate a second QUBO formulation. The operations may further include submitting the generated second QUBO formulation to an optimization solver machine and receiving a first solution of the second QUBO formulation. The operations may further include decoding the first solution to produce a second solution and publishing an output of the combinatorial optimization problem on a user device based on the second solution.