Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support secure training of machine learning (ML) models that preserves privacy in untrusted environments using distributed executable file packages. The executable file packages may include files, libraries, scripts, and the like that enable a cloud service provider configured to provide ML model training based on non-encrypted data to also support homomorphic encryption of data and ML model training with one or more clients, particularly for a diagnosis prediction model trained using medical data. Because the training is based on encrypted client data, private client data such as patient medical data may be used to train the diagnosis prediction model without exposing the client data to the cloud service provider or others. Using homomorphic encryption enables training of the diagnosis prediction model using encrypted data without requiring decryption prior to training.

Abstract: An Artificial Intelligence (AI) based filter apparatus includes an input filter and an output filter protecting a generative AI model and preventing restricted content from being transmitted to user devices. When a user query is received, the input filter determines if the user query can be transmitted to the generative AI model by generating an input risk score for the received user query. If the user query is transmitted and a model query response is received from the generative AI model, the output filter determines an output risk score based on which the model query response may be transmitted to the user. The input filter and the output filter each include a pre-trained language model as a base with additional layers trained to estimate the corresponding risk scores.

Abstract: A device may receive digital content and may process the digital content, with at least one of an optimization-based poisoning model or a statistical-based poisoning model, to generate at least one of first poisoning data or second poisoning data, respectively. The device may generate new digital content based on the digital content and the at least one of the first poisoning data or the second poisoning data. The device may provide the new digital content to one or more devices to be accessed by at least one deepfake model used to create fake digital content and may perform one or more actions based on the new digital content.

Abstract: A code translation apparatus receives a source code including one or more code vulnerabilities and automatically generates remediated code. The source code provided to the code translation apparatus is converted to a source directional graph. The edges of the source directional graph are augmented with additional edge attributes. The source directional graph thus augmented is further converted into a source graph vector representation. The source graph vector representation is provided to an encoder of a trained code transformer. The remediated code is obtained from the decoder of the trained code transformer.

Abstract: Systems and methods for defending an artificial intelligence model against an adversarial input are disclosed. The system may include an artificial intelligence model, such as a machine learning model. The system may include a transformation engine executable by one or more processors. The transformation engine may be configured to receive an input to the artificial intelligence model, and apply a pre-determined transformation set to the input to produce a transformed input. The transformation engine may be configured to generate a first output based on the input using the artificial intelligence model and may also apply the artificial intelligence model to the transformed input to produce a second output. The transformation engine may be configured to determine whether the input is associated with an adversarial attack based on a comparison of the first output and the second output. The system also facilitates generating transformation sets for defending against adversarial attacks.

Abstract: In some examples, source code differential pruning-based dataset creation may include receiving source code that includes at least one vulnerability and at least one remediation that remediates the at least one vulnerability, extracting at least one remediated section, and identifying each sentence of the remediated section. A plurality of clusters may be generated based on an analysis of each identified sentence of the remediated section to determine a score with respect to a specified cluster that includes the identified sentence. Further, a determination may be made as to whether the score is greater than a specified threshold. Each identified sentence for which the score is greater than the specified threshold may be designated as a relevant sentence. An auxiliary dataset may be generated based on a plurality of relevant sentences and include at least one relevant vulnerability and at least one relevant remediation that remediates the relevant vulnerability.

Abstract: Methods, systems and apparatus, including computer programs encoded on computer storage medium, for implementation of secret superposition protocols. In one aspect a method includes, performing, by a sender party, quantum operations on one or more qubits, comprising preparing, according to a predetermined secret superposition protocol, one or more qubits in respective uniform superposition quantum states; transmitting, by the sender party, to a recipient party, and through a secure channel, data indicating use of the predetermined secret superposition protocol; and transmitting, by the sender party and to the recipient party, one or more of the qubits, to wherein the recipient party performs one or more measurements on the qubits to verify use of the predetermined secret superposition protocol.

Abstract: In some examples, scalable source code vulnerability remediation may include receiving source code that includes at least one vulnerability, and receiving remediated code that remediates the at least one vulnerability associated with the source code. At least one machine learning model may be trained to analyze a vulnerable code snippet of the source code. The vulnerable code snippet may correspond to the at least one vulnerability associated with the source code. The machine learning model may be trained to generate, for the vulnerable code snippet, a remediated code snippet to remediate the at least one vulnerability associated with the source code. The remediated code snippet may be validated based on an analysis of whether the remediated code snippet remediates the at least one vulnerability associated with the source code.

Abstract: A code remediation system accesses a programming code including vulnerabilities such as potential secrets and remediates at least a subset of the potential secrets to generate modified programming code wherein the subset of potential secrets which are determined to be actual secrets are replaced with access mechanisms to storage locations on a vault wherein the actual secrets are secured. To identify the subset of potential secrets forming the actual secrets to be remediated, the code remediation system is configured to filter out false positives among the potential secrets and identify true positives. When an application executing the modified code encounters an access mechanism, it accesses the vault to retrieve the actual secrets.

Abstract: Methods, systems, and apparatus for transmitting qubits encoding quantum information with reduced risk of interception from an eavesdropper. In one aspect, a method includes encoding quantum information into an information qubit; encrypting the information qubit, comprising performing i) a parity operation on the information qubit and a parity control qubit and ii) a phase operation on the information qubit and a phase control qubit; performing, by a sender party, a sequence of one or more quantum logic gates on the phase control qubit; sending the information qubit, parity control qubit, and phase control qubit to a recipient party; and sending data identifying the sequence of one or more quantum logic gates to the recipient party, wherein the recipient party obtains the quantum information encoded into the information qubit using the information qubit, parity control qubit, phase control qubit, and data identifying the sequence of one or more quantum logic gates.

Abstract: A device may receive a machine learning model and training data utilized to train the machine learning model, and may perform a data veracity assessment of the training data to identify and remove poisoned data from the training data. The device may perform an adversarial assessment of the machine learning model to generate adversarial attacks and to provide defensive capabilities for the adversarial attacks, and may perform a membership inference assessment of the machine learning model to generate membership inference attacks and to provide secure training data as a defense for the membership inference attacks. The device may perform a model extraction assessment of the machine learning model to identify model extraction vulnerabilities and to provide a secure application programming interface as a defense to the model extraction vulnerabilities, and may perform actions based on results of one or more of the assessments.

Abstract: A device may receive content data identifying content created by users and metadata associated with the content. The device may receive rules data identifying rules associated with utilization of the content. The device may utilize the metadata to generate digital DNA signatures for the content in near-real time. The device may store, in a repository, the rules data, the content, the digital DNA signatures, and relationships between the digital DNA signatures. The device may receive, from a client device, new content that is generated based on particular content of the content data and new metadata associated with the new content. The device may utilize the new metadata to generate a new digital DNA signature for the new content. The device may process the new digital DNA signature, the rules data, and the digital DNA signatures to determine whether the new content violates one or more rules of the rules data.

Abstract: Methods, systems and apparatus, including computer programs encoded on computer storage medium, for implementation of secret superposition protocols. In one aspect a method includes, performing, by a sender party, quantum operations on one or more qubits, comprising preparing, according to a predetermined secret superposition protocol, one or more qubits in respective uniform superposition quantum states; transmitting, by the sender party, to a recipient party, and through a secure channel, data indicating use of the predetermined secret superposition protocol; and transmitting, by the sender party and to the recipient party, one or more of the qubits, wherein the recipient party performs one or more measurements on the qubits to verify use of the predetermined secret superposition protocol.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support secure training of machine learning (ML) models that preserves privacy in untrusted environments using distributed executable file packages. The executable file packages may include files, libraries, scripts, and the like that enable a cloud service provider configured to provide ML model training based on non-encrypted data to also support homomorphic encryption of data and ML model training with one or more clients, particularly for a diagnosis prediction model trained using medical data. Because the training is based on encrypted client data, private client data such as patient medical data may be used to train the diagnosis prediction model without exposing the client data to the cloud service provider or others. Using homomorphic encryption enables training of the diagnosis prediction model using encrypted data without requiring decryption prior to training.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support cooperative training of machine learning (ML) models that preserves privacy in untrusted environments using distributed executable file packages. The executable file packages may include files, libraries, scripts, and the like that enable a cloud service provider configured to provide server-side ML model training to also support cooperative ML model training with multiple clients, particularly for a fraud prediction model for financial transactions. Because the cooperative training includes the clients training respective ML models and the server aggregating the trained ML models, private client data such as financial transaction data may be used to train the fraud prediction model without exposing the client data to others. Such cooperative ML model training enables offloading of computing resource-intensive training from client devices to the server and may train a more robust fraud detection model.

Abstract: Methods, systems and apparatus, including computer programs encoded on computer storage medium, for implementation of secret superposition protocols. In one aspect a method includes, performing, by a sender party, quantum operations on one or more qubits, comprising preparing, according to a predetermined secret superposition protocol, one or more qubits in respective uniform superposition quantum states; transmitting, by the sender party, to a recipient party, and through a secure channel, data indicating use of the predetermined secret superposition protocol; and transmitting, by the sender party and to the recipient party, one or more of the qubits, wherein the recipient party performs one or more measurements on the qubits to verify use of the predetermined secret superposition protocol.

Abstract: Methods, systems, and apparatus for transmitting qubits encoding quantum information with reduced risk of interception from an eavesdropper. In one aspect, a method includes encoding quantum information into an information qubit; encrypting the information qubit, comprising performing i) a parity operation on the information qubit and a parity control qubit and ii) a phase operation on the information qubit and a phase control qubit; performing, by a sender party, a sequence of one or more quantum logic gates on the phase control qubit; sending the information qubit, parity control qubit, and phase control qubit to a recipient party; and sending data identifying the sequence of one or more quantum logic gates to the recipient party, wherein the recipient party obtains the quantum information encoded into the information qubit using the information qubit, parity control qubit, phase control qubit, and data identifying the sequence of one or more quantum logic gates.

Abstract: Methods, systems, and apparatus for transmitting qubits encoding quantum information with reduced risk of interception from an eavesdropper. In one aspect, a method includes encoding quantum information into an information qubit; encrypting the information qubit, comprising performing i) a parity operation on the information qubit and a parity control qubit and ii) a phase operation on the information qubit and a phase control qubit; performing, by a sender party, a sequence of one or more quantum logic gates on the phase control qubit; sending the information qubit, parity control qubit, and phase control qubit to a recipient party; and sending data identifying the sequence of one or more quantum logic gates to the recipient party, wherein the recipient party obtains the quantum information encoded into the information qubit using the information qubit, parity control qubit, phase control qubit, and data identifying the sequence of one or more quantum logic gates.

Abstract: Methods, systems, and apparatus for transmitting qubits encoding quantum information with reduced risk of interception from an eavesdropper. In one aspect, a method includes encoding quantum information into an information qubit; encrypting the information qubit, comprising performing i) a parity operation on the information qubit and a parity control qubit and ii) a phase operation on the information qubit and a phase control qubit; performing, by a sender party, a sequence of one or more quantum logic gates on the phase control qubit; sending the information qubit, parity control qubit, and phase control qubit to a recipient party; and sending data identifying the sequence of one or more quantum logic gates to the recipient party, wherein the recipient party obtains the quantum information encoded into the information qubit using the information qubit, parity control qubit, phase control qubit, and data identifying the sequence of one or more quantum logic gates.

Abstract: A device may receive content data identifying content created by users and metadata associated with the content. The device may receive rules data identifying rules associated with utilization of the content. The device may utilize the metadata to generate digital DNA signatures for the content in near-real time. The device may store, in a repository, the rules data, the content, the digital DNA signatures, and relationships between the digital DNA signatures. The device may receive, from a client device, new content that is generated based on particular content of the content data and new metadata associated with the new content. The device may utilize the new metadata to generate a new digital DNA signature for the new content. The device may process the new digital DNA signature, the rules data, and the digital DNA signatures to determine whether the new content violates one or more rules of the rules data.