57 results about "Adversary" patented technology

The invention relates to a system and a method for privacy preservation of sensitive attributes stored in a database. The invention reduces the complexity and enhances privacy preservation of the database by determining the distribution of sensitive data based on Kurtosis measurement. The invention further determines and compares the optimal value of k-sensitive attributes in k-anonymity data sanitization model with the optimal value of l sensitive attributes in l diversity data sanitization model using adversary information gain. The invention reduces the complexity of the method for preserving privacy by applying k anonymity only, when the distribution of the sensitive data is leptokurtic and optimal value of k is greater than the optimal value of l.

Methods and apparatus are provided for securing two-party computations against malicious adversaries. A method is provided for secure function evaluation. The disclosed method is performed by a garbled circuit evaluator for the transfer of private information, and comprises receiving from a constructor (i) s garbled circuits (GCs), wherein each of the GCs having a plurality of input wires; and (ii) commitments for each of the input wires, wherein the commitments comprise s² pair-wise cryptographic bindings of wire garblings of each given wire in the s GCs; requesting the constructor to reveal a selected check-set of s/2 of the s GCs; and verifying that the check-set was properly constructed using less than all of the commitments. In addition, the disclosed method optionally comprises the step of evaluating the remaining GCs that were not in the check-set.

The invention relates to an authenticable asymmetrical group secret key negotiation scheme in a mobile unbalanced network. The scheme comprises the steps that signature and authentication of group members are carried out, group members negotiate an asymmetrical shared secret key, and consistency verification of the shared secret key is carried out. According to the signature and authentication, a short signature mechanism of mobile equipment in a network is achieved by adopting a bilinear mapping technology, and members of mobile groups provide identify authentication by using the signature before the group secret key is negotiated to defense active attack of adversaries. According to step that the group members negotiate the asymmetrical shared secret key, all mobile equipment in the mobile unbalanced network negotiates a pair of asymmetrical group communication encryption/decryption secret keys, and members outside the group can send secret messages to members inside the group by encrypting the secret messages through a public key without the necessary of joining the group. According to the consistency verification of the shared secret key, after the group members negotiate the group secret key, the accuracy and the consistency of the calculated group secret key can be guaranteed. According to the authenticable asymmetrical group secret key negotiation scheme in the mobile unbalanced network, when secret communication is carried out among mobile group equipment in the mobile network environment, safe communication among the group equipment can be guaranteed, energy consumption of calculation and communication is quite low, and good flexibility, safety and practicability are achieved.

A method and apparatus for repeated communication sessions between a sender (e.g., RFID tag) and a receiver (RFID reader) that employs a proactive information security scheme is based on the assumption that the information exchanged during at least one of every n successive communication sessions is not exposed to an adversary. The sender and the receiver maintain a vector of n entries that is repeatedly refreshed by pairwise XORING entries, with a new vector of n entries that is randomly chosen by the sender and sent to the receiver as a part of each communication session. Also, a computational secure scheme based on the information secure scheme is employed to ensure that even in the case that the adversary listens to all the information exchanges, the communication between the sender and the receiver is secure. In particular, the scheme can be used in the domain of remote controls (e.g., for cars).

Identifying cyber adversary behavior on a computer network is provided. Individual security events are received from multiple threat intelligence data sources. A security incident corresponding to an attack on at least one element of the computer network, the security incident being described by the individual security events received from the multiple threat intelligence data sources, is matched to a defined cyber adversary objective in a structured framework of a plurality of defined cyber adversary objectives and a related technique associated with the defined cyber adversary objective used by a cyber adversary in the attack. A set of mitigation actions is performed on the computer network based on matching the security incident corresponding to the attack on the computer network to the defined cyber adversary objective and the related technique.

The invention discloses a trust negotiation building method based on history roles, which comprises the following steps that: (1) when a request party submits requests on a certain resource to a resource party, firstly, a chinchin stage of the trust negotiation is triggered, and the identities of the two parties and whether the historical negotiation success record information exists or not are verified; (2) after the chinchin stage is completed, the two parties enter a trust development period, a safe certificate disclosing sequence is found out to visit the requested resources through mutually disclosing visit control tactics and certificates; and (3) the two parties enter the trust maintenance stage after the trust negotiation is successful, relevant information of the other negotiation party in the current negotiation is stored for accelerating the subsequent negotiation process of the two parties. The method provided by the invention has the advantages that the successfully negotiated two parties update the roles for the other parties in self history information databases and record and obtain the certificate disclosing sequences of the two parties of the role. When the two parties negotiate again, the negotiation does not need to be carried out again, the disclosing can be directly carried out according to the certificate disclosing sequences of the role, the subsequent negotiation of the two parties is shortened, and the certificate collection activities of venomous negotiation adversaries is avoided.

A technique for providing cyber resilience by integrating autonomous adversary and defender agents, deep learning, and graph thinking. An automated competitive environment of autonomous adversary and defender agents is provided such that the adversary agent can emulate the adversary activities, patterns, and intentions using all available cybersecurity measurements and observations, and, the defender agent can generate and suggest the best possible appropriate actions to mitigate or prevent adversary activities while recovering or protecting assets. An automated cyber resilience system with autonomous agents is provided using machine learning and security analytics to first predict the current and future adversary activities and then provide an automated critical asset protection and recovery by enabling agents to take appropriate reactive and pro-active actions at each time step to prevent, recover, or mitigate adversary activities over enterprise and tactical networks.

A method, apparatus and computer program product to defend learning models that are vulnerable to adversarial example attack. It is assumed that data (a “dataset”) is available in multiple modalities (e.g., text and images, audio and images in video, etc.). The defense approach herein is premised on the recognition that the correlations between the different modalities for the same entity can be exploited to defend against such attacks, as it is not realistic for an adversary to attack multiple modalities. To this end, according to this technique, adversarial samples are identified and rejected if the features from one (the attacked) modality are determined to be sufficiently far away from those of another un-attacked modality for the same entity. In other words, the approach herein leverages the consistency between multiple modalities in the data to defend against adversarial attacks on one modality.

Techniques for providing innocent until proven guilty (IUPG) solutions for building and using adversary resistant and false positive resistant deep learning models are disclosed. In some embodiments, a system, process, and/or computer program product includes storing a set comprising one or more innocent until proven guilty (IUPG) models for static analysis of a sample; performing a static analysis of content associated with the sample, wherein performing the static analysis includes using at least one stored IUPG model; and determining that the sample is malicious based at least in part on the static analysis of the content associated with the sample, and in response to determining that the sample is malicious, performing an action based on a security policy.

A method, apparatus and computer program product to protect a deep neural network (DNN) having a plurality of layers including one or more intermediate layers. In this approach, a training data set is received. During training of the DNN using the received training data set, a representation of activations associated with an intermediate layer is recorded. For at least one or more of the representations, a separate classifier (model) is trained. The classifiers, collectively, are used to train an outlier detection model. Following training, the outliner detection model is used to detect an adversarial input on the deep neural network. The outlier detection model generates a prediction, and an indicator whether a given input is the adversarial input. According to a further aspect, an action is taken to protect a deployed system associated with the DNN in response to detection of the adversary input.

The invention discloses a certificate-free signature method based on a national secret algorithm. At the time of signature generation, B calculates the parameters participate in signature generation according to the secret parameters randomly selected to participate in signature generation, and send the parameters to A. A randomly selects the secret parameters involved in signature verification, and generates a message combining the secret parameters to send to B. During signature verification, B verifies whether A's signature public key is valid according to the secret parameters generated byparticipating signatures. If it is invalid, the verification fails, otherwise, the verification continues. B calculates a parameter and send it to A. A verifies whether B completes the signature verification according to the correct steps through this parameter, and sends a parameter containing the secret value if B conforms to the correct execution step, otherwise, terminates the signature. B receives the verification parameters of A and participates in the subsequent signature verification, and finally produces the result. The invention can prevent an adversary from controlling one of the parties to seek personal gain in the process of signature and verification in which the server cooperates with the mobile terminal.

In a method of security detection for a physical layer authentication a transmitter transmits a first signal, which, after passing over a wireless fading channel becomes a second signal. An active adversary receives the second signal and determines a first false alarm probability based. The adversary may determine a first optimal threshold by setting the first false alarm probability less than or equal to an upper bound, and then determine a first detection probability. A receiver device also receives the second signal, extracts an object signal from it, determines a residual signal from the object signal, and a second testing statistic. The receiver may then determine a second optimal threshold and a second detection probability. Based on the first and second detection probabilities, the system may determine a probability of security authentication and thus measure the safety of the physical layer authentication system.

Our machine architecture and machine procedures use robustness, unpredictability and variability to hinder malware infection. In some embodiments, our machine instruction opcodes are randomized. The computing behavior of our machine is structurally stable (invariant) to small changes made to its machine instructions. Our invention expands the engineering method of stability to a cryptographically stable machine that is resistant to malware sabotage by an adversary.

Our procedures use quantum randomness to build unpredictable stable instructions. Our machine procedures can execute just before running a program so that the computing task can be performed with a different representation of its instructions during each run. A process of hiding a key or data inside of random noise is described that protects the privacy of the machine instruction opcodes and operands. In some embodiments, quantum randomness generates random noise, using photonic emission with a light emitting diode.

The invention provides a target detection method based on a generative confrontation network, which includes designing a generator, generating various samples according to category labels, designing an agent, detecting the data of the generator, providing false true values, and applying the data generated by the agent to Training of target detectors, designing target detectors, judging whether the generated data is conducive to improving the accuracy of target detection, designing adversaries, in the training phase, judging whether the data comes from real data or generated data, the generator and the discriminator are trained alternately, In the testing phase, the data to be detected is directly input into the target detector to obtain the detection result. The combination of the samples generated by the generation network of the invention and the real samples can enrich the training data, improve the detection accuracy, the target detection network provides feedback to the generation network, so that the generated samples are more realistic, and the data generated by the agent is directly applied to the training of the target detector. There is no need to spend a lot of manpower and material resources for labeling, and the present invention has a simple structure and is easy to deploy.