Drexel University

Sas Institute Inc.

Politecnico Di Milano

Prathaban Mookiah

Kapil R Dandekar

Weichen Wang

Hamoon Azizsoltani

Jorge Manuel Gomes Da Silva

Ruiwen Zhang

Ye Liu

Thomas J O'Connell

Rachel Greenstadt

John MacLaren Walsh

Robert Winston Blanchard

Daniele Piazza

Varunraj Valsaraj

Mohammadreza Nazari

Amirhassan Fallah Dizche

Jonathan Lee Walker

Hardi Desai

Davood Hajinezhad

Michele D'Amico

Afshin Oroojlooyjadid

John Kountouriotis

Ian Holmes

John Robert Watkins

A computing device trains a machine state predictive model. A generative adversarial network with an autoencoder is trained using a first plurality of observation vectors. Each observation vector of the first plurality of observation vectors includes state variable values for state variables and an action variable value for an action variable. The state variables define a machine state, wherein the action variable defines a next action taken in response to the machine state. The first plurality of observation vectors successively defines sequential machine states to manufacture a product. A second plurality of observation vectors is generated using the trained generative adversarial network with the autoencoder. A machine state machine learning model is trained to predict a subsequent machine state using the first plurality of observation vectors and the generated second plurality of observation vectors. A description of the machine state machine learning model is output.

Automated control of a manufacturing process

(A) Conditional vectors are defined. (B) Latent observation vectors are generated using a predefined noise distribution function. (C) A forward propagation of a generator model is executed with the conditional vectors and the latent observation vectors as input to generate an output vector. (D) A forward propagation of a decoder model of a trained autoencoder model is executed with the generated output vector as input to generate a plurality of decoded vectors. (E) Transformed observation vectors are selected from transformed data based on the defined plurality of conditional vectors. (F) A forward propagation of a discriminator model is executed with the transformed observation vectors, the conditional vectors, and the decoded vectors as input to predict whether each transformed observation vector and each decoded vector is real or fake. (G) The discriminator and generator models are updated and (A) through (G) are repeated until training is complete.

Tabular data generation for machine learning model training system

Logical rules can be automatically generated for use with event detection systems according to some aspects of the present disclosure. For example, a computing device can extract a group of logical rules from trained decision trees and apply a test data set to the group of logical rules to determine count values corresponding to the logical rules. The computing device can then determine performance metric values based on the count values, select a subset of logical rules from among the group of logical rules based on the performance metric values, and provide at least one logical rule in the subset for use with an event detection system. The event detection system can be configured to detect an event in relation to a target data set that was not used to train the decision trees.

Automatically generating rules for event detection systems

Malicious activity can be detected and prevented in real-time or otherwise. For example, a system of the present disclosure can receive a request from a user to obtain access to an entity, determine data objects based on the request, and access data-object network definitions corresponding to the determined data objects. The system can also receive a profile for the user indicating behavioral information relating to the user. The system can then determine a likelihood that the request is associated with malicious activity based on (i) the data objects, (ii) the profile, and (iii) the data-object network definitions. The system can allow or deny the user access to the entity based on the likelihood that the request is associated with malicious activity.

Real-time detection and prevention of malicious activity

Channel based authentication schemes for intrusion detection that operates at the physical layer are described that apply the capabilities of a pattern reconfigurable antenna for improved performance. Performance gains are achieved by the schemes as a function of the number of antenna modes. The first scheme relies on a channel based fingerprint for differentiating between transmitters whereas another scheme poses the intruder detection problem as a generalized likelihood ratio (GLR) test problem that operates on the channel realizations corresponding to different modes present in a reconfigurable antenna. The benefits of these two schemes over single element antennas are demonstrated. General guidelines are provided on how to choose the different elements of the decision metric in order to realize better performance for physical layer based authentication schemes based on any diversity scheme.

Reconfigurable antenna based solutions for device authentication and intrusion detection in wireless networks

Reconfigurable antennas in an ad-hoc network are provided where all nodes employ MIMO/SIMO/MISO communication techniques. Three types of reconfigurable antennas: Reconfigurable Printed Dipole Array (RPDA), Reconfigurable Circular Patch Antenna (RCPA) and Two-Port Reconfigurable CRLH Leaky Wave Antennas are used. The RPDA, RCPA and the CRLH Leaky Wave antennas have a different number of configurations as well as different degrees of pattern diversity between possible configurations. To effectively use these antennas in a network, the performance of centralized and decentralized antenna configuration selection schemes are quantified for reconfiguration at one or both link ends. The sum capacity of the network is used as a metric to quantify the performance of these antennas in measured and simulated network channels.

Reconfigurable antennas and configuration selection methods for AD-HOC networks

Channel based authentication schemes for intrusion detection that operates at the physical layer are described that apply the capabilities of a pattern re-configurable antenna for improved performance. Performance gains are achieved by the schemes as a function of the number of antenna modes. The first scheme relies on a channel based fingerprint for differentiating between transmitters whereas another scheme poses the intruder detection problem as a generalized likelihood ratio (GLR) test problem that operates on the channel realizations corresponding to different modes present in a reconfigurable antenna. The benefits of these two schemes over single element antennas are demonstrated. General guidelines are provided on how to choose the different elements of the decision metric in order to realize better performance for physical layer based authentication schemes based on any diversity scheme.

A bi-directional antenna includes a plurality of unit cells stacked in two perpendicular planes (Y-X and Z-X planes) to form cube shaped unit cells whereby inductive loops are placed on four faces corresponding to the Y-X and Z-X planes. Each unit cell includes a magnetic permeability enhanced metamaterial. The resulting antenna has the ability to couple magnetic fields oriented in both the X and Y directions with increased permeabilities and can be used to realize a variety of different antenna architectures that do not have their magnetic field confined in a single direction.

Bi-directional magnetic permeability enhanced metamaterial (MPEM) substrate for antenna miniaturization

A magnetic permeability enhanced metamaterial is used to enhance the antenna array of a Multiple Input Multiple Output (MIMO) communication system. A rectangular patch antenna array is formed including a stack of a plurality of unit cells, where each unit cell includes an inductive loop of magnetic permeability enhanced metamaterials embedded in a host dielectric substrate. The use of such metamaterials permits the antenna arrays to be made smaller, and have less mutual coupling, when using a metamaterial substrate. The measured channel capacities of the antenna arrays are similar for the metamaterial and conventional substrates; however, the capacity improvement when using MIMO relative to single antenna communication systems is greater for antennas on metamaterial substrates.

Growing community of inventors

San Diego, CA, United States of America

Prathaban Mookiah