Patents by Inventor Bob Hu

Bob Hu has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Publication number: 20190156200
    Abstract: In some embodiments, anomaly detection may be facilitated via a multi-neural-network architecture. In some embodiments, a first neural network may be configured to generate hidden representations of data items corresponding to a concept. A second neural network may be configured to generate reconstructions of the data items from the hidden representations. The first neural network may be configured to assess the reconstructions against the data items and update configurations of the first neural network based on the assessment of the reconstructions. Subsequent to the update of the first neural network, the first neural network may generate a hidden representation of a first data item from the first data item. The second neural network may generate a reconstruction of the first data item from the hidden representation. An anomaly in the first data item may be detected based on differences between the first data item and the reconstruction.
    Type: Application
    Filed: June 18, 2018
    Publication date: May 23, 2019
    Inventor: Bob HU
  • Patent number: 10121104
    Abstract: In some embodiments, anomaly detection may be facilitated via a multi-neural-network architecture. In some embodiments, a first neural network may be configured to generate hidden representations of data items corresponding to a concept. A second neural network may be configured to generate reconstructions of the data items from the hidden representations. The first neural network may be configured to assess the reconstructions against the data items and update configurations of the first neural network based on the assessment of the reconstructions. Subsequent to the update of the first neural network, the first neural network may generate a hidden representation of a first data item from the first data item. The second neural network may generate a reconstruction of the first data item from the hidden representation. An anomaly in the first data item may be detected based on differences between the first data item and the reconstruction.
    Type: Grant
    Filed: July 9, 2018
    Date of Patent: November 6, 2018
    Assignee: AIVITAE LLC
    Inventor: Bob Hu
  • Publication number: 20180120401
    Abstract: This disclosure provides a computer-implemented method for sequencing magnetic resonance imaging waveforms using a multistage sequencing hardware. The method comprises creating, with the aid of a computer processor, an active memory region that includes waveforms and schedules being played, and creating one or more buffer memory regions that contain waveforms and schedules not currently being played. Next, the waveforms and schedules in the one or more buffer memory regions may be updated while waveforms may be played in the active memory region. Upon completion of the waveform playback in the active memory region, the active and buffer memory regions may be swapped so that the former buffer memory region becomes the active memory region, and the former active memory region becomes the buffer memory region. The method may be repeated as needed until the imaging process is completed or otherwise halted.
    Type: Application
    Filed: December 6, 2017
    Publication date: May 3, 2018
    Inventors: Taehoon SHIN, Bob HU, William OVERALL, Juan Santos
  • Publication number: 20130274592
    Abstract: A method for acquiring a volumetric scan from at least a portion of a body of a subject suspected of exhibiting an observable manifestation of a disease or adverse health condition comprises, with the aid of a radio frequency (RF) source of a magnetic resonance imaging (MRI) system, applying a first RF pulse to the at least the portion of a body of the subject. A detector coil of the MRI system can then detect magnetic resonance (MR) signals from the at least the portion of the body of the subject. The MR signals can be detected upon a time delay subsequent to the application of the first RF pulse. The MR signals can be stored in a memory location as non-Cartesian data in k-space.
    Type: Application
    Filed: February 28, 2013
    Publication date: October 17, 2013
    Inventors: Taehoon SHIN, Bob Hu
  • Publication number: 20050177054
    Abstract: A device and software system with input and output capability for manipulating real and virtual objects in 3-dimensional space. The device consists of a six degree-of-freedom mechanical armature that has sensors to determine the location and orientation of a stylus and planar surface. In the input mode, manipulation of the physical armature will result in a corresponding two-dimensional, virtual image of the stylus and surface on a computer screen. The armature also has motors to automatically change the armature location and orientation in order to generate a physical representation in the real world of the location and orientation of a virtual object. The armature is built so that it maintains balance at any location and orientation to statically maintain the armature location and orientation without drifting to a null rest position.
    Type: Application
    Filed: February 10, 2004
    Publication date: August 11, 2005
    Inventors: Dingrong Yi, Graham Wright, Bob Hu