Patents by Inventor Jonathan Yedidia

Jonathan Yedidia 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).

  • Patent number: 10572807
    Abstract: A method and device determines an optimization solution for an optimization problem. The method includes receiving the optimization problem having cost functions and variables in which each of the cost functions has a predetermined relationship with select ones of the variables. The variables comprise a sub-solution of a spline indicative of a curved path along an estimated trajectory. The method includes generating a first message for each of the cost functions for each corresponding variable based upon the relationship and a second message for each of the variables for each corresponding cost function based upon the relationship. The method includes generating a disagreement variable for each corresponding pair of variables and cost functions measuring a disagreement value between the first and second beliefs. The method includes forming a consensus between the first and second messages until the optimization solution is determined.
    Type: Grant
    Filed: March 31, 2016
    Date of Patent: February 25, 2020
    Assignee: Disney Enterprises, Inc.
    Inventors: Jonathan Yedidia, Nate Derbinsky, Jose Bento Ayres Pereira, Dan Schmidt, Charles Mathy
  • Patent number: 9639813
    Abstract: A method and device determines an optimization solution for an optimization problem. The method includes receiving the optimization problem having cost functions and variables in which each of the cost functions has a predetermined relationship with select ones of the variables. The method includes generating a first message for each of the cost functions for each corresponding variable based upon the respective predetermined relationship and a second message for each of the variables for each corresponding cost function based upon the respective predetermined relationship. The method includes generating a disagreement variable for each corresponding pair of variables and cost functions measuring a disagreement value between the first and second beliefs. The method includes repeating steps (b), (c), and (d) until a consensus is formed between the first and second messages until the optimization solution is determined based upon the consensus.
    Type: Grant
    Filed: December 16, 2013
    Date of Patent: May 2, 2017
    Assignee: DISNEY ENTERPRISES, INC.
    Inventors: Jonathan Yedidia, Nate Derbinsky, Jose Bento Ayres Pereira, Veit Elser
  • Publication number: 20160217380
    Abstract: A method and device determines an optimization solution for an optimization problem. The method includes receiving the optimization problem having cost functions and variables in which each of the cost functions has a predetermined relationship with select ones of the variables. The variables comprise a sub-solution of a spline indicative of a curved path along an estimated trajectory. The method includes generating a first message for each of the cost functions for each corresponding variable based upon the relationship and a second message for each of the variables for each corresponding cost function based upon the relationship. The method includes generating a disagreement variable for each corresponding pair of variables and cost functions measuring a disagreement value between the first and second beliefs. The method includes forming a consensus between the first and second messages until the optimization solution is determined.
    Type: Application
    Filed: March 31, 2016
    Publication date: July 28, 2016
    Inventors: Jonathan YEDIDIA, Nate DERBINSKY, Jose Bento Ayres PEREIRA, Dan SCHMIDT, Charles MATHY
  • Patent number: 9349106
    Abstract: There is provided a system and method for training and utilizing a boundary graph machine learning algorithm. The system including a processor configured to receive a plurality of entry nodes, each of the plurality of entry nodes including an entry node input and an entry node output, add each of the plurality of entry nodes to a graph using the entry node input and the entry node output, receiving a plurality of training nodes, each of the plurality of training nodes including a training node input and a training node output, add each of the plurality of training nodes to the graph when the training node input for each of the plurality of training nodes is similar to the training node output of a closest node and the training node output of each of the plurality of training nodes is different than the training node output of the closest node.
    Type: Grant
    Filed: January 27, 2014
    Date of Patent: May 24, 2016
    Assignee: Disney Enterprises, Inc.
    Inventor: Jonathan Yedidia
  • Publication number: 20150149394
    Abstract: There is provided a system and method for training and utilizing a boundary graph machine learning algorithm. The system including a processor configured to receive a plurality of entry nodes, each of the plurality of entry nodes including an entry node input and an entry node output, add each of the plurality of entry nodes to a graph using the entry node input and the entry node output, receiving a plurality of training nodes, each of the plurality of training nodes including a training node input and a training node output, add each of the plurality of training nodes to the graph when the training node input for each of the plurality of training nodes is similar to the training node output of a closest node and the training node output of each of the plurality of training nodes is different than the training node output of the closest node.
    Type: Application
    Filed: January 27, 2014
    Publication date: May 28, 2015
    Applicant: Disney Enterprises, Inc.
    Inventor: Jonathan Yedidia
  • Publication number: 20140324761
    Abstract: A method and device determines an optimization solution for an optimization problem. The method includes receiving the optimization problem having cost functions and variables in which each of the cost functions has a predetermined relationship with select ones of the variables. The method includes generating a first message for each of the cost functions for each corresponding variable based upon the respective predetermined relationship and a second message for each of the variables for each corresponding cost function based upon the respective predetermined relationship. The method includes generating a disagreement variable for each corresponding pair of variables and cost functions measuring a disagreement value between the first and second beliefs. The method includes repeating steps (b), (c), and (d) until a consensus is formed between the first and second messages until the optimization solution is determined based upon the consensus.
    Type: Application
    Filed: December 16, 2013
    Publication date: October 30, 2014
    Applicant: Disney Enterprises, Inc.
    Inventors: Jonathan Yedidia, Nate Derbinsky, Jose Bento Ayres Pereira, Veit Elser
  • Patent number: 8595589
    Abstract: A system and a method for determining a quasi-cyclic (QC) low-density parity-check (LDPC) code, such that the QC LDPC code has no trapping sets are disclosed. A set of matrices representing a family of QC LDPC codes are acquired, wherein each QC LDPC code is a tail-biting spatially-coupled code of girth not less than eight, and wherein each column of each matrix in the set has a weight not less than four. Based on a trapping set pattern, a matrix from the set of matrices is selected such that the matrix represents the QC LDPC code with no trapping sets. The matrix can be stored into a memory.
    Type: Grant
    Filed: September 30, 2011
    Date of Patent: November 26, 2013
    Assignee: Mitsubishi Electric Research Laboratories, Inc.
    Inventors: Jonathan Yedidia, Yige Wang
  • Patent number: 8499218
    Abstract: A method estimates parameters of 3-phase voltage signals to synchronize a power grid in a presence of a voltage unbalance by transforming the 3-phase voltage signals to ??-reference signals using a Clarke transformation matrix, and estimating sinusoidal signals and corresponding quadrature signals of the ??-reference signals using an extended Kalman filter, and determining a phase angle of a positive sequence based on a relationship of the phase angle to the estimated the sinusoidal signals.
    Type: Grant
    Filed: September 30, 2011
    Date of Patent: July 30, 2013
    Assignee: Mitsubishi Electric Research Laboratories, Inc.
    Inventors: Jonathan Yedidia, Stark Draper, Yige Wang
  • Publication number: 20130086445
    Abstract: A system and a method for determining a quasi-cyclic (QC) low-density parity-check (LDPC) code, such that the QC LDPC code has no trapping sets are disclosed. A set of matrices representing a family of QC LDPC codes are acquired, wherein each QC LDPC code is a tail-biting spatially-coupled code of girth not less than eight, and wherein each column of each matrix in the set has a weight not less than four. Based on a trapping set pattern, a matrix from the set of matrices is selected such that the matrix represents the QC LDPC code with no trapping sets. The matrix can be stored into a memory.
    Type: Application
    Filed: September 30, 2011
    Publication date: April 4, 2013
    Inventors: Jonathan Yedidia, Yige Wang
  • Publication number: 20130086456
    Abstract: A system and a method determine a quasi-cyclic (QC) low-density parity-check (LDPC) code corresponding to a protograph and having a predetermined girth. At least several elements of the connectivity matrix representing the protograph are duplicated along at least one line of duplication to produce an inflated connectivity matrix, wherein values of at least several elements in the connectivity matrix form a pattern indicated by the predetermined girth. A hierarchical quasi-cyclic (HQC) LDPC code corresponding to the inflated connectivity matrix is determined and at least several elements of a matrix representing the HQC LDPC code are removed, such that the matrix is squashed along the line of duplication to produce the QC LDPC code.
    Type: Application
    Filed: September 30, 2011
    Publication date: April 4, 2013
    Inventors: Jonathan Yedidia, Stark Draper, Yige Wang
  • Publication number: 20110041033
    Abstract: A code to be decoded by message-passing is represented by a factor graph. The factor graph includes variable nodes indexed by i and constraint nodes indexed by a connected by edges for transferring messages mi?a outgoing from the variable nodes to the constraint nodes and messages ma?i incoming from the constraint nodes to the variable nodes. The messages mi?a are initialized based on beliefs bi of a received codeword. The messages ma?i are generated by overshooting the messages mi?a at the constraint nodes. The beliefs bi are updated at the variable nodes using the messages ma?i. The codeword is outputted if found, otherwise, the messages mi?a are updated using a correction for the overshooting.
    Type: Application
    Filed: June 21, 2010
    Publication date: February 17, 2011
    Inventors: Jonathan Yedidia, Yige Wang, Stark Draper
  • Patent number: 7372393
    Abstract: A method converts an input image of noisy wrapped phases to an output image of absolute unwrapped phases. The noisy wrapped phases in the input image are represented as a set of re-wrapped phases and a set of phase shifts. The set of re-wrapped phases are partitioned into a first group and a second group. Integer differences between the phase shifts are optimized while holding the re-wrapped phases fixed. Then, the first group of re-wrapped phases are optimized, while holding the integer differences between the phase shifts, and the second group of re-wrapped phases fixed. The integer differences between the phase shifts are optimized again, while holding the re-wrapped phases fixed. Then, the second group of re-wrapped phases are optimized, while holding the integer differences between the phase shifts, and the first group of re-wrapped phases fixed. The optimizing steps are repeated until the re-wrapped phase converge.
    Type: Grant
    Filed: July 7, 2006
    Date of Patent: May 13, 2008
    Assignee: Mitsubishi Electric Research Laboratories, Inc.
    Inventors: Jonathan Yedidia, Ali Azarbayejani, Emin Martinian, Ying-Zong Huang
  • Publication number: 20080007450
    Abstract: A method converts an input image of noisy wrapped phases to an output image of absolute unwrapped phases. The noisy wrapped phases in the input image are represented as a set of re-wrapped phases and a set of phase shifts. The set of re-wrapped phases are partitioned into a first group and a second group. Integer differences between the phase shifts are optimized while holding the re-wrapped phases fixed. Then, the first group of re-wrapped phases are optimized, while holding the integer differences between the phase shifts, and the second group of re-wrapped phases fixed. The integer differences between the phase shifts are optimized again, while holding the re-wrapped phases fixed. Then, the second group of re-wrapped phases are optimized, while holding the integer differences between the phase shifts, and the first group of re-wrapped phases fixed. The optimizing steps are repeated until the re-wrapped phase converge.
    Type: Application
    Filed: July 7, 2006
    Publication date: January 10, 2008
    Inventors: Jonathan Yedidia, Ali Azarbayejani, Emin Martinian, Ying-Zong Huang
  • Publication number: 20070217432
    Abstract: A system and method for communicating information in a wireless cooperative relay network of nodes, the nodes including a source, a set of relays, and a destination. The source broadcasts a code word encoded as a data stream using a rateless code. The relays receive the data stream, decode the data stream to recover the code word, and reencode and transmit the recovered code word as the data stream with the rateless code. The destination receives and decodes the reencoded data streams to recover the code word.
    Type: Application
    Filed: March 16, 2006
    Publication date: September 20, 2007
    Inventors: Andreas Molisch, Neelesh Mehta, Jonathan Yedidia, Jinyun Zhang
  • Publication number: 20070174633
    Abstract: Biometric parameters acquired from human forces, voices, fingerprints, and irises are used for user authentication and access control. Because the biometric parameters are continuous and vary from one reading to the next, syndrome codes are applied to determine biometric syndrome vectors. The biometric syndrome vectors can be stored securely while tolerating an inherent variability of biometric data. The stored biometric syndrome vector is decoded during user authentication using biometric parameters acquired at that time. The syndrome codes can also be used to encrypt and decrypt data.
    Type: Application
    Filed: November 29, 2006
    Publication date: July 26, 2007
    Inventors: Stark Draper, Ashish Khisti, Emin Martinian, Anthony Vetro, Jonathan Yedidia
  • Publication number: 20060161830
    Abstract: A method generates a combined-replica group-shuffled iterative decoder, comprising. First. an error-correcting code and an iterative decoder for an error-correcting code is received by the method. Multiple group-shuffled sub-decoders for the error-correcting code are constructed, based on the iterative decoder. Then, the multiple group-shuffled sub-decoders are combined into a combined-replica group-shuffled iterative decoder.
    Type: Application
    Filed: January 14, 2005
    Publication date: July 20, 2006
    Inventors: Jonathan Yedidia, Marc Fossorier, Juntan Zhang, Yige Wang
  • Publication number: 20060123241
    Abstract: First biometric parameters are acquired from a user. Input data are encrypted according to the biometric parameters to produce ciphertext. The biometric parameters are encoded using a syndrome encoder to produce a syndrome code. The ciphertext and the syndrome code are associated with each other and stored in a computer readable media so that only the same user can subsequently decrypt the cipher text.
    Type: Application
    Filed: September 1, 2005
    Publication date: June 8, 2006
    Inventors: Emin Martinian, Anthony Vetro, Sergey Yekhanin, Jonathan Yedidia
  • Publication number: 20060067382
    Abstract: A method encodes a sequence of blocks of input bits to be transmitted over a wireless channel. Each block of input bits is converted to a codeword, and each codeword is mapped to multiple sub-blocks of complex numbers. Each sub-block is multiplied by a disambiguating spreading transform to obtain a sub-block of transformed symbols, which can be modulated and transmitted to a receiver, where each received block is demodulated to a block of complex numbers, which are partitioned into sub-blocks. The sub-block of complex numbers are converted to a set of likelihood ratios that correspond to decoded codewords. The input bits can be decoded unambiguously even if only one of the blocks of symbols corresponding to the block of input bits is received correctly.
    Type: Application
    Filed: September 27, 2004
    Publication date: March 30, 2006
    Inventors: Jonathan Yedidia, Andreas Molisch, Karunakar Pedagani
  • Publication number: 20060045184
    Abstract: A method represents a correlated set of images. The correlation can be spatial or temporal. A lossy operation is applied to each image in the correlated set to generate a coarse image. The coarse image is encoded losslessly to yield an encoded coarse image. Each image is also represented by syndrome bits. The combination of the encoded coarse images and the syndrome bits represent the correlated set of images.
    Type: Application
    Filed: August 27, 2004
    Publication date: March 2, 2006
    Inventors: Anthony Vetro, Jonathan Yedidia
  • Publication number: 20060048038
    Abstract: A method compresses a set of correlated signals by first converting each signal to a sequence of integers, which are further organized as a set of bit-planes. This can be done by signal transformation and quantization. An inverse accumulator is applied to each bit-plane to produce a bit-plane of shifted bits, which are permuted according to a predetermined permutation to produce bit-planes of permuted bits. Each bit-plane of permuted bits is partitioned into a set of blocks of bits. Syndrome bits are generated for each block of bits according to a rate-adaptive base code. Subsequently, the syndrome bits can be decompressed in a decoder to recover the original correlated signals. For each bit-plane of the corresponding signal, a bit probability estimate is generated. Then, the bit-plane is reconstructed using the syndrome bits and the bit probability estimate.
    Type: Application
    Filed: August 27, 2004
    Publication date: March 2, 2006
    Inventors: Jonathan Yedidia, Anthony Vetro, Ashish Khisti, Dmitry Malioutov