Samsung Electronics Co., Ltd.

Marvell International Limited

Marvellworld Trade Ltd.

Leland Stanford Junior University

Marvell Semiconductor, Inc.

Jungwon Lee

Hui-Ling Lou

Mostafa El-Khamy

Inyup Kang

Dongwoon Bai

Yakun Sun

Jihwan P Choi

Rohit U Nabar

Jiwoong Choi

Dimitrios-Alexandros Toumpakaris

Hyukjoon Kwon

Yoo Jin Choi

Jung Hyun Bae

Leilei Song

Qing Zhao

Haoyu Ren

Woong Jun Jang

Pranav Dayal

Gennady Feygin

Sungsoo Kim

Adina Matache

Mojtaba Rahmati

Jaeyoung Kim

Hongbing Cheng

Binnan Zhuang

Linbo Li

Chris Cheng-Chieh Lee

Niranjan Nayak Ratnakar

Seonkeun Park

Duyeong Choi

Seungjoon Lee

Jaewoong Chung

Ethan Lee

Shuangquan Wang

Jinhong Wu

Sayandev Mukherjee

Hessam Mahdavifar

Wonkyu Sung

Yingqun Yu

Shahab Sanayei

Hongyuan Zhang

Sandeep H Krishnamurthy

Songping Wu

Dahyun Lee

Seyed Mohsen Hosseinian

Arvind Yedla

Sahnghee Bahn

Jongwoo Jung

Kee-Bong Song

Donghyun Yeom

Kangmoon Kim

Junho Jin

Dohyung Ha

Konstantinos Sarrigeorgidis

Arul Durai Murugan Palanivelu

Jinhoo Lee

Vladan Petrovic

Raj Mani Misra

Hsien-Ping Lin

Insoo Hwang

Vignesh Sethuraman

Jongchul Choi

Seongjun Park

Changryong Heo

Jong-Chul Choi

Minwoo Yoo

Min-Sung Lee

Kwang-Tai Kim

Hyun-Ju Hong

Daehyeong Park

Hyungsup Byeon

Bhaskar Nallapureddy

Sungyoon Cho

Min-Su Jung

Chun Kin Au Yeung

Federico Penna

Min-Woo Yoo

Yonghwa Han

Hairuo Zhuang

Hyunseok Ryu

John Matthew Cioffi

Namsu Kim

Kedar Shirali

Junyoung Choi

Youngjin Park

Nan Zhang

Jieun Myung

Wooyong Lee

Bosoon Kang

Yuhan Zhou

Inkyu Lee

Junghyuck Im

Sangkyu Park

Kisung Kim

Seong Taek Chung

Sanguk Park

Wonseok Oh

Ruchen Duan

Raul Herman Etkin

Jiwon Steve Han

Methods and apparatuses for deep learning training are provided which include receiving a candidate unit for classification. The candidate unit including an intersection area between a ground-truth bounding box and a detection box. The candidate unit is classified by assigning a label that is a probability value that a given feature is observed in the intersection area. Deep learning training is performed using the assigned label of the classified candidate unit.

System and method for training deep learning classification networks

A method includes: computing noise data by subtracting, by a processing circuit, a noisy image from a corresponding ground truth image; clustering, by the processing circuit, a plurality of noise values of the noise data based on intensity values of the corresponding ground truth image; permuting, by the processing circuit, a plurality of locations of the noise values of the noise data within each cluster; generating, by the processing circuit, a synthetic noise image based on the permuted locations of the noise values; adding, by the processing circuit, the synthetic noise image to the corresponding ground truth image to generate a synthetic noisy image; and augmenting an image dataset for training a neural network to perform image denoising with the synthetic noisy image.

Systems and methods for image denoising using deep convolutional networks

A method of enabling communication between a user equipment (UE) and a non-terrestrial network (NTN), is described. The UE may be able to calculate a timing advance compensation. The timing advance compensation may be a differential or full timing advance compensation. An offset value, K, may be indicated to the UE. The Kvalue may be used for timing relationships.

Methods for timing advance indication and timing relationships indication for non-terrestrial networks

A method and user equipment (UE) are provided. The method includes transmitting, from a first UE, an assistance request to at least one neighboring UE, receiving, by the first UE and from the at least one neighboring UE, assistance information including an indication of at least one resource for transmission, and transmitting, from the first UE, over the indicated at least one resource.

User equipment assistance for resource selection in new radio vehicle to everything

A convolutional neural network (CNN) system for generating a classification for an input image is presented. The CNN system comprises circuitry running on clock cycles and configured to compute a product of two received values, and at least one non-transitory computer-readable medium that stores instructions for the circuitry to derive a feature map based on at least the input image; puncture at least one selection among the feature map and a kernel by setting the value of an element at an index of the at least one selection to zero and cyclic shifting a puncture pattern to achieve a 1/d reduction in number of clock cycles, where d is an integer and puncture interval value >1. The feature map is convolved with the kernel to generate an output, and a classification of the input image is generated based on the output.

Method and apparatus for reducing computational complexity of convolutional neural networks

A method and system for training a neural network are provided. The method includes receiving an input image, selecting at least one data augmentation method from a pool of data augmentation methods, generating an augmented image by applying the selected at least one data augmentation method to the input image, and generating a mixed image from the input image and the augmented image.

Method and apparatus for data efficient semantic segmentation

Methods and devices are provided for performing federated learning. A global model is distributed from a server to a plurality of client devices. At each of the plurality of client devices: model inversion is performed on the global model to generate synthetic data; the global model is on an augmented dataset of collected data and the synthetic data to generate a respective client model; and the respective client model is transmitted to the server. At the server: client models are received from the plurality of client devices, where each client model is received from a respective client device of the plurality of client devices; model inversion is performed on each client model to generate a synthetic dataset; the client models are averaged to generate an averaged model; and the averaged model is trained using the synthetic dataset to generate an updated model.

Method and apparatus for federated learning

Apparatuses and methods of manufacturing same, systems, and methods for performing network parameter quantization in deep neural networks are described. In one aspect, multi-dimensional vectors representing network parameters are constructed from a trained neural network model. The multi-dimensional vectors are quantized to obtain shared quantized vectors as cluster centers, which are fine-tuned. The fine-tuned and shared quantized vectors/cluster centers are then encoded. Decoding reverses the process.

Method and apparatus for neural network quantization

A method and system are provided. The method includes performing channel estimation on a reference signal (RS), compressing, with a neural network, the channel estimation of the RS, decompressing, with the neural network, the compressed channel estimation, and interpolating the decompressed channel estimation.

Deep learning-based channel buffer compression

A method for training a generator, by a generator training system including a processor and memory, includes: extracting training statistical characteristics from a batch normalization layer of a pre-trained model, the training statistical characteristics including a training mean μ and a training variance σ; initializing a generator configured with generator parameters; generating a batch of synthetic data using the generator; supplying the batch of synthetic data to the pre-trained model; measuring statistical characteristics of activations at the batch normalization layer and at the output of the pre-trained model in response to the batch of synthetic data, the statistical characteristics including a measured mean {circumflex over (μ)}and a measured variance {circumflex over (σ)}; computing a training loss in accordance with a loss function Lbased on μ, σ, {circumflex over (μ)}, and {circumflex over (σ)}; and iteratively updating the generator parameters in accordance with the training loss until a training completion condition is met to compute the generator.

Method and apparatus for data-free post-training network quantization and generating synthetic data based on a pre-trained machine learning model

A method and system for providing Gaussian weighted self-attention for speech enhancement are herein provided. According to one embodiment, the method includes receiving an input noise signal, generating a score matrix based on the received input noise signal, and applying a Gaussian weighted function to the generated score matrix.

Transformer with Gaussian weighted self-attention for speech enhancement

A method and system are provided. The method includes receiving an input frame, performing high frequency noise reduction (HFNR) on the received input frame, performing low frequency noise reduction (LFNR) on the received input frame, and obtaining a hybrid denoised clean frame by fusing an output of the performed HFNR and an output of the performed LFNR.

Method and apparatus for adaptive hybrid fusion

A method of performing scene-dependent lens shading correction (SD-LSC) is provided. The method includes collecting scene information from a Bayer thumbnail of an input image; generating a standard red green blue (sRGB) thumbnail by processing the Bayer thumbnail of the input image to simulate white balance (WB) and pre-gamma blocks; determining a representative color channel ratio of the input image based on the scene information and the sRGB thumbnail; determining an ideal grid gain of the input image based on the representative color channel ratio and a grid gain of the input image; merging the ideal grid gain and the grid gain of the input image to generate a new grid gain; and applying the new grid gain to the input image.

Method and apparatus based on scene dependent lens shading correction

A method and system are provided. The method includes obtaining a difference map between a reference frame and a non-reference frame, estimating a noise variance of the obtained difference map, obtaining merging weights based on the estimated noise variance and the obtained different map, and merging the non-reference frame with the reference frame using the obtained merging weights.

Method and apparatus for adaptive multi-frame image fusion

A system for performing echo cancellation includes: a processor configured to: receive a far-end signal; record a microphone signal including: a near-end signal; and an echo signal corresponding to the far-end signal; extract far-end features from the far-end signal; extract microphone features from the microphone signal; compute estimated near-end features by supplying the microphone features and the far-end features to an acoustic echo cancellation module including a recurrent neural network including: an encoder including a plurality of gated recurrent units; and a decoder including a plurality of gated recurrent units; compute an estimated near-end signal from the estimated near-end features; and transmit the estimated near-end signal to the far-end device. The recurrent neural network may include a contextual attention module; and the recurrent neural network may take, as input, a plurality of error features computed based on the far-end features, the microphone features, and acoustic path parameters.

System and method for acoustic echo cancelation using deep multitask recurrent neural networks

Growing community of inventors

San Diego, CA, United States of America

Jungwon Lee