Explanipedia

Robust Training of Neural Networks at Arbitrary Precision and Sparsity Open

Chengxi Ye, Grace Chu, Yanfeng Liu, Yichi Zhang, Łukasz Lew , et al. · 2024

The discontinuous operations inherent in quantization and sparsification introduce a long-standing obstacle to backpropagation, particularly in ultra-low precision and sparse regimes. The standard Straight-Through Estimator (STE) is widely…

MobileNetV4 -- Universal Models for the Mobile Ecosystem Open

Danfeng Qin, Chas Leichner, Manolis Delakis, Marco Fornoni, Shixin Luo , et al. · 2024

We present the latest generation of MobileNets, known as MobileNetV4 (MNv4), featuring universally efficient architecture designs for mobile devices. At its core, we introduce the Universal Inverted Bottleneck (UIB) search block, a unified…

Exploiting Invariance in Training Deep Neural Networks Open

Chengxi Ye, Xiong Zhou, Tristan McKinney, Yanfeng Liu, Qinggang Zhou , et al. · 2022

Inspired by two basic mechanisms in animal visual systems, we introduce a feature transform technique that imposes invariance properties in the training of deep neural networks. The resulting algorithm requires less parameter tuning, train…

Strategies for Controlling or Releasing the Influence Due to the Volume Expansion of Silicon inside Si−C Composite Anode for High-Performance Lithium-Ion Batteries Open

Xian Zhang, Jingzheng Weng, Chengxi Ye, Mengru Liu, Chenyu Wang , et al. · 2022

Currently, silicon is considered among the foremost promising anode materials, due to its high capacity, abundant reserves, environmental friendliness, and low working potential. However, the huge volume changes in silicon anode materials …

Exploiting Invariance in Training Deep Neural Networks Open

Chengxi Ye, Xiong Zhou, Tristan McKinney, Yanfeng Liu, Qinggang Zhou , et al. · 2021

Inspired by two basic mechanisms in animal visual systems, we introduce a feature transform technique that imposes invariance properties in the training of deep neural networks. The resulting algorithm requires less parameter tuning, train…

Research progress in electrochemical properties of lithium batteries with PAA binders Open

Mengru Liu, Chengxi Ye, Libo Peng, Jingzheng Weng · 2021

Electrode binder is an essential part to maintain the integrity of electrode, and it is very important to improve the specific capacity and cycle stability of the battery. Polyacrylic acid (PAA) is widely used as the binder of anode or cat…

From asymmetrical to balanced genomic diversification during rediploidization: Subgenomic evolution in allotetraploid fish Open

Jing Luo, Jing Chai, Yanling Wen, Min Tao, Guoliang Lin , et al. · 2020

Allotetraploid cyprinids have a unique strategy for balancing subgenomic stabilization and diversification in rediploidization.

Network Deconvolution Open

Chengxi Ye, Matthew Evanusa, Hua He, Anton Mitrokhin, Thomas Goldstein , et al. · 2019

Convolution is a central operation in Convolutional Neural Networks (CNNs), which applies a kernel to overlapping regions shifted across the image. However, because of the strong correlations in real-world image data, convolutional kernels…

EV-IMO: Motion Segmentation Dataset and Learning Pipeline for Event Cameras Open

Anton Mitrokhin, Chengxi Ye, Cornelia Fermüller, Yiannis Aloimonos, Tobi Delbrück · 2019

We present the first event-based learning approach for motion segmentation in indoor scenes and the first event-based dataset - EV-IMO - which includes accurate pixel-wise motion masks, egomotion and ground truth depth. Our approach is bas…

LEARNING OF DENSE OPTICAL FLOW, MOTION AND DEPTH, FROM SPARSE EVENT CAMERAS Open

Chengxi Ye · 2019

With recent advances in the field of autonomous driving, autonomous agents need to safely navigate around humans or other moving objects in unconstrained, highly dynamic environments. In this thesis, we demonstrate the feasibility of recon…

EV-IMO: Motion Segmentation Dataset and Learning Pipeline for Event Cameras Open

Anton Mitrokhin, Chengxi Ye, Cornelia Fermüller, Yiannis Aloimonos, Tobi Delbrück · 2019

We present the first event-based learning approach for motion segmentation in indoor scenes and the first event-based dataset - EV-IMO - which includes accurate pixel-wise motion masks, egomotion and ground truth depth. Our approach is bas…

Unsupervised Learning of Dense Optical Flow and Depth from Sparse Event Data. Open

Chengxi Ye, Anton Mitrokhin, Chethan M. Parameshwara, Cornelia Fermüller, James A. Yorke , et al. · 2018

In this work we present a lightweight, unsupervised learning pipeline for \textit{dense} depth, optical flow and egomotion estimation from sparse event output of the Dynamic Vision Sensor (DVS). To tackle this low level vision task, we use…

Unsupervised Learning of Dense Optical Flow, Depth and Egomotion from Sparse Event Data Open

Chengxi Ye, Anton Mitrokhin, Cornelia Fermüller, James A. Yorke, Yiannis Aloimonos · 2018

In this work we present a lightweight, unsupervised learning pipeline for \textit{dense} depth, optical flow and egomotion estimation from sparse event output of the Dynamic Vision Sensor (DVS). To tackle this low level vision task, we use…

Publisher Correction: The sea lamprey germline genome provides insights into programmed genome rearrangement and vertebrate evolution Open

Jeramiah J. Smith, Nataliya Timoshevskaya, Chengxi Ye, Carson Holt, Melissa C. Keinath , et al. · 2018

When published, this article did not initially appear open access. This error has been corrected, and the open access status of the paper is noted in all versions of the paper. Additionally, affiliation 16 denoting equal contribution was m…

Evenly Cascaded Convolutional Networks Open

Chengxi Ye, Chinmaya Devaraj, Michael Maynord, Cornelia Fermüller, Yiannis Aloimonos · 2018

We introduce Evenly Cascaded convolutional Network (ECN), a neural network taking inspiration from the cascade algorithm of wavelet analysis. ECN employs two feature streams - a low-level and high-level steam. At each layer these streams i…

Publisher Correction: The sea lamprey germline genome provides insights into programmed genome rearrangement and vertebrate evolution Open

Jeramiah J. Smith, Nataliya Timoshevskaya, Chengxi Ye, Carson Holt, Melissa C. Keinath , et al. · 2018

The sea lamprey germline genome provides insights into programmed genome rearrangement and vertebrate evolution Open

Jeramiah J. Smith, Nataliya Timoshevskaya, Chengxi Ye, Carson Holt, Melissa C. Keinath , et al. · 2018

The sea lamprey ( Petromyzon marinus ) serves as a comparative model for reconstructing vertebrate evolution. To enable more informed analyses, we developed a new assembly of the lamprey germline genome that integrates several complementar…

On the Importance of Consistency in Training Deep Neural Networks Open

Chengxi Ye, Yezhou Yang, Cornelia Fermüller, Yiannis Aloimonos · 2017

We explain that the difficulties of training deep neural networks come from a syndrome of three consistency issues. This paper describes our efforts in their analysis and treatment. The first issue is the training speed inconsistency in di…

LightNet Open

Chengxi Ye, Chen Zhao, Yezhou Yang, Cornelia Fermüller, Yiannis Aloimonos · 2016

LightNet is a lightweight, versatile, purely Matlab-based deep learning framework. The idea underlying its design is to provide an easy-to-understand, easy-to-use and efficient computational platform for deep learning research. The impleme…

DBG2OLC: Efficient Assembly of Large Genomes Using Long Erroneous Reads of the Third Generation Sequencing Technologies Open

Chengxi Ye, C. Hill, Shigang Wu, Jue Ruan, Zhanshan Ma · 2016

Peer Review #2 of "Sparc: a sparsity-based consensus algorithm for long erroneous sequencing reads (v0.3)" Open

Chengxi Ye · 2016