arXiv subjects

Dongkyu Kim

Publications and source records attributed to Dongkyu Kim.

21 records · Page 2Linked to original sources

arXiv2018

Smallest Neural Network to Learn the Ising Criticality

Learning with an artificial neural network encodes the system behavior in a feed-forward function with a number of parameters optimized by data-driven training. An open question is whether one can minimize the network complexity without loss of performance to reveal how and why it works. Here we investigate the learning of the phase transition in the Ising model and find that having two hidden neurons can be enough for an accurate prediction of critical temperature. We show that the networks learn the scaling dimension of the order parameter while being trained as a phase classifier, demonstrating how the machine learning exploits the Ising universality to work for different lattices of the same criticality within a single set of trainings in one lattice geometry.

cond-mat.stat-mech↗

arXiv2017

A Robust Blind Watermarking Using Convolutional Neural Network

This paper introduces a blind watermarking based on a convolutional neural network (CNN). We propose an iterative learning framework to secure robustness of watermarking. One loop of learning process consists of the following three stages: Watermark embedding, attack simulation, and weight update. We have learned a network that can detect a 1-bit message from a image sub-block. Experimental results show that this learned network is an extension of the frequency domain that is widely used in existing watermarking scheme. The proposed scheme achieved robustness against geometric and signal processing attacks with a learning time of one day.

cs.MM↗

arXiv2017

Nonlinear Self-Interference Cancellation for Full-Duplex Radios: From Link- and System-Level Performance Perspectives

One of the promising technologies for LTE Evolution is full-duplex radio, an innovation is expected to double the spectral efficiency. To realize full-duplex in practice, the main challenge is overcoming self-interference, and to do so, researchers have developed self-interference cancellation techniques. Since most wireless transceivers use power amplifiers, especially in cellular systems, researchers have revealed the importance of nonlinear self-interference cancellation. In this article, we first explore several nonlinear digital self-interference cancellation techniques. We then propose a low complexity pre-calibration-based nonlinear digital self-interference cancellation technique. Next we discuss issues about reference signal allocation and the overhead of each technique. For performance evaluations, we carry out extensive measurements through a real-time prototype and link-/system-level simulations. For link-level analysis, we measure the amount of cancelled self-interference for each technique. We also evaluate system-level performances through 3D ray-tracing-based simulations. Numerical results confirm the significant performance improvement over a half-duplex system even in interference-limited indoor environments.

cs.IT↗