• Smart Media Journal
• /
• v.12 no.9
• /
• pp.89-94
• /
• 2023

Recommender systems provide users with the most favorable items by analyzing explicit or implicit feedback of users on items. Recently, as the size of deep-learning-based models employed in recommender systems has increased, many studies have focused on reducing inference time while maintaining high recommendation accuracy. As one of them, a study on recommender systems with a knowledge distillation (KD) technique is actively conducted. By KD, a small-sized model (i.e., student) is trained through knowledge extracted from a large-sized model (i.e., teacher), and then the trained student is used as a recommendation model. Existing studies on KD for recommender systems have been mainly performed only for implicit feedback settings. Thus, in this paper, we try to investigate the performance and accuracy when applied to explicit feedback settings. To this end, we leveraged a total of five state-of-the-art KD methods and three real-world datasets for recommender systems.

• Journal of Institute of Control, Robotics and Systems
• /
• v.10 no.12
• /
• pp.1295-1304
• /
• 2004

In this paper, the Bayesian recurrent neural network is proposed to predict time series data. A neural network predictor requests proper learning strategy to adjust the network weights, and one needs to prepare for non-linear and non-stationary evolution of network weights. The Bayesian neural network in this paper estimates not the single set of weights but the probability distributions of weights. In other words, the weights vector is set as a state vector of state space method, and its probability distributions are estimated in accordance with the particle filtering process. This approach makes it possible to obtain more exact estimation of the weights. In the aspect of network architecture, it is known that the recurrent feedback structure is superior to the feedforward structure for the problem of time series prediction. Therefore, the recurrent neural network with Bayesian inference, what we call Bayesian recurrent neural network (BRNN), is expected to show higher performance than the normal neural network. To verify the proposed method, the time series data are numerically generated and various kinds of neural network predictor are applied on it in order to be compared. As a result, feedback structure and Bayesian learning are better than feedforward structure and backpropagation learning, respectively. Consequently, it is verified that the Bayesian reccurent neural network shows better a prediction result than the common Bayesian neural network.

• Convergence Security Journal
• /
• v.16 no.5
• /
• pp.17-22
• /
• 2016

This study examined the current state of ICT utilization in physical education and presented the available physical education ICT contents configuration model. Using ICT in physical education is expected to be an alternative of overcome the bad physical education facilities and environment, also highly utilized as a valuable material that can provide specific feedback in learning motor skill. The types of ICT use in physical education classes are being utilized divided by web-use learning and application program-use learning. In composing physical education ICT contents, the server presents a problem to be solved by each student and encourage cooperative learning. If one team determines the solving idea about learning problem through team discussion, they solve the problem by repeating the process of giving teacher and other team's feedback on determining opinion. On the other hand, the class begins after learning the principle of specific movement utilizing the VTR or computer S/W in the practical training lesson of physical education. If the good hardware and software environment combine with the transformation of the recognition on physical education which has been away from the ICT, it will be able expected a broad using of ICT in physical education.

• Communications of Mathematical Education
• /
• v.31 no.4
• /
• pp.457-485
• /
• 2017

The Self-reported Evaluation tool developed in this study allows the learners to check and evaluate their own learning by determining the details that are self-assessed. Also this tool allows learners to receive feedback on their self - evaluation results. In this study pre - post test was performed to investigate the effect of self - assessment on the learners' tendency of studying math. The result showed that Self-reported evaluation improved self - confidence, self - strategy on learning mathematics, and meta-cognitive ability. Also by conducting a qualitative analysis of the Self-reported evaluation, students practiced the cognitive activities such as summarizing the contents they have learned that day. They also tried to understand and improve the learning habit, attitude, and learning state. Teachers were also able to communicate with students by providing individual questions and feedback through student's individual Self-reported Evaluation.

• International Journal of Precision Engineering and Manufacturing
• /
• v.1 no.1
• /
• pp.79-83
• /
• 2000

Although it is well known that fuzzy learning controller is powerful for nonlinear systems, it is very difficult to apply a learning method if they are unstable. An unstable system diverges for impulse input. This divergence makes it difficult to learn the rules unless we can find the initial rules to make the system table prior to learning. Therefore, we introduced LQR(Linear Quadratic Regulator) technique to stabilize the system. It is a state feedback control to move unstable poles of a linear system to stable ones. But, if the system is nonlinear or complicated to get a liner model, we cannot expect good results with only LQR. In this paper, we propose that the LQR law is derived from a roughly approximated linear model, and next the fuzzy controller is tuned by the adaptive on-line learning with the real nonlinear plant. This hybrid controller of LQR and fuzzy learning was superior to the LQR of a linearized model in unstable nonlinear systems.

• 제어로봇시스템학회:학술대회논문집
• /
• 1994.10a
• /
• pp.411-415
• /
• 1994

An adaptive learning control scheme by use of multilayer neural networks for compensating for uncertainties in nonlinear dynamic system is examined. Multilayer neural networks are introduced to map the uncertainties in nonlinear dynamics and perform nonlinear state feedback. Parameters of neural networks are adjusted by conventional back-propagation algorithms modified with the projection operation. Effectiveness of the proposed scheme for tracking control are demonstrated through computer simulations.

• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.297-298
• /
• 2007

Repetitive systems stand for a kind of systems that perform a simple task on a fixed pattern repetitively and are widely spread in industrial fields. Hence, those systems have been of much interests by many researchers, especially in the field of iterative learning control (ILC). In this paper, we propose a finite-horizon tracking control scheme for linear time-varying repetitive systems with uncertain initial conditions. The scheme is derived both analytically and numerically for state-feedback systems and only numerically for output-feedback systems. Then, it is extended to stable systems with input constraints. All numerical schemes are developed in the forms of linear matrix inequalities. A distinguished feature of the proposed scheme from the existing iterative learning control is that the scheme guarantees the tracking performance exactly even under uncertain initial conditions. The simulation results demonstrate the good performance of the proposed scheme.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.39 no.12
• /
• pp.1306-1316
• /
• 1990

A controller using a multilayer neural network is proposed to the dynamic control of a PUMA 560 robot arm. This controller is developed based on an error back-propagation (BP) neural network. Since the neural network can model an arbitrary nonlinear mapping, it is used as a commanded feedforward torque generator. A Proportional Derivative (PD) feedback controller is used in parallel with the feedforward neural network to train the system. The neural network was trained by the current state of the manipulator as well as the PD feedback error torque. No a priori knowledge on system dynamics is needed and this information is rather implicitly stored in the interconnection weights of the neural network. In another experiment, the neural network was trained with the current, past and future positions only without any use of velocity sensors. Form this thim window of position values, BP network implicitly filters out the velocity and acceleration components for each joint. Computer simulation demonstrates such powerful characteristics of the neurocontroller as adaptation to changing environments, robustness to sensor noise, and continuous performance improvement with self-learning.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2003.06a
• /
• pp.1012-1017
• /
• 2003

The development of a fast, accurate, and inexpensive position-controlled pneumatic actuator that may be applied to a variety of practical positioning applications with various external loads is described in this paper. A novel modified pulso width modulation (MPWM) valve pulsing algorithm allows on/off solenoid valves to be used in place of costly servo valves. A comparison between the system response of standard PWM technique and that of the novel modified PWM technique shows that the control performance is significantly increased. A state feedback controller with position, velocity and acceleration feedback is successfully implemented as the continuous controller. Switching algorithm of control parameter using learning vector quantization neural network (LVQNN) is newly proposed. which estimates the external loads of the pneumatic actuator. The effectiveness of the proposed control algorithms are demonstrated through experiments with various loads.

• Journal of the Korean Society for Precision Engineering
• /
• v.20 no.8
• /
• pp.86-93
• /
• 2003

The development of a fast, accurate, and inexpensive position-controlled pneumatic actuator that may be applied to a variety of practical positioning applications with various external loads is described in this paper. A novel modified pulse width modulation (MPWM) valve pulsing algorithm allows on/off solenoid valves to be used in place of costly servo valves. A comparison between the system response of standard PWM technique and that of the novel modified PWM technique shows that the control performance is significantly increased. A state feedback controller with position, velocity and acceleration feedback is successfully implemented as the continuous controller. Switching algorithm of control parameter using learning vector quantization neural network (LVQNN) is newly proposed, which estimates the external loads of the pneumatic actuator. The effectiveness of the proposed control algorithms are demonstrated through experiments with various loads.