• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.252-255
• /
• 2020

Unsupervised deep learning methods have shown impressive results for the challenging monocular depth estimation task, a field of study that has gained attention in recent years. A common approach for this task is to train a deep convolutional neural network (DCNN) via an image synthesis sub-task, where additional views are utilized during training to minimize a photometric reconstruction error. Previous unsupervised depth estimation networks are trained within a fixed depth estimation range, irrespective of its possible range for a given image, leading to suboptimal estimates. To overcome this suboptimal limitation, we first propose an unsupervised adaptive depth estimation method guided by minimum and maximum (min-max) depth priors for a given input image. The incorporation of min-max depth priors can drastically reduce the depth estimation complexity and produce depth estimates with higher accuracy. Moreover, we propose a novel network architecture for adaptive depth estimation, called the AdaMM-DepthNet, which adopts the min-max depth estimation in its front side. Intensive experimental results demonstrate that the adaptive depth estimation can significantly boost up the accuracy with a fewer number of parameters over the conventional approaches with a fixed minimum and maximum depth range.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.830-841
• /
• 2023

Tokenization is the process of segmenting the input text into smaller units of text, and it is a preprocessing task that is mainly performed to improve the efficiency of the machine learning process. Various tokenization methods have been proposed for application in the field of natural language processing, but studies have primarily focused on efficiently segmenting text. Few studies have been conducted on the Korean language to explore what tokenization methods are suitable for document classification task. In this paper, an exploratory study was performed to find the most suitable tokenization method to improve the performance of a representative spatio-temporal document classifier in Korean. For the experiment, a convolutional neural network model was used, and for the final performance comparison, tasks were selected for document classification where performance largely depends on the tokenization method. As a tokenization method for comparative experiments, commonly used Jamo, Character, and Word units were adopted. As a result of the experiment, it was confirmed that the tokenization of word units showed excellent performance in the case of representative spatio-temporal document classification task where the semantic embedding ability of the token itself is important.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.40 no.2
• /
• pp.19-25
• /
• 2003

For a class of multi-input multi-output nonlinear systems which perform a given task repetitively, an extended type of a direct leaning control (DLC) is proposed using the information on the (vector) relative degree of a multi-input multi-output system. Existing DLC methods are observed to be applied to a limited class of systems with the relative degree one and a new DLC law is suggested which can be applied to systems having higher relative degree. Using the proposed control law, the control input corresponding to the new desired output trajectory is synthesized directly based on the control inputs obtained from the learning process for other output trajectories. To show the validity and the performance of the proposed DLC, simulations are performed for trajectory tracking control of a two-axis SCARA robot.

• Journal of the Ergonomics Society of Korea
• /
• v.31 no.1
• /
• pp.41-48
• /
• 2012

Objective: Companies want to expand their business by providing their services at other devices and new services based upon existing services. Therefore, they look for building brand identity by providing same experience throughout devices and services. Background: Many services are available to use at multiple devices including mobile phones, tablet, personal computers, and televisions, thanks to proliferation of n-screen and cloud technology. Method: It was discussed that consistency, which emphasizes the regularity and has been one of essential aspects of user interface design, seems not effective to be applied to n-screen services, owing to different screen size, input and output peripherals, usage environment and users' attitude. Results: A new definition of same experience among different devices and services, called coherence, was introduced and abstraction levels of user interfaces were proposed as the denominator of defining coherence. Then types of users' task knowledge at each abstraction level were discussed with examples. Conclusion: This paper concluded by discussing design requirements for designing coherent user interfaces among devices and services.

• The Journal of Korea Robotics Society
• /
• v.15 no.4
• /
• pp.398-403
• /
• 2020

Robot's throwing control is difficult to accurately calculate because of air resistance and rotational inertia, etc. This complexity can be solved by using machine learning. Reinforcement learning using reward function puts limit on adapting to new environment for robots. Therefore, this paper applied deep reinforcement learning using neural network without reward function. Throwing is evaluated as a success or failure. AI network learns by taking the target position and control policy as input and yielding the evaluation as output. Then, the task is carried out by predicting the success probability according to the target location and control policy and searching the policy with the highest probability. Repeating this task can result in performance improvements as data accumulates. And this model can even predict tasks that were not previously attempted which means it is an universally applicable learning model for any new environment. According to the data results from 520 experiments, this learning model guarantees 75% success rate.

• Journal of information and communication convergence engineering
• /
• v.9 no.2
• /
• pp.161-165
• /
• 2011

Recently, According to increasing interest to original sound Karaoke instrument, MIDI type karaoke manufacturer attempt to make more cheap method instead of original recoding method. Separating technique for singing voice from music accompaniment is very useful in such equipment. We propose a system to separate singing voice from music accompaniment for stereo recordings. Our system consists of three stages. The first stage is a spectral change detector. The second stage classifies an input into vocal and non vocal portions by using GMM classifier. The last stage is a selective frequency separation stage. The results of removed by listening test from the results for computer based extraction simulation, spectrogram results show separation task successfully. Listening test with extracted MR from proposed system show vocal separating and removal task successfully.

• 제어로봇시스템학회:학술대회논문집
• /
• 1987.10a
• /
• pp.905-908
• /
• 1987

A human operator has the ability to control a complicated system such as a gantry crane, an aircraft and a remote manipulator after enough training and learning. In this article, we attempt the positioning experiment of a flexible arm by a human operator. Flexible arm has nonlinearlity and infinite-degrees of freedom in general; thus it is difficult to obtain a control input. The operator interprets a given task and finds the procedure of operations. He devises an effective way of achieving the goal on the basis of his experience and knowledge about the task.

• English Language & Literature Teaching
• /
• v.12 no.4
• /
• pp.31-68
• /
• 2006

The present research investigates the acquisition of the English locative alternation by Korean EFL learners, which poses a learnability paradox, taking Pinker's framework of learnability theory as its basis. It addresses two questions (1) how lexical knowledge is represented initially and at different levels of interlanguage development and (2) what kinds of difficulty Korean learners find in the acquisition of English locative verbs and their constructions. Three groups of learners at different proficiency levels with a control group of English native speakers are examined by two instruments: elicited production task and grammaticality judgment task. According to different levels of proficiency, the learners exhibit gradual sensitivity to a change-of-state meaning and obtain complete perception of the meanings of locative verbs (manner-of-motion and change-of-state) and their constructions. Overgeneralization errors are observed in their performance. The errors are due to misinterpretations of particular lexical items in conjunction with the universal linking rules. More fundamental cause of difficulty is accounted for by partial use of learning mechanisms, caused by insufficient L2 input.

• Asia pacific journal of information systems
• /
• v.3 no.2
• /
• pp.117-144
• /
• 1993

MIS researches involve, in particular, the problems of developing appropriate IS for given MIS environments and lay emphasis on identifying prominant IS characteristics pertaining to the system in issue. This study deals with the IS characteristics for the management task of "planning". It investigates empirically the impacts of IS characteristics, under a set of proposed hypotheses, with particular respect to users' information satisfaction and performance in carrying out the task of planning. Futhermore, it examines whether the IS characteristics have different effects depending on the planning modes on the part of IS users. Based on a group of hypotheses accepted, the IS characteristics for planning decision making are not only proposed according to planning modes: preactive, reactive and proactive, but also classified into six subsystem categories: input, process, ouput, storage, interface and communication. Finally, the implications of the findings are discussed.

• Journal of the Ergonomics Society of Korea
• /
• v.24 no.1
• /
• pp.19-26
• /
• 2005

Mobile internet access using such devices as PDAs and cellular phones becomes popular as mobile technologies grow. However, the characteristics of small screen devices such as small screen size and pen-based input cause many usability problems. In this study, a human factors experiment was conducted to identify the factors affecting the usability of information search on a PDA. Factors manipulated in the experiment included use of wheel equipment, scroll dimension, and screen orientation. Task completion time, error frequency, and subjective satisfaction level were measured. In addition, various users' behavioral patterns such as scanning routes and mainly used scrolling methods were analyzed. Scroll dimension has a significant effect on task completion time. Scroll equipment and screen orientation affect subjective satisfaction level. The results could be applied to designing information structure of web sites for mobile use by providing vertical scroll and using external scroll equipments.