Learning bias of phonological alternation in children learning English

조진영 한국음운론학회 2017 음성·음운·형태론 연구 Vol.23 No.2

The present study aims to test the hypothesis in McCarthy (1998) and Hayes (2004) that output-to-output faithfulness constraints are highly ranked a priori which leads children to initially favor non-alternating forms but to eventually produce alternating forms as adults do. Using spontaneous speech data of six mother-child dyads (Demuth et al. 2006), I examined how word-medial /t/ and /d/ in the flapping environment were realized in suffixed words such as eating (base: eat) and hiding (base: hide). The results show that children indeed produced suffixed words with variants that are faithful to the corresponding segment in the base (e.g., ea[t]ing and hi[d]ing), especially in earlier stages of learning. Such characteristic speech of children cannot be fully explained by the influence of the input data that they receive from their caregivers. Thus, I conclude that children’s production of variants that are phonologically similar to the base is due to highly ranked output-to-output faithfulness constraints at the initial state of learning. Based on a set of OT constraints, the children’s grammar is modeled with maximum entropy grammar (Goldwater and Johnson 2003).

전자파 방사 내성 시험을 위한 최적화된 균일장 측정 방법에 관한 연구

조진영,문준철,정창규,Cho Jin-Young,Mun Jun-Chul,Jung Chang-Kyoo 한국전자파학회 2005 한국전자파학회논문지 Vol.16 No.6

본 논문에서는 전자파 방사 내성 시험의 불확실성 요인이 되는 '오버 및 언더 편차'와 그로 인해 발생하는 '오버 및 언더 테스팅'을 줄이고자 기존 균일장 측정 방법의 문제점을 비교 분석하고 '오버 및 언더 편차'를 최소화할 수 있는 새로운 균일장 측정 방법을 제안하였다. 성능이 다른 4개의 국내 시험 기관 무반사실에 이를 적용하여 기존 측정 방법과 비교 실험하였으며 6 dB을 초과하는 필드에 대해 최소 편차인 FDI(Field Deviation Index)와 비교 분석하였다. 실험에 대한 분석 결과 '오버 및 언더 편차'는 제안 방법이 FDI와 동일하게 나타났으며 특정 주파수에선 현재 국내 규격 및 국제 규격보다 4.9 dB, 2.37 dB 개선되었고 전체 주파수에서는 평균적으로 1.193 dB, 0.017 dB 개선되었다. '오버 및 언더 테스팅 포인트'는 제안 방법이 기존 방법보다 최대 926개, 10개 포인트가 적게 측정되었으며 균일 영역 중심부 4포인트에서는 최대 118개, 9개 적게 측정되었으며, 다른 포인트에 대한 비율 또한 가장 낮게 측정되어 시험의 불확실성의 요인을 최소화하였음을 확인하였다. In this paper, we reduce 'over and under deviation' and 'over and under testing' which have a cause of uncertainty in electromagnetic radiated immunity test. We analyze continued measurement method of field uniformity and propose new estimate method having the minimum uncertainty condition. We test comparison in 4 anechoic chambers and analyze FDI(Field Deviation Index) which is the minimum deviation excess 6 dB on fields. We obtain that the proposed test method improve 1.193 dB than domestic standard(KN 61000-4-3) and 0.017 dB than international standard (IEC61000-4-3) in every frequency. In the test result 'over and under testing point' have 926 point at every field area and have 118 point at center uniform area. These result know that proposed measurement method have 10 point less at every field area and 9 point less at center uniform area than continued measurement method. The proposed test method have few point more than continued measurement method. Therefore, we confirmed that the proposed test method have the minimum uncertainty in electromagnetic radiated immunity test.

공유 기억장치를 이용한 분산-다중프로세서의 구현

조진영 호남대학교 1988 호남대학교 학술논문집 Vol.9 No.-

This paper discribed interprocessor communication through a shared random access memory (RAM) in distributed multiprocessor system is presented. An arbiter which empolies a simple hardware and requires no external clock is proposed. The arbiter consists of a controller and scheduler. The controller receives a shared memory request signal from a processor and sends Grant or Wait signal back to the microprocessor according to a command from a scheduler. The scheduler schedules the shared memory accessing for each microprocessor. The microprocessors in the system need not be identical, but are required to have a Wait state and a fetch state indicator. The system can be increased up to 16 microprocessors and is good for real-time data acquisition and control application. A front-end communication system using the propsed system is also presented here.

디지탈 임피던스 미터의 설계

조진영 호남대학교 정보통신연구소 1992 정보통신연구 Vol.1 No.-

This paper discribes the digitalized impedence meter which measures the voltages accross a voltages divider consisting of a reference register and the unknown impedance converted so that their magnitude and phase angle are digitally determined. This method introduces impedance measurement where the current through the unknown impedance is not necessarily constant. Also this method can modified to the admittance measurement.

메모리 직접 전송법을 이용한 주파수 및 속도 측정 시스템

조진영 호남대학교 정보통신연구소 1993 정보통신연구 Vol.2 No.-

본 논문은 마이크로 프로세서의 DMA(직접 메모리전송)법에 의한 주파수와 속도 측정 시스템의 설계및 구현에 관한 연구이다. 제안한 주파수 측정법은 일정하거나 연장된 샘플링 시간 동안의 펄스 계수와 시간 측정에 관한 것으로, 자주 타이머(Free Run Timer)의 내용은 입력 펄스의 각 상승에지에서 트리거되어 주기억장치로 DMA전송법에 의해 전송된다. 제안한 DMA전송법은 단일칩으로 직접회로화한 마이크로컨트롤러와 함께 실시간 측정 시스템에 적합하며 여러 측정 과정상 예상되는 문제점을 상세하게 검토하였다. This paper discribes a microprocessor-based wide-range frequency and speed measurement system which was designed and implemented for wide-range frequency and speed measurement using DMA transfer method. A new frequency measurement method is based on both pulse counting and time measurement during constant or extended sampling time. The content of a free-running timer is transfered by a direct memory access(DMA) to a memory location after each rising edge of a synchronized input pulse,so the DMA word counter operates as a pulse counter. Several sources of measurement errors are discussed in detail. The DMA method is suitable for real-time speed control systems with an integrated DMA controller.

네트웍기반 실시간 피드백 제어시스템의 성능개선의 관한 연구

조진영 호남대학교 정보통신연구소 2001 정보통신연구 Vol.11 No.-

This paper described the performance of network based feedback control system in industrial realtime control. TCP and UDP protocol is known as a protocol that can be commonly used in non-realtime system because of non-predictability of network latency. The mathematical model of the network latency according to each network layer is defined. The minimum feedback loop time period of the proposed system is determined by its result. The CAN networks can achieve the real time communication for the industrial control applications then Ethernet. In the proposed system architecture, we analyse its traffic for the real-time control system of all messages, determine the time period of a transducer to be satisfied with the required real time specification. The result of this paper provide a useful information to design network based realtime feedback control system.

상호작용적 신매체의 선택 및 이용모델 연구

조진영 成均館大學校 大學院 1996 首善論集 Vol.21 No.-

Electrochemical Properties of Novel PEGDA Gel Elecrolytes

조진영,송민규,조병원,이희우 한국고분자학회 1998 한국고분자학회 학술대회 연구논문 초록집 Vol.1998 No.10

대학생의 치매에 대한 지식, 태도 및 교육요구도 분석

조진영 한국디지털정책학회 2016 디지털융복합연구 Vol.14 No.12

순차형 연관 메모리의 설계

조진영 호남대학교 정보통신연구소 1996 정보통신연구 Vol.5 No.-

본 논문은 순차식 연관메모리( sequential CAM )의 연산 기능 확장에 관한 것으로, 메모리 내의 워드중 크기(대소) 비교, 범의 비교, 최대치/최소치를 고속으로 찾기 위하여 메모리를 구성하는 모든 워드들을 동시에 한 비트씩 병렬(parallel by words and serial by bits) 로 모든 비교 대상의 대응 비트의 논리 연산으로 주어진 결과와 순차적 비교하므로, 어드레스를 입력하여 엑세스 할 수 있는 RAM(serial by words and parallel by bits)에서 보다 엑세스 횟수를 대폭 줄여 고속으로 처리할 수 있는 방법과 장치에 관한 것이다. This paper cescribes a design of multi-fumction sequential CAM(Comtent Addressable Memory). The designed sequential CAM can be variably configured to a configuration funtilon input: The CAM selects one item of the group which is composed of 'greater than'. 'less than'. 'equal to'. a mininum and a maximum. and detects a relationship any fields of data in memory. The designed sequential CAM is more suitable for realtime image processing and high speed sorting than conventional RAM(Random Access Memory) which is accessed by address