• Title/Summary/Keyword: NEGATIVE

Search Result 31,339, Processing Time 0.05 seconds

Improving passage retrieval via negative sampling from semantic feature space (의미론적 feature 공간상에서의 negative sampling을 통한 검색 성능 개선)

  • Jeong-Doo Lee;Beomseok Hong;Wonseok Choi;Youngsub Han;Byoung-Ki Jeon;Seung-Hoon Na
    • Annual Conference on Human and Language Technology
    • /
    • 2022.10a
    • /
    • pp.146-149
    • /
    • 2022
  • 최근 검색 태스크에서는 좋은 negative sample을 얻는 방법론들이 적용되어 큰 성능 향상을 이뤘다. 하지만 좋은 negative sample 대부분의 방법들은 큰 계산 비용이 든다. 따라서 본 논문에서는 계산 비용이 적고 효과적인 negative sample을 얻기 위해 Mixed Gaussian Recurrent Chain (MGRC) sampling을 사용하여 feature 공간상에서 의미론적으로 유사한 feature를 얻고 이를 negative sample로 활용하여 기존 baseline 모델보다 좋은 성능을 얻었다.

  • PDF

ON THE LIMIT BEHAVIOR OF EXTENDED NEGATIVE QUADRANT DEPENDENCE

  • Baek, Jong-Il;Lee, Gil-Hwan
    • Honam Mathematical Journal
    • /
    • v.32 no.4
    • /
    • pp.689-699
    • /
    • 2010
  • We discuss in this paper the notions of extended negative quadrant dependence and its properties. We study a class of bivariate uniform distributions having extended negative quadrant dependence, which is derived by generalizing the uniform representation of a well-known Farlie-Gumbel-Morgenstern distribution. Finally, we also study the limit behavior on the extended negative quadrant dependence.

STRONG LAW OF LARGE NUMBERS FOR ASYMPTOTICALLY NEGATIVE DEPENDENT RANDOM VARIABLES WITH APPLICATIONS

  • Kim, Hyun-Chull
    • Journal of applied mathematics & informatics
    • /
    • v.29 no.1_2
    • /
    • pp.201-210
    • /
    • 2011
  • In this paper, we obtain the H$\`{a}$jeck-R$\`{e}$nyi type inequality and the strong law of large numbers for asymptotically linear negative quadrant dependent random variables by using this inequality. We also give the strong law of large numbers for the linear process under asymptotically linear negative quadrant dependence assumption.

A Programmable CMOS Negative Resistor using Bump Circuit (Bump 회로를 이용한 Programmable CMOS Negative Resistor)

  • Song, Han-Jung
    • Proceedings of the KIEE Conference
    • /
    • 2002.11c
    • /
    • pp.253-256
    • /
    • 2002
  • A programmable CMOS negative resistor has been designed and fabricated in a 0.5um double poly double metal technology. The proposed CMOS negative resistor consists of a positive feedback OTA and a bump circuit with Gaussian-like I-V curve. Measurements of the fabricated chip confirm that the proposed CMOS resistor shows various negative resistance according to control voltage.

  • PDF

The Study on Ending Repetition Construction ("-고 자시고"류 어미 반복 구성에 관한 연구)

  • Cho, Mihee
    • Korean Linguistics
    • /
    • v.76
    • /
    • pp.213-241
    • /
    • 2017
  • Korean verb 'Casi-', which is treated as negative element in dictionaries, is actually a placeholder shows indefinite use, and '-ko casiko' construction is understood as a member of a class of 'Ending repetition construction'. However, unlike among others, '-ko casiko' construction is now undergoing developement of its negative polarity. Because of its indefiniteness, '-ko casiko' construction change its meaning into a universal negative when it is used in negative sentences.

On the Paneitz-Branson Operator in Manifolds with Negative Yamabe Constant

  • Ali, Zouaoui
    • Kyungpook Mathematical Journal
    • /
    • v.62 no.4
    • /
    • pp.751-767
    • /
    • 2022
  • This paper deals with the Paneitz-Branson operator in compact Riemannian manifolds with negative Yamabe invariant. We start off by providing a new criterion for the positivity of the Paneitz-Branson operator when the Yamabe invariant of the manifold is negative. Another result stated in this paper is about the existence of a metric on a manifold of dimension 5 such that the Paneitz-Branson operator has multiple negative eigenvalues. Finally, we provide new inequalities related to the upper bound of the mean value of the Q-curvature.

Penalized Likelihood Regression with Negative Binomial Data with Unknown Shape Parameter

  • Kim, Young-Ju
    • Communications for Statistical Applications and Methods
    • /
    • v.14 no.1
    • /
    • pp.23-32
    • /
    • 2007
  • We consider penalized likelihood regression with data from the negative binomial distribution with unknown shape parameter. Smoothing parameter selection and asymptotically efficient low dimensional approximations are employed for negative binomial data along with shape parameter estimation through several different algorithms.