Northwestern Polytechnical
Audio Speech & Language Processing Group
Digital Signal Processing
  • English
您是第counter free hit unique web位访客


Wireless Communications Speech Processing Medical Applications


      近期,实验室在故事分割(Story Segmentation)研究方面的一篇论文“Broadcast News Story Segmentation Using Manifold Learning on Latent Topic Distributions”(作者:Xiaoming Lu, Lei Xie, Cheung-Chi Leung, Bin Ma and Haizhou Li)被第51届计算语言学联合会年度会议(The 51st Annual Meeting of the Association for Computational Linguistics,ACL2013)录用。ACL是自然语言处理领域的顶级会议之一,论文录用率在20%左右,会议享有很高声望。本届会议将于2013年8月4日至9日在保加利亚首都索菲亚召开。ACL2013大会网址:。

      论文题目:Broadcast News Story Segmentation Using Manifold Learning on Latent Topic Distributions

      论文摘要:We present an efficient approach forbroadcast news story segmentation using amanifold learning algorithm on latent topic distributions. The latent topic distribution estimated by Latent Dirichlet Allocation (LDA) is used to represent each textblock. We employ LaplacianEigenmaps (LE) to project the latent topic distributions into low-dimensional semantic representations while preserving the intrinsic local geometric structure. We evaluate two approaches employing LDA and probabilistic latent semantic analysis (PLSA)distributions respectively. The effects ofdifferent amounts of training data and different numbers of latent topics on the twoapproaches are studied. Experimental results show that our proposed LDA-basedapproach can outperform the corresponding PLSA-based approach. The proposedapproach provides the best performancewith the highest F1-measure of 0.7860.



  • 校园风光