The search functionality is under construction.
The search functionality is under construction.

Author Search Result

[Author] Tingting HE(3hit)

1-3hit
  • A Global Deep Reranking Model for Semantic Role Classification

    Haitong YANG  Guangyou ZHOU  Tingting HE  Maoxi LI  

     
    LETTER-Natural Language Processing

      Pubricized:
    2021/04/15
      Vol:
    E104-D No:7
      Page(s):
    1063-1066

    The current approaches to semantic role classification usually first define a representation vector for a candidate role and feed the vector into a deep neural network to perform classification. The representation vector contains some lexicalization features like word embeddings, lemmar embeddings. From linguistics, the semantic role frame of a sentence is a joint structure with strong dependencies between arguments which is not considered in current deep SRL systems. Therefore, this paper proposes a global deep reranking model to exploit these strong dependencies. The evaluation experiments on the CoNLL 2009 shared tasks show that our system can outperforms a strong local system significantly that does not consider role dependency relations.

  • A Controlled Retransmission Scheme for Burst Segmentation in OBS Networks on the Consideration of Path Relevance

    Rui HOU  Tingting HE  Mingming ZHENG  Tengyue MAO  

     
    PAPER-Systems and Control

      Vol:
    E98-A No:2
      Page(s):
    676-683

    In this paper, we propose a controlled retransmission scheme in optical burst switching (OBS) networks. Different from previous works in the literature, we set a different value to retransmission probability at each contention and propose a retransmission analytical model for burst segmentation contention resolution scheme. In addition, we consider the effect of relevance in traffic come from multiple paths. We take into account the load at each link (include the given links and the other correlated links taking traffic) due to both the fresh and the retransmitted traffic and calculate the path blocking probability and the byte loss probability (ByLP) in cases of without and with full- wavelength conversion to evaluate the network performance. An extensive simulation is proposed to validate our analytical model, and results have shown that both path blocking probability and ByLP are affected by the load and the retransmission probability in each contention along a path and the correlated traffic carried links on the path.

  • Adversarial Domain Adaptation Network for Semantic Role Classification

    Haitong YANG  Guangyou ZHOU  Tingting HE  Maoxi LI  

     
    PAPER-Natural Language Processing

      Pubricized:
    2019/09/02
      Vol:
    E102-D No:12
      Page(s):
    2587-2594

    In this paper, we study domain adaptation of semantic role classification. Most systems utilize the supervised method for semantic role classification. But, these methods often suffer severe performance drops on out-of-domain test data. The reason for the performance drops is that there are giant feature differences between source and target domain. This paper proposes a framework called Adversarial Domain Adaption Network (ADAN) to relieve domain adaption of semantic role classification. The idea behind our method is that the proposed framework can derive domain-invariant features via adversarial learning and narrow down the gap between source and target feature space. To evaluate our method, we conduct experiments on English portion in the CoNLL 2009 shared task. Experimental results show that our method can largely reduce the performance drop on out-of-domain test data.