Document (#43892)

Author

Yu, C.
Xue, H.
An, L.
Li, G.

Title

¬A lightweight semantic-enhanced interactive network for efficient short-text matching

Source

Journal of the Association for Information Science and Technology. 74(2023) no.2, S.283-300

Year

2023

Abstract

Knowledge-enhanced short-text matching has been a significant task attracting much attention in recent years. However, the existing approaches cannot effectively balance effect and efficiency. Effective models usually consist of complex network structures leading to slow inference speed and the difficulties of applications in actual practice. In addition, most knowledge-enhanced models try to link the mentions in the text to the entities of the knowledge graphs-the difficulties of entity linking decrease the generalizability among different datasets. To address these problems, we propose a lightweight Semantic-Enhanced Interactive Network (SEIN) model for efficient short-text matching. Unlike most current research, SEIN employs an unsupervised method to select WordNet's most appropriate paraphrase description as the external semantic knowledge. It focuses on integrating semantic information and interactive information of text while simplifying the structure of other modules. We conduct intensive experiments on four real-world datasets, that is, Quora, Twitter-URL, SciTail, and SICK-E. Compared with state-of-the-art methods, SEIN achieves the best performance on most datasets. The experimental results proved that introducing external knowledge could effectively improve the performance of the short-text matching models. The research sheds light on the role of lightweight models in leveraging external knowledge to improve the effect of short-text matching.

Content

Vgl.: https://asistdl.onlinelibrary.wiley.com/doi/10.1002/asi.24731.

Similar documents (content)

1. Agarwal, B.; Ramampiaro, H.; Langseth, H.; Ruocco, M.: ¬A deep network model for paraphrase detection in short text messages (2018) 0.19

   0.18529361 = sum of:
     0.18529361 = product of:
       0.6617629 = sum of:
         0.0236352 = weight(abstract_txt:performance in 5043) [ClassicSimilarity], result of:
           0.0236352 = score(doc=5043,freq=1.0), product of:
             0.0816693 = queryWeight, product of:
               1.0974133 = boost
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.016071936 = queryNorm
             0.28940126 = fieldWeight in 5043, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.0625 = fieldNorm(doc=5043)
         0.19114946 = weight(abstract_txt:paraphrase in 5043) [ClassicSimilarity], result of:
           0.19114946 = score(doc=5043,freq=3.0), product of:
             0.18108575 = queryWeight, product of:
               1.1554947 = boost
               9.7509775 = idf(docFreq=6, maxDocs=44218)
               0.016071936 = queryNorm
             1.0555743 = fieldWeight in 5043, product of:
               1.7320508 = tf(freq=3.0), with freq of:
                 3.0 = termFreq=3.0
               9.7509775 = idf(docFreq=6, maxDocs=44218)
               0.0625 = fieldNorm(doc=5043)
         0.06113581 = weight(abstract_txt:network in 5043) [ClassicSimilarity], result of:
           0.06113581 = score(doc=5043,freq=3.0), product of:
             0.122144274 = queryWeight, product of:
               1.6437017 = boost
               4.6236176 = idf(docFreq=1179, maxDocs=44218)
               0.016071936 = queryNorm
             0.5005213 = fieldWeight in 5043, product of:
               1.7320508 = tf(freq=3.0), with freq of:
                 3.0 = termFreq=3.0
               4.6236176 = idf(docFreq=1179, maxDocs=44218)
               0.0625 = fieldNorm(doc=5043)
         0.06031489 = weight(abstract_txt:semantic in 5043) [ClassicSimilarity], result of:
           0.06031489 = score(doc=5043,freq=2.0), product of:
             0.15251142 = queryWeight, product of:
               2.1208355 = boost
               4.4743214 = idf(docFreq=1369, maxDocs=44218)
               0.016071936 = queryNorm
             0.39547786 = fieldWeight in 5043, product of:
               1.4142135 = tf(freq=2.0), with freq of:
                 2.0 = termFreq=2.0
               4.4743214 = idf(docFreq=1369, maxDocs=44218)
               0.0625 = fieldNorm(doc=5043)
         0.07792405 = weight(abstract_txt:text in 5043) [ClassicSimilarity], result of:
           0.07792405 = score(doc=5043,freq=2.0), product of:
             0.21801163 = queryWeight, product of:
               3.3543985 = boost
               4.0438666 = idf(docFreq=2106, maxDocs=44218)
               0.016071936 = queryNorm
             0.3574307 = fieldWeight in 5043, product of:
               1.4142135 = tf(freq=2.0), with freq of:
                 2.0 = termFreq=2.0
               4.0438666 = idf(docFreq=2106, maxDocs=44218)
               0.0625 = fieldNorm(doc=5043)
         0.1159433 = weight(abstract_txt:short in 5043) [ClassicSimilarity], result of:
           0.1159433 = score(doc=5043,freq=1.0), product of:
             0.32000965 = queryWeight, product of:
               3.434728 = boost
               5.79699 = idf(docFreq=364, maxDocs=44218)
               0.016071936 = queryNorm
             0.36231187 = fieldWeight in 5043, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               5.79699 = idf(docFreq=364, maxDocs=44218)
               0.0625 = fieldNorm(doc=5043)
         0.13166024 = weight(abstract_txt:matching in 5043) [ClassicSimilarity], result of:
           0.13166024 = score(doc=5043,freq=1.0), product of:
             0.34831253 = queryWeight, product of:
               3.5834005 = boost
               6.047913 = idf(docFreq=283, maxDocs=44218)
               0.016071936 = queryNorm
             0.37799457 = fieldWeight in 5043, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               6.047913 = idf(docFreq=283, maxDocs=44218)
               0.0625 = fieldNorm(doc=5043)
       0.28 = coord(7/25)
   

2. Draper, S.W.; Dunlop, M.D.: New IR - new evaluation : the impact of interaction and multimedia on information retrieval and its evaluation (1997) 0.12

   0.1171003 = sum of:
     0.1171003 = product of:
       0.48791793 = sum of:
         0.0236352 = weight(abstract_txt:performance in 2462) [ClassicSimilarity], result of:
           0.0236352 = score(doc=2462,freq=1.0), product of:
             0.0816693 = queryWeight, product of:
               1.0974133 = boost
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.016071936 = queryNorm
             0.28940126 = fieldWeight in 2462, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.0625 = fieldNorm(doc=2462)
         0.03810854 = weight(abstract_txt:effect in 2462) [ClassicSimilarity], result of:
           0.03810854 = score(doc=2462,freq=1.0), product of:
             0.112296835 = queryWeight, product of:
               1.28684 = boost
               5.4296865 = idf(docFreq=526, maxDocs=44218)
               0.016071936 = queryNorm
             0.3393554 = fieldWeight in 2462, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               5.4296865 = idf(docFreq=526, maxDocs=44218)
               0.0625 = fieldNorm(doc=2462)
         0.06120772 = weight(abstract_txt:interactive in 2462) [ClassicSimilarity], result of:
           0.06120772 = score(doc=2462,freq=1.0), product of:
             0.17630063 = queryWeight, product of:
               1.9747556 = boost
               5.5548496 = idf(docFreq=464, maxDocs=44218)
               0.016071936 = queryNorm
             0.3471781 = fieldWeight in 2462, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               5.5548496 = idf(docFreq=464, maxDocs=44218)
               0.0625 = fieldNorm(doc=2462)
         0.047612824 = weight(abstract_txt:models in 2462) [ClassicSimilarity], result of:
           0.047612824 = score(doc=2462,freq=1.0), product of:
             0.16412646 = queryWeight, product of:
               2.2001135 = boost
               4.6415744 = idf(docFreq=1158, maxDocs=44218)
               0.016071936 = queryNorm
             0.2900984 = fieldWeight in 2462, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.6415744 = idf(docFreq=1158, maxDocs=44218)
               0.0625 = fieldNorm(doc=2462)
         0.23942961 = weight(abstract_txt:lightweight in 2462) [ClassicSimilarity], result of:
           0.23942961 = score(doc=2462,freq=1.0), product of:
             0.4376917 = queryWeight, product of:
               3.1115055 = boost
               8.752448 = idf(docFreq=18, maxDocs=44218)
               0.016071936 = queryNorm
             0.547028 = fieldWeight in 2462, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               8.752448 = idf(docFreq=18, maxDocs=44218)
               0.0625 = fieldNorm(doc=2462)
         0.07792405 = weight(abstract_txt:text in 2462) [ClassicSimilarity], result of:
           0.07792405 = score(doc=2462,freq=2.0), product of:
             0.21801163 = queryWeight, product of:
               3.3543985 = boost
               4.0438666 = idf(docFreq=2106, maxDocs=44218)
               0.016071936 = queryNorm
             0.3574307 = fieldWeight in 2462, product of:
               1.4142135 = tf(freq=2.0), with freq of:
                 2.0 = termFreq=2.0
               4.0438666 = idf(docFreq=2106, maxDocs=44218)
               0.0625 = fieldNorm(doc=2462)
       0.24 = coord(6/25)
   

3. Potha, N.; Stamatatos, E.: Improving author verification based on topic modeling (2019) 0.12

   0.1160076 = sum of:
     0.1160076 = product of:
       0.41431287 = sum of:
         0.0236352 = weight(abstract_txt:performance in 5385) [ClassicSimilarity], result of:
           0.0236352 = score(doc=5385,freq=1.0), product of:
             0.0816693 = queryWeight, product of:
               1.0974133 = boost
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.016071936 = queryNorm
             0.28940126 = fieldWeight in 5385, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.0625 = fieldNorm(doc=5385)
         0.02920366 = weight(abstract_txt:most in 5385) [ClassicSimilarity], result of:
           0.02920366 = score(doc=5385,freq=1.0), product of:
             0.11848249 = queryWeight, product of:
               1.8693167 = boost
               3.943693 = idf(docFreq=2328, maxDocs=44218)
               0.016071936 = queryNorm
             0.24648081 = fieldWeight in 5385, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               3.943693 = idf(docFreq=2328, maxDocs=44218)
               0.0625 = fieldNorm(doc=5385)
         0.042649068 = weight(abstract_txt:semantic in 5385) [ClassicSimilarity], result of:
           0.042649068 = score(doc=5385,freq=1.0), product of:
             0.15251142 = queryWeight, product of:
               2.1208355 = boost
               4.4743214 = idf(docFreq=1369, maxDocs=44218)
               0.016071936 = queryNorm
             0.2796451 = fieldWeight in 5385, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.4743214 = idf(docFreq=1369, maxDocs=44218)
               0.0625 = fieldNorm(doc=5385)
         0.08158978 = weight(abstract_txt:external in 5385) [ClassicSimilarity], result of:
           0.08158978 = score(doc=5385,freq=1.0), product of:
             0.21353722 = queryWeight, product of:
               2.1733177 = boost
               6.113391 = idf(docFreq=265, maxDocs=44218)
               0.016071936 = queryNorm
             0.38208693 = fieldWeight in 5385, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               6.113391 = idf(docFreq=265, maxDocs=44218)
               0.0625 = fieldNorm(doc=5385)
         0.047612824 = weight(abstract_txt:models in 5385) [ClassicSimilarity], result of:
           0.047612824 = score(doc=5385,freq=1.0), product of:
             0.16412646 = queryWeight, product of:
               2.2001135 = boost
               4.6415744 = idf(docFreq=1158, maxDocs=44218)
               0.016071936 = queryNorm
             0.2900984 = fieldWeight in 5385, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.6415744 = idf(docFreq=1158, maxDocs=44218)
               0.0625 = fieldNorm(doc=5385)
         0.111698315 = weight(abstract_txt:enhanced in 5385) [ClassicSimilarity], result of:
           0.111698315 = score(doc=5385,freq=1.0), product of:
             0.28977466 = queryWeight, product of:
               2.923385 = boost
               6.167458 = idf(docFreq=251, maxDocs=44218)
               0.016071936 = queryNorm
             0.38546613 = fieldWeight in 5385, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               6.167458 = idf(docFreq=251, maxDocs=44218)
               0.0625 = fieldNorm(doc=5385)
         0.07792405 = weight(abstract_txt:text in 5385) [ClassicSimilarity], result of:
           0.07792405 = score(doc=5385,freq=2.0), product of:
             0.21801163 = queryWeight, product of:
               3.3543985 = boost
               4.0438666 = idf(docFreq=2106, maxDocs=44218)
               0.016071936 = queryNorm
             0.3574307 = fieldWeight in 5385, product of:
               1.4142135 = tf(freq=2.0), with freq of:
                 2.0 = termFreq=2.0
               4.0438666 = idf(docFreq=2106, maxDocs=44218)
               0.0625 = fieldNorm(doc=5385)
       0.28 = coord(7/25)
   

4. Jones, S.; Hancock-Beaulieu, M.: Support strategies for interactive thesaurus navigation (1994) 0.11

   0.11290596 = sum of:
     0.11290596 = product of:
       0.4704415 = sum of:
         0.035452798 = weight(abstract_txt:performance in 7734) [ClassicSimilarity], result of:
           0.035452798 = score(doc=7734,freq=1.0), product of:
             0.0816693 = queryWeight, product of:
               1.0974133 = boost
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.016071936 = queryNorm
             0.43410188 = fieldWeight in 7734, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.09375 = fieldNorm(doc=7734)
         0.04357259 = weight(abstract_txt:improve in 7734) [ClassicSimilarity], result of:
           0.04357259 = score(doc=7734,freq=1.0), product of:
             0.093706 = queryWeight, product of:
               1.175505 = boost
               4.9599204 = idf(docFreq=842, maxDocs=44218)
               0.016071936 = queryNorm
             0.46499252 = fieldWeight in 7734, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.9599204 = idf(docFreq=842, maxDocs=44218)
               0.09375 = fieldNorm(doc=7734)
         0.077647865 = weight(abstract_txt:difficulties in 7734) [ClassicSimilarity], result of:
           0.077647865 = score(doc=7734,freq=1.0), product of:
             0.13773516 = queryWeight, product of:
               1.4251586 = boost
               6.0133076 = idf(docFreq=293, maxDocs=44218)
               0.016071936 = queryNorm
             0.5637476 = fieldWeight in 7734, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               6.0133076 = idf(docFreq=293, maxDocs=44218)
               0.09375 = fieldNorm(doc=7734)
         0.091811575 = weight(abstract_txt:interactive in 7734) [ClassicSimilarity], result of:
           0.091811575 = score(doc=7734,freq=1.0), product of:
             0.17630063 = queryWeight, product of:
               1.9747556 = boost
               5.5548496 = idf(docFreq=464, maxDocs=44218)
               0.016071936 = queryNorm
             0.52076715 = fieldWeight in 7734, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               5.5548496 = idf(docFreq=464, maxDocs=44218)
               0.09375 = fieldNorm(doc=7734)
         0.048041735 = weight(abstract_txt:knowledge in 7734) [ClassicSimilarity], result of:
           0.048041735 = score(doc=7734,freq=1.0), product of:
             0.14423744 = queryWeight, product of:
               2.5260413 = boost
               3.5527887 = idf(docFreq=3442, maxDocs=44218)
               0.016071936 = queryNorm
             0.33307394 = fieldWeight in 7734, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               3.5527887 = idf(docFreq=3442, maxDocs=44218)
               0.09375 = fieldNorm(doc=7734)
         0.17391494 = weight(abstract_txt:short in 7734) [ClassicSimilarity], result of:
           0.17391494 = score(doc=7734,freq=1.0), product of:
             0.32000965 = queryWeight, product of:
               3.434728 = boost
               5.79699 = idf(docFreq=364, maxDocs=44218)
               0.016071936 = queryNorm
             0.5434678 = fieldWeight in 7734, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               5.79699 = idf(docFreq=364, maxDocs=44218)
               0.09375 = fieldNorm(doc=7734)
       0.24 = coord(6/25)
   

5. Hoeber, O.: ¬A study of visually linked keywords to support exploratory browsing in academic search (2022) 0.11

   0.107718185 = sum of:
     0.107718185 = product of:
       0.5385909 = sum of:
         0.029544 = weight(abstract_txt:performance in 644) [ClassicSimilarity], result of:
           0.029544 = score(doc=644,freq=1.0), product of:
             0.0816693 = queryWeight, product of:
               1.0974133 = boost
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.016071936 = queryNorm
             0.3617516 = fieldWeight in 644, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               4.63042 = idf(docFreq=1171, maxDocs=44218)
               0.078125 = fieldNorm(doc=644)
         0.06152411 = weight(abstract_txt:effectively in 644) [ClassicSimilarity], result of:
           0.06152411 = score(doc=644,freq=1.0), product of:
             0.13318118 = queryWeight, product of:
               1.4014004 = boost
               5.913062 = idf(docFreq=324, maxDocs=44218)
               0.016071936 = queryNorm
             0.461958 = fieldWeight in 644, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               5.913062 = idf(docFreq=324, maxDocs=44218)
               0.078125 = fieldNorm(doc=644)
         0.108200975 = weight(abstract_txt:interactive in 644) [ClassicSimilarity], result of:
           0.108200975 = score(doc=644,freq=2.0), product of:
             0.17630063 = queryWeight, product of:
               1.9747556 = boost
               5.5548496 = idf(docFreq=464, maxDocs=44218)
               0.016071936 = queryNorm
             0.61372995 = fieldWeight in 644, product of:
               1.4142135 = tf(freq=2.0), with freq of:
                 2.0 = termFreq=2.0
               5.5548496 = idf(docFreq=464, maxDocs=44218)
               0.078125 = fieldNorm(doc=644)
         0.040034775 = weight(abstract_txt:knowledge in 644) [ClassicSimilarity], result of:
           0.040034775 = score(doc=644,freq=1.0), product of:
             0.14423744 = queryWeight, product of:
               2.5260413 = boost
               3.5527887 = idf(docFreq=3442, maxDocs=44218)
               0.016071936 = queryNorm
             0.2775616 = fieldWeight in 644, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               3.5527887 = idf(docFreq=3442, maxDocs=44218)
               0.078125 = fieldNorm(doc=644)
         0.29928702 = weight(abstract_txt:lightweight in 644) [ClassicSimilarity], result of:
           0.29928702 = score(doc=644,freq=1.0), product of:
             0.4376917 = queryWeight, product of:
               3.1115055 = boost
               8.752448 = idf(docFreq=18, maxDocs=44218)
               0.016071936 = queryNorm
             0.683785 = fieldWeight in 644, product of:
               1.0 = tf(freq=1.0), with freq of:
                 1.0 = termFreq=1.0
               8.752448 = idf(docFreq=18, maxDocs=44218)
               0.078125 = fieldNorm(doc=644)
       0.2 = coord(5/25)