Title | ||
---|---|---|
A recurrent neural network language modeling framework for extractive speech summarization |
Abstract | ||
---|---|---|
Extractive speech summarization, with the purpose of automatically selecting a set of representative sentences from a spoken document so as to concisely express the most important theme of the document, has been an active area of research and development. A recent school of thought is to employ the language modeling (LM) approach for important sentence selection, which has proven to be effective for performing speech summarization in an unsupervised fashion. However, one of the major challenges facing the LM approach is how to formulate the sentence models and accurately estimate their parameters for each spoken document to be summarized. This paper presents a continuation of this general line of research and its contribution is two-fold. First, we propose a novel and effective recurrent neural network language modeling (RNNLM) framework for speech summarization, on top of which the deduced sentence models are able to render not only word usage cues but also long-span structural information of word co-occurrence relationships within spoken documents, getting around the need for the strict bag-of-words assumption. Second, the utilities of the method originated from our proposed framework and several widely-used unsupervised methods are analyzed and compared extensively. A series of experiments conducted on a broadcast news summarization task seem to demonstrate the performance merits of our summarization method when compared to several state-of-the-art existing unsupervised methods. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1109/ICME.2014.6890220 | ICME |
Keywords | Field | DocType |
signal representation,language modeling,speech recognition,word cooccurrence relationship,bag-of-word assumption,word processing,parameter estimation,rnnlm framework,extractive speech summarization,feature extraction,recurrent neural network language modeling framework,lm approach,broadcast news summarization,sentence selection model,recurrent neural nets,spoken document summarization,feature selection,recurrent neural network,sentence representation,long-span structural information,speech summarization,word usage cues,data models,speech,vectors,measurement,recurrent neural networks | Word usage,Data modeling,Computer science,Continuation,Recurrent neural network,Time delay neural network,Artificial intelligence,Natural language processing,Language model,Automatic summarization,Pattern recognition,Sentence,Machine learning | Conference |
ISSN | Citations | PageRank |
1945-7871 | 2 | 0.36 |
References | Authors | |
0 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kuan-Yu Chen | 1 | 450 | 55.78 |
Shih-Hung Liu | 2 | 66 | 14.53 |
Berlin Chen | 3 | 151 | 34.59 |
Hsin-min Wang | 4 | 1201 | 129.62 |
Wen-Lian Hsu | 5 | 1701 | 198.40 |
Hsin-hsi Chen | 6 | 2267 | 233.93 |