=Paper=
{{Paper
|id=Vol-1175/CLEF2009wn-QACLEF-BarraganEt2009
|storemode=property
|title=INAOE at QAST 2009: Evaluating the Usefulness of a Phonetic Codification of Transcriptions
|pdfUrl=https://ceur-ws.org/Vol-1175/CLEF2009wn-QACLEF-BarraganEt2009.pdf
|volume=Vol-1175
|dblpUrl=https://dblp.org/rec/conf/clef/Reyes-BarraganVM09
}}
==INAOE at QAST 2009: Evaluating the Usefulness of a Phonetic Codification of Transcriptions==
https://ceur-ws.org/Vol-1175/CLEF2009wn-QACLEF-BarraganEt2009.pdf
INAOE at QAST 2009: Evaluating the usefulness of a
phonetic codification of transcriptions
Alejandro Reyes-Barragán, Luis Villaseñor-Pineda and Manuel Montes-y-Gómez
Laboratory of Language Technologies
National Institute of Astrophysics, Optics and Electronics (INAOE), Mexico.
alejandroe@hotmail.com; {villasen, mmontesg}@inaoep.mx
Abstract
This paper describes the participation of the Laboratory of Language Technologies of INAOE at
the QAST 2009 track. This participation was mainly focused on the evaluation of an enriched rep-
resentation of transcriptions based on their phonetic codification. The idea that motivated the de-
velopment of this new representation was to reduce the impact of the transcription errors by char-
acterizing words with similar pronunciations through the same phonetic code. The achieved results
were unsatisfactory since they indicate the phonetic information had no impact on the answer ac-
curacy, even for those cases considering the transcriptions of lower quality.
Categories and Subject Descriptors
H.3 [Information Storage and Retrieval]: H.3.1 Content Analysis and Indexing; H.3.3 Information Search and
Retrieval; H.3.4 Systems and Software; H.3.7 Digital Libraries
General Terms
Measurement, Performance, Experimentation
Keywords
Question Answering, Spontaneous Speech Transcriptions, Phonetic Codification.
1 Introduction
Question Answering (QA) has become a promising research field whose aim is to provide more natural access to
information than traditional document retrieval techniques. In essence, a QA system is a kind of search engine
that allows users to pose questions using natural language instead of an artificial query language, and that returns
exact answers to the questions instead of a list of entire documents (Burger et al., 2001).
Most QA developments have focused on searching answers for written questions from collections containing
written texts. However, due to the great amount of information appearing in spoken documents, such as tele-
phone conversations, meetings, speeches and broadcast news, some recent approaches have begun to address the
task of QA in automatic speech transcriptions1 (Turno et al., 2008).
The initial idea to tackle this problem was to apply the same techniques used in traditional textual-based QA.
Nevertheless, this idea did not seem completely appropriate due to existence of (several) mistakes on the auto-
matic transcriptions.
The mistakes generated by an automatic speech recognition (ASR) system may be caused by different cir-
cumstances, ranging from the presence of noise in the recording to the presence of unknown terms (e.g., proper
nouns that are out of its dictionary), but in general, most of them are approximations of the phonetic production
of the original words. Based on this fact, in this paper we propose to enrich the representation of the transcrip-
tions by including their phonetic codification. The purpose of this representation is to reduce the impact of the
transcription errors by characterizing words with similar pronunciations through the same phonetic code. In
particular, we propose using the Soundex codes (Odell and Russell, 1918) to enrich the representation of tran-
scriptions.
It is important to mention that the idea of using a phonetic codification for indexing automatic transcriptions
of speech is not completely new. There was some early attempt to use the Soundex codes for indexing names
with the aim of finding all their pronunciation variants (Raghavan and Allan, 2004). Continuing this idea, we
applied the phonetic codification indiscriminately to all transcription words, and evaluated the impact of using
1 These transcriptions are the result of applying an automatic speech recognition (ASR) system.
Oral Document Question
Transcriptions Transcription
Document processing Question processing
NE recognition Question classification
Passage segmantation Phonetic codification
Phonetic codification Query representation
Passage representation
Information retrieval
Indexing
Enriched Enriched
representation representation
of passages of the question
Retrieval
Answer extraction
Answer selection
Candidate
Answers
Figure 1. General architecture of the proposed QAST system
this codification in the task of spoken document retrieval (Reyes-Barragán et al., 2008). In this paper we go a
step forward and evaluate the usefulness of this codification in the task of QA in speech transcriptions (QAST).
The rest of the paper is organized as follows. Section 2 describes the main modules of the proposed system.
Then, Section 3 presents the evaluation results. Finally, Section 4 gives our conclusions and describes some
ideas for future work.
2 System Description
Figure 1 shows the general architecture of our QAST system. This system follows the traditional architecture,
including modules for question processing, passage retrieval and answer extraction; nevertheless, it considers the
phonetic codification from queries and passages as additional information.
Given that our participation at QAST 2009 was mainly oriented to evaluated the usefulness of considering the
phonetic codification, we intentionally restricted the scope of system to English (tasks T1a and T1b), and to
respond questions whose answer is a named entity. In particular, we considered only three types of named enti-
ties: dates, quantities and general proper names.
The following sections describe the main modules of the proposed system.
2.1 Named Entity Recognition
The first step in the processing of the transcriptions considered the labeling of named entities. This process was
achieved by FreeLing (Carreras et al., 2004), a tool that allow recognizing dates, quantities and proper names.
Moreover, given that some documents are speeches from the European Parliament, it was necessary to add a set
of regular expressions to identify some colloquial expressions related to dates, such as: today, in the future, now,
a week ago, in the years to come, etc.
It is important to clarify that we did not apply the same NE recognition process for all given transcriptions2. In
particular, it was impossible to use FreeLing on C-transcriptions. However, given that our interest was to evalu-
ate the impact of using a phonetic codification, especially from transcriptions of low quality, we decided to apply
a naïve method assuming that the NE recognition problem was resolved. This method considered the identifica-
tion of named entities using an ad-hoc catalogue built from the entities recognized from the automatic transcrip-
tions A and B.
2 Each task has four different transcriptions, one manual and three automatic (A, B and C), being the transcription A the one
with the best quality and the C the one with the lowest.
2.2 Passage segmentation
Once finished the process of NE recognition, the transcriptions were segmented into passages of 24 words
(named entities were considered as single words), and, after that, adjacent passages were overlapped by 12
words. We considered passages of 24 words because it corresponds to the average length of the sentences of the
manual transcriptions.
It is important to comment that the application of this process was motivated by our aim of simplifying the fi-
nal process of answer extraction.
2.3 Phonetic codification
Phonetic codifications attempt to represent words with similar pronunciations by the same code. Among all
existing phonetic codification algorithms, Soundex is the most widely known. It was originally proposed for
dealing with the problem of having different spelling variations of the same name (e.g., Lewinsky vs. Lewin-
skey) (Odell and Russell, 1918), and since then it has been applied in several database applications for indexing
surnames, for instance, it has been used in the U.S. census.
The Soundex algorithm is based on the phonetic classification of human speech sounds (bilabial, labiodental,
dental, alveolar, velar, and glottal), which in turn are based on where we put our lips and tongue to make sounds.
The algorithm itself is straightforward since it does not require of backtracking or multiple passes over the input
word. This algorithm is as follows:
1. Capitalize all letters in the word and drop all punctuation marks.
2. Retain the first letter of the word.
3. Change all occurrence of the following letters to '0' (zero):
'A', E', 'I', 'O', 'U', 'H', 'W', 'Y'.
4. Change letters from the following sets into the given digit:
1 = 'B', 'F', 'P', 'V'
2 = 'C', 'G', 'J', 'K', 'Q', 'S', 'X', 'Z'
3 = 'D','T'
4 = 'L'
5 = 'M','N'
6 = 'R'
5. Remove all pairs of equal digits occurring beside each other from the string resulted after step (4).
6. Remove all zeros from the string that results from step (5)
7. Pad the string resulted from step (6) with trailing zeros and return only the first six positions. The output
code will be of the form .
Using this algorithm, both "Robert" and "Rupert" return the same string "R16300", whereas "Rubin" yields
"R15000". In our case (refer to Figure 1), we applied the Soundex algorithm to both passages and queries in
order to construct their phonetic codifications.
2.4 Query and passage representation
Once we obtained the phonetic codification from the passages and the query, we constructed their enriched rep-
resentations. These representations exclude non-content words but include information about named entities as
well as some phonetic codes. Table 1 shows an example of the enriched representation from a passage.
Table 1. Example of an enriched passage representation
Manual transcription I am starting to know what Frank Sinatra must have
(show it for reference) felt like as his farewell appearances staggered on
into his seventies
Automatic transcription I am starting to work I do not know what Frank Sinatra
must have felt like as his fellow appearances decided on in
the seventies
After elimination of starting work Frank Sinatra felt like fellow appear-
stopwords ances decided seventies
Phonetic codification S36352 W62000 F65200 S53600 F43000 L20000 F40000 A16522
D23300 S15320
Enriched representation starting work Frank Sinatra felt like fellow appear-
ances decided seventies S36352 W62000 F65200
S53600 F43000 L20000 F40000 A16522 D23300 S15320
The enriched representation of a query is very similar to that from a passage; nevertheless, it also includes in-
formation about the expected answer type. In our implementation we only consider three basic types of expected
answers (date, quantity and proper name), which exactly correspond to the kinds of NEs recognized across the
passages. The determination of the expected answer type was carried out in the question classification module
(refer to Figure 1) using a set of manually constructed regular expressions.
2.5 Passage retrieval and answer extraction
In order to take advantage of all information contained in the enriched representations of passages and queries
we decided using Indri (Strohman et al., 2005) as search engine. Indri allowed us to determine a difference im-
portance of each term by assigning them different weights in accordance to their type. For example, for the ques-
tion “when do we need this strategic program for cohesion?”, whose expected answer type is DATE, we built the
following query: (8.0 DATE) (2.0 strategic program cohesion) (1.0 S36322 P62650 C25000), giving much more
importance to the expected answer type (a weigh of 8.0), followed by the terms of the query (with a weight of
2.0) and finally to the phonetic codes (a weight of 1.0). These weights were determined empirically from the
available training data.
Once retrieved the passages by Indri, we extracted from each passage the entities corresponding to the ex-
pected answer type. All extracted answers were ordered in accordance to the position of their source passage (we
put first answers from the first passage, then answers extracted from the second passage and so on), and the top
five answers were selected as candidate answers.
3 Results
This section presents the experiments that were carried out and discuses the achieved results. As was previously
mentioned, our system was limited to English tasks T1a and T1b. The difference between these tasks is the kind
of queries, while queries from task T1a are written questions, the ones from task T1b correspond to transcrip-
tions from oral spontaneous questions.
In order to evaluate the relevance of the phonetic codifications we sent two runs for each transcription of both
tasks: run INAOE1 did not include phonetic information, whereas run INAOE2 employed the proposed enriched
representations.
Results corresponding to manual transcriptions are shown in Table 2. For task T1a, that uses written ques-
tions, the results are practically the same with and without the phonetic codifications. In other words, the inclu-
sion of the phonetic information was not advantageous, but neither it was prejudicial. In the case of the task T1b,
where the questions are oral transcriptions, it is possible to observe an improvement using the phonetic codifica-
tion, which indicates that the phonetic codification is useful and allow retrieving a greater amount of relevant
passages.
Table 2. Results for manual transcriptions
Without phonetic transcription With phonetic transcription
#Questions #Questions
Task Run id with at least 1 MRR ACC Run id with at least 1 MRR ACC
correct answer correct answer
T1a (written
INAOE1 54 0.36 27 % INAOE2 51 0.36 28 %
questions)
T1b (spontaneous
INAOE1 35 0.27 22 % INAOE2 47 0.34 26 %
oral questions)
Tables 3 and 4 show the results of including the phonetic information when automatic transcriptions were
used. The results indicate that the phonetic information had no impact on the achieved accuracy, even for those
cases considering the transcriptions of lower quality or oral questions3. This unexpected behavior may be due to
the fact that the phonetic codification reduces the size of the vocabulary, that is, there are less phonetic codes
than words. For example, in transcription A there are 4,032 different words and only 2,415 phonetic codes. This
reduction increased the number of “irrelevant” passages associated to each query, and, therefore, it produced an
unfavorable effect.
3
It is important to mention that due to the particular treatment of the named entities in transcription C, the results using this
transcription are not directly comparable with those from transcriptions A and B. However, it is possible to observe the ef-
fect caused by the inclusion of the phonetic codification, which is the main goal of this work.
Table 3. Results for automatic transcriptions from Task T1a
Without phonetic transcription With phonetic transcription
#Questions #Questions
Transcription Run id with at least 1 MRR ACC Run id with at least 1 MRR ACC
correct answer correct answer
A INAOE1 41 0.3 23 % INAOE2 42 0.29 22 %
B INAOE1 29 0.22 17 % INAOE2 30 0.22 17 %
C INAOE1 34 0.28 25 % INAOE2 35 0.28 24 %
Table 3. Results for automatic transcriptions from Task T1b
Without phonetic transcription With phonetic transcription
#Questions #Questions
Transcription Run id with at least 1 MRR ACC Run id with at least 1 MRR ACC
correct answer correct answer
A INAOE1 40 0.3 24 % INAOE2 41 0.29 23 %
B INAOE1 30 0.22 16 % INAOE2 31 0.22 16 %
C INAOE1 33 0.28 25 % INAOE2 34 0.27 23 %
4 Conclusions
The present work explored the impact of the phonetic codification in the task of question answering in speech
transcriptions. The proposed system enriched the representation of the automatic transcriptions with their corre-
sponding phonetic codes (Soundex codes). To focus our participation in the evaluation of the phonetic codifica-
tion, it was limited to questions whose answer was a named entity, simplifying in this way the processes of pas-
sage retrieval and answer extraction.
As it can be noticed from the results section, the phonetic codification –at least in the way we implemented–
did not produce any improvement on the answer accuracy. Despite these results, the use of the phonetic codifica-
tion should not be discarded completely; it is necessary to make a deeper analysis in order to identify its possible
advantages. This is supported by the fact that we achieved some improvement when the phonetic codification
was applied to manual transcriptions from the task T1b. In this case the performance was almost the same to that
achieved with written questions, indicating that it was possible to decrease –in some degree–the error introduced
by the automatic speech recognition system.
Acknowledgements. This work was done under partial support of CONACYT (Project Grant 83459). We also
like to thank to the CLEF organizing committee for the resources provided.
References
1. Burger, J. et al. Issues, Tasks and Program Structures to Roadmap Research in Question & Answering
(Q&A), in TREC 10, NIST 2001.
2. Carreras, X., Chao, I., Padró, Ll., Padró, M. FreeLing: An Open-Source Suite of Language Analyzers. Pro-
ceedings of the 4th International Conference on Languages Resources and Evaluation (LREC 2004). Euro-
pean Language Resources Association, 2004.
3. Odell, M.K., Russell, R.C.: U.S. Patent Numbers 1261167 (1918) and 1435663 (1922). Washington, D.C.:
U.S. Patent Office, 1918.
4. Raghavan, H., Allan, J. Using Soundex Codes for Indexing Names in ASR documents. In Proceedings of the
Workshop on Interdisciplinary Approaches to Speech Indexing and Retrieval at Human Language Technol-
ogy Conference and North American chapter of Association of Computational Linguistics, Boston, MA,
USA, pp. 22–27, 2004.
5. Reyes-Barragán, M. A., Villaseñor-Pineda, L., and Montes-y-Gómez, M. A Soundex-based Approach for
Spoken Document Retrieval. Mexican International Conference of Artificial Intelligence, MICAI 2008. Lec-
ture Notes in Artificial Intelligence 5317, Springer 2008.
6. Strohman, T., Metzler, D., Turtle, H., Croft, W.B. Indri: A Language-Model based Search Engine for Com-
plex Queries. In Proceedings of the International Conference on Intelligence Analysis, McLean, VA, May
2005.
7. Turmo, J., Comas, P.R., Rosset, S., Lamel, L., Moureau, N., and Mostefa, D. Overview of QAST 2008.
Proceedings of the CLEF 2008 Workshop on Cross-Language Information Retrieval and Evaluation, 2008.