2. Stats
•173 long papers
•105 as oral
•68 as poster presentations
•692 total long paper submissions
•145 short papers
•50 as oral
•95 as poster presentations
•648 total short paper submissions
•13 TACL papers
•7 Student Research Workshop papers
•25 system demonstrations
•8 tutorials
•15 workshops
4. A Computational Approach to Automatic
Prediction of Drunk-Texting
Alcohol abuse may lead to unsociable behavior such
as crime, drunk driving, or privacy leaks. We introduce
automatic drunk-texting prediction as the task of
identifying whether a text was written when under
the influence of alcohol. We experiment with tweets
labeled using hashtags as distant supervision. Our
classifiers use a set of N-gram and stylistic features to
detect drunk tweets. Our observations present the
first quantitative evidence that text contains signals
that can be exploited to detect drunk-texting.
• Dataset 1 (2435 drunk, 762 sober)
• #drunk, #drank, #imdrunk
• #notdrunk, #imnotdrunk, #sober
• Dataset 2 (2435 drunk, 5644 sober)
• Dataset H (193 drunk, 317 sober) http://ej.uz/Drunk-Texting
6. Modeling Argument Strength in Student Essays
While recent years have seen a surge of interest in automated essay grading, including
work on grading essays with respect to particular dimensions such as prompt
adherence, coherence, and technical quality, there has been relatively little work on
grading the essay dimension of argument strength, which is arguably the most
important aspect of argumentative essays. We introduce a new corpus of
argumentative student essays annotated with argument strength scores and propose a
supervised, feature-rich approach to automatically scoring the essays along this
dimension. Our approach significantly outperforms a baseline that relies solely on
heuristically applied sentence argument function labels by up to 16.1%.
http://ej.uz/ArgStrInStudEssays
8. Driving ROVER with Segment-based
ASR Quality Estimation
ROVER is a widely used method to combine the output of multiple
automatic speech recognition (ASR) systems. Though effective, the basic approach
and its variants suffer from potential drawbacks: i) their results depend on the
order in which the hypotheses are used to feed the combination process, ii) when
applied to combine long hypotheses, they disregard possible differences in
transcription quality at local level, iii) they often rely on word confidence information.
We address these issues by proposing a segment-based ROVER in which hypothesis
ranking is obtained from a confidence-independent ASR quality estimation method.
Our results on English data from the IWSLT2012 and IWSLT2013 evaluation
campaigns significantly outperform standard ROVER and approximate two strong
oracles.
http://ej.uz/ROVER-SegASR-QEst
9. Driving ROVER with Segment-based
ASR Quality Estimation
1. Split the utterance into segments (ideally at sentence level);
2. For each segment, automatically estimate the quality (e.g. in terms of WER) of the
corresponding M (segment-level) hypotheses;
3. Use the estimates to rank the hypotheses and feed ROVER based on the ranking;
4. Reconstruct the entire utterance transcription by concatenating the combined
segment level transcriptions produced by ROVER;
5. Measure the overall WER differences against standard ROVER and other oracles.
http://ej.uz/ROVER-SegASR-QEst
10. Multi-level Translation Quality Prediction
with QUEST++
This paper presents QUEST++ , an open source tool for quality estimation which
can predict quality for texts at word, sentence and document level. It also provides
pipelined processing, whereby prediction smade at a lower level (e.g. for words) can
be used as input to build models for predictions at a higher level (e.g. sentences).
QUEST++ allows the extraction of a variety of features, and provides machine
learning algorithms to build and test quality estimation models. Results on recent
datasets show that QUEST++ achieves state-of-the-art performance.
http://ej.uz/QUESTpp
• 148 sentence level features
• 40 word level features
• 67 document level features
13. Unsupervised Decomposition of a Multi-Author
Document Based on Naive-Bayesian Model
This paper proposes a new unsupervised method for decomposing a multi-author
document into authorial components. We assume that we do not know anything about
the document and the authors, except the number of the authors of that document.
The key idea is to exploit the difference in the posterior probability of the Naive-
Bayesian model to increase the precision of the clustering assignment and the accuracy
of the classification process of our method. Experimental results show that the
proposed method outperforms two state-of-the-art methods.
http://ej.uz/Mult-AuthDocDecomposition
14. Unsupervised Decomposition of a Multi-Author
Document Based on Naive-Bayesian Model
http://ej.uz/Mult-AuthDocDecomposition
15. Decomposition of a Multi-Author Document
• Step 1 Divide the document into segments of fixed length.
• Step 2 Represent the resulted segments as vectors using an appropriate feature set which can differentiate
the writing styles among authors.
• Step 3 Cluster the resulted vectors into l clusters using an appropriate clustering algorithm targeting on
achieving high recall rates.
• Step 4 Re-vectorize the segments using a different feature set to more accurately discriminate the segments
in each cluster.
• Step 5 Apply the ”Segment Elicitation Procedure” to select the best segments from each cluster to increase
the precision rates.
• Step 6 Re-vectorize all selected segments using another feature set that can capture the differences among
the writing styles of all sentences in a document.
• Step 7 Train the classifier using the Naive-Bayesian model.
• Step 8 Classify each sentence using the learned classifier.
• Step 9 Apply the ”Probability Indication Procedure” to increase the accuracy of the classification results
using five criteria.
http://ej.uz/Mult-AuthDocDecomposition
16. Automatic Identification of
Age-Appropriate Ratings of Song Lyrics
This paper presents a novel task, namely the
automatic identification of age-appropriate ratings
of a musical track, or album, based on its lyrics.
Details are provided regarding the construction of a
dataset of lyrics from 12,242 tracks across 1,798
albums along with age-appropriate ratings
obtained from various web resources, along with
results from various text classification experiments.
The best accuracy of 71.02% for classifying albums
by age groups is achieved by combining vector
space model and psycholinguistic features.
http://ej.uz/IDofSongAgeRatings
Statistics of the dataset:
17. Linguistic Harbingers of Betrayal:
A Case Study on an Online Strategy Game
Interpersonal relations are fickle, with close friendships
often dissolving into enmity. In this work, we explore
linguistic cues that presage such transitions by studying
dyadic interactions in an on-line strategy game where
players form alliances and break those alliances through
betrayal. We characterize friendships that are unlikely to
last and examine temporal patterns that foretell betrayal.
We reveal that subtle signs of imminent betrayal are
encoded in the conversational patterns of the dyad, even if
the victim is not aware of the relationship’s fate. In
particular, we find that lasting friendships exhibit a form of
balance that manifests itself through language. In contrast,
sudden changes in the balance of certain conversational
attributes—such as positive sentiment, politeness, or
focus on future planning—signal impending betrayal.
http://ej.uz/LinguisticBetrayal
18. Linguistic Harbingers of Betrayal
http://ej.uz/LinguisticBetrayal
Features for recognizing imminent betrayal:
in decreasing order
19. An analysis of the user occupational class
through Twitter content
Social media content can be used as a complementary source
to the traditional methods for extracting and studying
collective social attributes. This study focuses on the
prediction of the occupational class for a public user profile.
Our analysis is conducted on a new annotated corpus of
Twitter users, their respective job titles, posted textual
content and platform-related attributes. We frame our task as
classification using latent feature representations such as
word clusters and embeddings. The employed linear and,
especially, non-linear methods can predict a user’s
occupational class with strong accuracy for the coarsest level
of a standard occupation taxonomy which includes nine
classes. Combined with a qualitative assessment, the derived
results confirm the feasibility of our approach in inferring a
new user attribute that can be embedded in a multitude of
downstream applications. http://ej.uz/occupationalClass-Twitter
20. Occupational class through Twitter content
User level attributes for a Twitter user:Topics, represented by their most central and most frequent 10 words: