Available via license: CC BY 4.0
Content may be subject to copyright.
Sister Help: Data Augmentation for Frame-Semantic Role Labeling
Ayush Pancholy♦Miriam R. L. Petruck♥Swabha Swayamdipta♣
♦University of California, Berkeley
♥International Computer Science Institute, Berkeley
♣Allen Institute for Artificial Intelligence
{ayush.pancholy@,miriamp@icsi.}berkeley.edu
swabhas@allenai.org
Abstract
While FrameNet is widely regarded as a rich
resource of semantics in natural language pro-
cessing, a major criticism concerns its lack of
coverage and the relative paucity of its labeled
data compared to other commonly used lexi-
cal resources such as PropBank and VerbNet.
This paper reports on a pilot study to address
these gaps. We propose a data augmentation
approach, which uses existing frame-specific
annotation to automatically annotate other lex-
ical units of the same frame which are unan-
notated. Our rule-based approach defines the
notion of a sister lexical unit and generates
frame-specific augmented data for training.
We present experiments on frame-semantic
role labeling which demonstrate the impor-
tance of this data augmentation: we obtain a
large improvement to prior results on frame
identification and argument identification for
FrameNet, utilizing both full-text and lexico-
graphic annotations under FrameNet. Our find-
ings on data augmentation highlight the value
of automatic resource creation for improved
models in frame-semantic parsing.
1 Introduction
Among the challenges to Natural Language Pro-
cessing (NLP) systems is access to sufficient and
accurate information about the mapping between
form and meaning in language (Bender and Koller,
2020). While the number of resources that provide
such information has increased in recent decades,
producing these resources remains labor-intensive
and costly. This is particularly true for FrameNet
(
FN
;Ruppenhofer et al.,2016), a rich resource
of semantic annotations, popularized for semantic
role labeling (
SRL
) since the pioneering work of
Gildea and Jurafsky (2002).
Nevertheless, FrameNet still relies (almost ex-
clusively) on expert manual linguistic annotation,
resulting in far fewer annotations than other re-
sources that are able to scale up the annotation pro-
cess easily (FitzGerald et al.,2018). Moreover, the
carChuck bought a from for $2,000 .Jerry
BUYER
GOODS
SELLER
MONEY
carChuck purchased a from for $2,000 .
Jerry
BUYER
GOODS
SELLER
MONEY
buy.v
purchase.v
Commerce_buy
Figure 1: Illustration of Sister Lexical Units (Sister
LUs) with FrameNet annotation for the verbs buy
and purchase, which are defined in terms of the
Commerce_buy frame. The arguments of both LUs
are identical, both realizing the frame elements (FEs)
BUYER,GO OD S,SELLER, and MONEY. Targets in
the sentence that trigger the frame are the words bought
and purchased, respectively.
fine-grained annotations in FrameNet make them
harder to produce compared to other lexical re-
sources such as PropBank (Palmer et al.,2005)
and VerbNet (Kipper et al.,2000). As a result,
while SRL has continued to be a mainstay of core
NLP, with applications in event extraction, partic-
ipant tracking, machine translation, and question-
answering, far fewer SRL systems use FrameNet
for training, in contrast to PropBank.1
Our work seeks to address this low resource
problem in FrameNet in English by automatic data
augmentation. We leverage the FrameNet hierar-
chy where multiple lexical units are associated with
the same frame, but might have a different number
of annotations associated with them (§2). We pro-
pose a rule-based approach to transfer annotations
among sister LUs; see Fig. 1for an illustration. We
implement and extend the proposal by Petruck and
Swayamdipta (2019) to generate new annotation
for previously unannotated LUs in the same frame,
by exploiting the already identified target and frame
of a given lemma. We hypothesize that our rule-
based approach to transfer frame-specific annota-
tion to generate semantic role labels for example
1
PropBank annotations were also included in CoNLL
shared tasks in 2005 (Carreras and Màrquez,2005) and 2012
(Pradhan et al.,2012), and in OntoNotes (Pradhan et al.,2013).
arXiv:2109.07725v1 [cs.CL] 16 Sep 2021
sentences of lexical units (
LU
s) for which none
exist, can indeed result in a FrameNet with higher
coverage annotations, a state of affairs that can help
in training frame-semantic role labeling systems
(§3). Our experiments show that this is indeed
the case; models trained on augmented data are
able to surpass a strong baseline (OPEN-SESAME;
Swayamdipta et al.,2017) for frame and argument
identification (§4). We also provide a summary of
the types of linguistic errors in the augmented data
and the SRL task.
While our work provides a proof of concept, its
novelty lies in exploiting frame-specific annotation
to produce augmented data yielding new annota-
tion. Our implementation and augmented data are
publicly available.2
2 Background
2.1 FrameNet
FrameNet (Ruppenhofer et al.,2016) is a research
and resource development project in corpus-based
computational lexicography grounded in the prin-
ciples of
frame semantics
(Fillmore,1985). One
of the goals of this effort is documenting the
va-
lences
, i.e., the syntactic and semantic combina-
torial possibilities of each item analyzed. These
valence descriptions provide critical information
on the mapping between form and meaning; NLP
and more broadly, natural language understanding
(NLU) require such mapping.
At the heart of the work is the
semantic frame
, a
script-like knowledge structure that facilitates infer-
ence within and across events, situations, relations,
etc. FrameNet defines a semantic frame in terms
of its
frame elements
(FEs), or participants in the
scene that the frame captures; a
lexical unit
(LU)
is a pairing of a lemma and a frame, thus character-
izing that LU in terms of the frame that it evokes.
Valence descriptions derive from the annotation
of FEs, i.e.
semantic roles
, on example sentences
that illustrate the linguistic manifestation of the par-
ticipants in a scene for the target of analysis. The
first sentence in Figure 1illustrates annotation with
respect to the verb buy, which FN defines in terms
of the
Commerce_buy
frame, whose FEs are a
BUYER,SELLER,GOODS, and MONEY.3
2https://github.com/ayush-pancholy/
sister-help
3
Frame names appear in
Typewriter
font; FE names
are in SM AL L CAPS; and LUs are in italics.
The FrameNet
4
database (version 1.7) holds
1,224 frames, 10,478 frame-specific FEs (in lexical
frames),
5
over 13,500 LUs. The primary annota-
tions are in the form of nearly 203,000 manually
annotated sentences providing information about
the mapping between form and meaning in English;
these are called the
lexicographic
annotations. A
subset of the sentences with lexicographic anno-
tations contain annotations for all possible frames
of meaning in them; FrameNet calls this subset
full-text annotations. Despite FrameNet’s costly
annotation efforts, 38% of the over 13,500 LUs in
the database remain without annotation. That is, a
significant amount of LUs in the database have no
associated annotated sentences to serve as training
data for downstream NLP applications.
2.2 Frame Semantic Role Labeling
Semantic role labeling (
SRL)
involves automat-
ically labeling who did what to whom in a text
document. SRL systems facilitate developing NLP
applications, like question answering (Shen and La-
pata,2007), machine translation (Pedersen,2001),
and text summarization (Han et al.,2016), to name
but a few. Frame semantic role labeling (
frame-
SRL)
is a special type of SRL where the task is
to identify target tokens (which evoke a frame of
meaning), the frame itself (of target LUs) and FEs,
or semantic roles, in text. The relationship between
frames and their respective FEs requires that the
system first identify the correct frame (FrameID)
to identify the correct FEs (ArgID). Identifying the
frame incorrectly necessarily means incorrect FE
identification.
6
In practice, frame-SRL is usually
implemented as a pipeline of structured prediction
(Smith,2011) tasks: identifying targets and LUs,
frames, and finally the FEs (Das et al.,2010).
Using an early version of FN data (Johnson
et al.,2001), Gildea and Jurafsky (2002) devel-
oped the first SRL system, which also initiated
SRL as a now well-recognized task in the field.
Recent years have seen the development of sev-
eral (relatively) high-performing SRL systems. SE-
MAFOR (Das et al.,2014) uses a pipeline of dis-
crete, manually designed feature-based linear clas-
4https://framenet.icsi.berkeley.edu/
fndrupal/framenet_request_data
5
FrameNet distinguishes between lexical and non-lexical
frames, where the latter do not have associated annotation
sets.
6
Baker et al. (2007) also considered correct identification
of the spans of a dependent of a lexical unit, even if the frame
and by definition the FEs were identified incorrectly.
sifiers for target identification, frame identification
(FrameID), and argument identification (ArgID).
The core constraints of frame-semantic analyses,
e.g. not repeating core FEs, are satisfied via in-
teger linear programming. PathLSTM (Roth and
Lapata,2016) uses neural features to embed path
relationships between frames and FEs, in a pipeline
similar to that of SEMAFOR. Yang and Mitchell
(2017) use a joint model to leverage Propbank SRL
annotations to improve frame-SRL. OPEN-SESA ME
(Swayamdipta et al.,2017,2018) uses an uncon-
strained, neural approach in a pipeline like SE-
MAFOR and PathLSTM. Continuous representa-
tions and a sophisticated global model based on
semi-Markov conditional random fields (Sarawagi
and Cohen,2004) improve identifying arguments.
We employ OPEN-SESAM E
7
as the baseline of
choice in our experiments (§4). With the exception
of a few approaches, like that of Yang and Mitchell
(2017), most prior work uses only the full-text an-
notations for frame-SRL; we present experiments
considering the larger set of lexicographic annota-
tions as well.
3 Method: Data Augmentation with
Sister LUs
The relationship among LUs in a frame motivates
examining their use for paraphrasing, where one
LU replaces another resulting in an alternative
phrasing of a sentence. Ellsworth and Janin (2007)
produced paraphrases based on the relationship
among LUs of a frame or closely related frames,
augmenting a meager data set for use in speech
recognition.
Most importantly, LUs in a frame tend over-
whelmingly to follow the same annotation structure:
the arguments of a lexical unit in a frame consist of
the same FEs, regardless of the LU.8Note the par-
allel annotation for the two sentences in Figure 1
for the verbs, buy and purchase. We use this critical
insight in our work, hypothesizing that existing an-
notation would inform the automated generation of
annotation for unannotated LUs in a given frame.
To this end, Petruck and Swayamdipta (2019)
defined
Sister Lexical Unit
as any LU for which
annotation exists in a frame, and whose annotation
can serve as a model to generate new annotation. A
7https://github.com/swabhs/open-sesame
8
Two or more FEs might be part of a core set, yet only
one will be realized for valid annotation. For a given frame,
different LUs may not realize all the same FEs in all examples.
Sister
is the LU with annotation and an
EmptyLU
lacks annotation; the Sister LU and the Empty LU
must be of the same part of speech.
To generate augmented labeled data, we first
identified all EmptyLUs. For each such LU, we
identified a corresponding Sister, specifically, the
LU with the greatest number of annotation sets
and of the same POS as the EmptyLU in a given
frame. We replaced each occurrence of a Sister
in an annotated sentence with the EmptyLU. The
replacement process included steps to ensure that
the newly generated instance included the correct
word forms, i.e., singular and plural forms of nouns
or conjugated verbs of a Sister LU, following the
strategy that Rastogi and Van Durme (2014) em-
ployed, which augmented FN using a paraphrase
database (Ganitkevitch et al.,2013) to add new lem-
mas and then rewrote existing annotated sentences
with those lemmas.
We transferred each sentence in the annotation
set of the Sister into that of each EmptyLU, and
replaced instances of the Sister with those of the
EmptyLU. The replacement process ensured that
newly generated data included correct word forms,
i.e., singular and plural forms of nouns, conjugated
verbs, or those with tense marking.
Ensuring agreement of tense, person, and num-
ber between instances of a sister and an EmptyLU
required replacing the original word forms with
those of the same grammatical number for nouns
and tense, person, and number for verbs. However,
this does not always yield correct results, e.g., for
lemmas with irregular plural or past tense forms,
as in ox, oxen and bring, brought.
3.1 Candidates for Augmentation
The aforementioned process identified 2,805 pre-
viously unannotated LUs, of which approximately
500 either were multiword expressions (MWEs) or
had an MWE as a potential Sister. Thus, the pro-
cess added at least one sentence to each previously
empty annotation set, yielding the potential to pro-
vide new annotation for (approximately) 45% of
the unannotated LUs in FrameNet 1.7. See Table 2
in the Appendix Afor a few pairs of sister and
empty LU pairs. The replacement process did not
always yield grammatically correct results. Some-
times FrameNet (arbitrarily) treats MWEs as single
LUs, and using a MWE (as a Sister or an EmptyLU)
often resulted in ungrammatical sentences. Thus,
we also eliminated MWEs from the data in this
work, resulting in 2,300 previously unannotated
LUs, eligible for augmentation.
4 Experiments
To test the idea of leveraging frame-specific an-
notation to generate new annotation for LUs lack-
ing any, we wish to compare the performance of
our approach with a baseline model. This facili-
tates determining the extent to which our augmen-
tation algorithm would improve SRL performance,
and thus (potentially) contribute to the FrameNet
corpus. Our augmentation approach is applied to
FrameNet version 1.7.
4.1 Training Data
In principle, our data augmentation algorithm (§3)
could provide new annotation for 2,300 LUs (§3.1),
all of which could be used for training. However,
we wish to determine whether using augmented
data improves performance on the existing test sets.
Additionally, care must be taken to avoid the log-
ical flaw of using augmentation algorithm from
producing labeled sentences specifically for the
test set: the process would circularly assume that
the algorithm offers accurate labelling.
We addressed this issue by adjusting the train-
ing data for the baseline: we removed annota-
tion from 1,500 randomly selected LUs from the
lexicographic portion of the training data for FN
1.7. Each of these 1,500 LUs functions as Empty-
LUs when stripped of their annotation. Our base-
line model was trained with the remaining lexico-
graphic annotations. Our full model was trained
on data given by the augmentation algorithm to
re-annotate the 1,500 LUs, in addition to the data
used for the baseline. Note that not all LUs are
eligible for training in OPEN-SESAME, which in-
volves a pre-processing step to remove those LUs
incompatible with the parser; Table 1provides the
final training data statistics.
In addition to the lexicographic annotations, both
the baseline and the augmented setup use all the
full-text annotations in FrameNet 1.7. Our training
data adjustments above allow using the standard
validation and test sets for FrameNet 1.7, following
Swayamdipta et al. (2017).
4.2 Model and Hyperparameters
All of our experiments were performed using the
non-syntactic model provided in OPEN -SESAME
(Swayamdipta et al.,2017,2018). We follow the
basic setup in OPEN-SESAME which assumes gold
targets for FrameID and gold frames for ArgID. We
left most hyperparameters associated with OP EN-
SESAME at their default values, but decreased the
learning rate for training the ArgID stage to
0.0001
from its original value of
0.0005
. The decrease al-
lowed training ArgID on the entire dataset in three
epochs without error, which serves reproducibility
requirements.
4.3 Results
Table 1presents test and validation set perfor-
mances comparing two OP EN-SESAME models,
trained with and without the augmented data for
ArgID and FrameID. All measures in the experi-
mental setting outperformed the baseline, highlight-
ing the value of data augmentation. Although our
main interest was testing the system’s performance
for ArgID (i.e., identifying semantic roles / FEs
for given frame-LU annotations) of the augmented
data, we also observed improvement in the identifi-
cation of frames. The results in the experimental
setting show an absolute improvement of 2.7 in F1
on ArgID and 1.6 in F1 on FrameID.9
4.4 Error Analysis and Discussion
At times the augmentation algorithm produced un-
grammatical sentences. We do not believe that er-
rors occurred because of ungrammatical sentences.
Still, FrameNet would not conjure producing data
with ungrammatical sentences.
10
A manual analy-
sis of a random sample of LUs showed three cate-
gories of errors: (a) word form mismatch; (b) incor-
rect or missing marker; and (c) semantic mismatch,
examples of which appear in (a) - (c):
(a) *The moon was now occlude by clouds.
(b) *And he complained the endless squeeze on
cash.
(c) ?The faint flash from a street light showed him
the outline of a hedge.
The semantic mismatch in (c) holds between faint.a
and flash.n, yielding a semantically odd sentence
not an ungrammatical one. By definition, a flash
would not be faint.11
9
We further experimented with cross validation settings
for both FrameID and ArgID; however, the runtime of OP EN-
SESAME was prohibitively slow in this setting.
10
A complete manual analysis of the data remains to be
done. In future extensions of this work, we will report such an
analysis, including the percentage of grammatical sentences.
11
The results showed several types of semantic mismatch.
Space limitations preclude discussing all of them here.
Dev Test
Training Data ArgID FrameID ArgID FrameID
#LUs #Lexicographic Annotations F1 F1 P R F1 F1
OPEN-SES AM E 6905 164372 59.3 84.7 61.8 56.9 59.2 81.1
7944 260292 61.9 85.3 63.6 60.3 61.9 82.7
Table 1: Results comparing the O PE N-S ES AM E models trained on the original FrameNet 1.7 data for frame iden-
tification (FrameID) and argument identification (ArgID) tasks, with OPEN-SESAME models trained on our aug-
mented data. Data augmentation improves performance on both tasks across all metrics; boldface indicates best
performance. Note that under both settings we consider full-text annotations, a subset of the lexicographic annota-
tions. Gold-standard frames were used for ArgID, and gold-standard targets for FrameID, in both the baseline and
our models. Development set results only include F1 scores; (by default) OPEN-SESAME just reports that score.
Recall that incorrectly identifying a frame
necessarily yields incorrect FEs (ArgID). The
analysis showed that errors occurred with highly
polysemous lemmas, such as make.v, where
FrameNet has 10 LUs; in (d), OPEN-SES AME
misidentified making.v as
Manufacturing
,
not
Earnings_and_losses
. Similarly,
(e) is a misidentification of gathering as
Food_gathering
, not
Come_together
.
Note the metaphorical use of the verb in (e),
compared to literal senses of the lemma gather.v.
(d) I like working and making money.
(e) ...the debate on welfare reform is gathering
like a storm at sea...
5 Related Work
Data augmentation in the context of FrameNet
and SRL is not novel. Pavlick et al. (2015) cre-
ated an expanded FN via automatic paraphrasing
and crowdsourcing to confirm frame assignment
of LUs. Hartmann et al. (2017) used automatically
generated training data by linking FrameNet, Prop-
Bank, and VerbNet to study the differences among
those resources in frame assignment and seman-
tic role classification. The results showed that the
augmented data performed as well as the training
data for these tasks on German texts. With similar
goals to those of this work, Carvalho et al. (2020)
proposed augmenting FN annotation using notions
of lexical, syntactic, or semantic equivalence, and
FN’s frame-to-frame relations (Inheritance, Using,
SubFrame, etc.) to infer annotations across related
frames, producing a 13% increase in annotation.
6 Conclusion and Future Work
We outlined an approach to Frame-SRL that seeks
to leverage existing annotation in a frame to gen-
erate new annotation for previously unannotated
LUs in the same frame, by exploiting the already
identified target and frame of a given lemma. Fur-
ther, we demonstrated that augmentation of data
improves Frame-SRL performance substantially.
While we show the advantages of this approach
on the open-SESAME baseline, future work might
involve studying if using this type of data augmen-
tation will yield improvements on a state-of-the-art
SRL system, e.g. Fei et al. (2021).
Several other interesting directions for future
work exist, including the following: refining the
augmentation algorithm to produce more gram-
matical sentences; streamlining open-SESAME’s
code to run more efficiently; use active-learning
to inform the data generation and to assist in the
SRL task; and experimentation with other com-
putational techniques, as suggested by Petruck
and Swayamdipta (2019). These plans will ad-
vance defining processes to include Frame-SRL
into FrameNet’s development process, the ultimate
goal of this work.12
Acknowledgements
The authors are grateful to the three anonymous re-
viewers for their feedback, and to Nathan Schneider
for his input on various aspects of the experiment,
including very early conversations about pursuing
the idea. Collin Baker and Michael Ellsworth have
contributed to the completion of the paper in count-
less ways.
References
Collin F. Baker, Michael Ellsworth, and Katrin Erk.
2007. Semeval-2007 task 19: Frame semantic struc-
ture extraction. In Proceedings of the 4th Inter-
12
Rigorous testing of multiple factors will be required to
evaluate the viability of including Frame-SRL into FrameNet’s
existing primarily manual process.
national Workshop on Semantic Evaluations, Se-
mEval@ACL 2007, Prague, Czech Republic, June
23-24, 2007, pages 99–104. The Association for
Computer Linguistics.
Emily M. Bender and Alexander Koller. 2020. Climb-
ing towards NLU: On meaning, form, and under-
standing in the age of data. In Proceedings of the
58th Annual Meeting of the Association for Compu-
tational Linguistics, pages 5185–5198, Online. As-
sociation for Computational Linguistics.
Xavier Carreras and Lluís Màrquez. 2005. Introduc-
tion to the CoNLL-2005 shared task: Semantic
role labeling. In Proceedings of the Ninth Confer-
ence on Computational Natural Language Learning
(CoNLL-2005), pages 152–164, Ann Arbor, Michi-
gan. Association for Computational Linguistics.
Breno Carvalho, Aline Paes, and Bernardo Gonçalves.
2020. Augmenting linguistic semi-structured data
for machine learning: A Case study using FrameNet.
In Proceedings of the International Conference on
Machine Learning Techniques and NLP, volume
10.12, pages 1–13.
Dipanjan Das, Desai Chen, André F. T. Martins,
Nathan Schneider, and Noah A. Smith. 2014.
Frame-semantic parsing.Computational Linguis-
tics, 40(1):9–56.
Dipanjan Das, Nathan Schneider, Desai Chen, and
Noah A. Smith. 2010. Probabilistic frame-semantic
parsing. In Human Language Technologies: The
2010 Annual Conference of the North American
Chapter of the Association for Computational Lin-
guistics, pages 948–956, Los Angeles, California.
Association for Computational Linguistics.
Michael Ellsworth and Adam Janin. 2007. Mutaphrase:
Paraphrasing with FrameNet. In Proceedings of the
ACL-PASCAL Workshop on Textual Entailment and
Paraphrasing, pages 143–150, Prague. Association
for Computational Linguistics.
Hao Fei, Shengqiong Wu, Yafeng Ren, and Donghong
Ji. 2021. Second-order semantic role labeling with
global structural refinement.IEEE/ACM Transac-
tions on Audio, Speech, and Language Processing,
29:1966–1976.
Charles J. Fillmore. 1985. Frames and the semantics
of understanding.Quaderni di Semantica, 6(2):222–
254.
Nicholas FitzGerald, Julian Michael, Luheng He, and
Luke Zettlemoyer. 2018. Large-scale QA-SRL pars-
ing. In Proceedings of the 56th Annual Meeting of
the Association for Computational Linguistics (Vol-
ume 1: Long Papers), pages 2051–2060, Melbourne,
Australia. Association for Computational Linguis-
tics.
Juri Ganitkevitch, Benjamin Van Durme, and Chris
Callison-Burch. 2013. PPDB: The paraphrase
database. In Proceedings of the 2013 Conference of
the North American Chapter of the Association for
Computational Linguistics: Human Language Tech-
nologies, pages 758–764, Atlanta, Georgia. Associa-
tion for Computational Linguistics.
Daniel Gildea and Daniel Jurafsky. 2002. Automatic
labeling of semantic roles.Computational Linguis-
tics, 28(3):245–288.
Xu Han, Tao Lv, Zhirui Hu, Xinyan Wang, and Cong
Wang. 2016. Text summarization using FrameNet-
based semantic graph model.Scientific Program-
ming, 2016:1–10.
Silvana Hartmann, Éva Mújdricza-Maydt, Ilia
Kuznetsov, Iryna Gurevych, and Anette Frank.
2017. Assessing SRL frameworks with automatic
training data expansion. In Proceedings of the 11th
Linguistic Annotation Workshop, pages 115–121,
Valencia, Spain. Association for Computational
Linguistics.
Christopher R. Johnson, Charles J. Fillmore, Esther J.
Wood, Josef Ruppenhofer, Margaret Urban, Miriam
R. L. Petruck, and Collin F. Baker. 2001. The
FrameNet Project: Tools for Lexicon Building.
ICSI Online Publication.
Karin Kipper, Hoa Trang Dang, and Martha Palmer.
2000. Class-based construction of a verb lexicon.
In Proceedings of the Seventeenth National Confer-
ence on Artificial Intelligence and Twelfth Confer-
ence on Innovative Applications of Artificial Intelli-
gence, pages 691–696. AAAI Press.
Martha Palmer, Daniel Gildea, and Paul Kingsbury.
2005. The proposition bank: An annotated cor-
pus of semantic roles.Computational Linguistics,
31(1):71–106.
Ellie Pavlick, Travis Wolfe, Pushpendre Rastogi,
Chris Callison-Burch, Mark Dredze, and Benjamin
Van Durme. 2015. FrameNet+: Fast paraphrastic
tripling of FrameNet. In Proceedings of the 53rd An-
nual Meeting of the Association for Computational
Linguistics and the 7th International Joint Confer-
ence on Natural Language Processing (Volume 2:
Short Papers), pages 408–413, Beijing, China. As-
sociation for Computational Linguistics.
Bolette S. Pedersen. 2001. Lexical ambiguity in ma-
chine translation: Using frame semantics for ex-
pressing systemacies in polysemy. In Recent Ad-
vances in Natural Language Processing.
Miriam R L Petruck and Swabha Swayamdipta. 2019.
Automating FrameNet annotation. Pre-Proposal for
NSF CISE EAGER Grant.
Sameer Pradhan, Alessandro Moschitti, Nianwen Xue,
Hwee Tou Ng, Anders Björkelund, Olga Uryupina,
Yuchen Zhang, and Zhi Zhong. 2013. Towards ro-
bust linguistic analysis using OntoNotes. In Pro-
ceedings of the Seventeenth Conference on Computa-
tional Natural Language Learning, pages 143–152,
Sofia, Bulgaria. Association for Computational Lin-
guistics.
Sameer Pradhan, Alessandro Moschitti, Nianwen Xue,
Olga Uryupina, and Yuchen Zhang. 2012. CoNLL-
2012 shared task: Modeling multilingual unre-
stricted coreference in OntoNotes. In Joint Confer-
ence on EMNLP and CoNLL - Shared Task, pages
1–40, Jeju Island, Korea. Association for Computa-
tional Linguistics.
Pushpendre Rastogi and Benjamin Van Durme. 2014.
Augmenting FrameNet via PPDB. In Proceedings
of the Second Workshop on EVENTS: Definition, De-
tection, Coreference, and Representation, Baltimore,
Maryland, USA. Association for Computational Lin-
guistics.
Michael Roth and Mirella Lapata. 2016. Neural seman-
tic role labeling with dependency path embeddings.
In Proceedings of the 54th Annual Meeting of the
Association for Computational Linguistics (Volume
1: Long Papers), pages 1192–1202.
Josef Ruppenhofer, Michael Ellsworth, Miriam R. L.
Petruck, Christopher R. Johnson, Collin F. Baker,
and Jan Scheffczyk. 2016. FrameNet II: Extended
Theory and Practice. ICSI: Berkeley.
Sunita Sarawagi and William W. Cohen. 2004. Semi-
markov conditional random fields for information
extraction. In Proc. of NeurIPS, pages 1185–1192.
Dan Shen and Mirella Lapata. 2007. Using seman-
tic roles to improve question answering. In Pro-
ceedings of the 2007 Joint Conference on Empirical
Methods in Natural Language Processing and Com-
putational Natural Language Learning (EMNLP-
CoNLL), pages 12–21, Prague, Czech Republic. As-
sociation for Computational Linguistics.
Noah A. Smith. 2011. Linguistic Structure Prediction.
Synthesis Lectures on Human Language Technolo-
gies. Morgan and Claypool.
Swabha Swayamdipta, Sam Thomson, Chris Dyer, and
Noah A. Smith. 2017. Frame-semantic parsing with
softmax-margin segmental rnns and a syntactic scaf-
fold.
Swabha Swayamdipta, Sam Thomson, Kenton Lee,
Luke Zettlemoyer, Chris Dyer, and Noah A. Smith.
2018. Syntactic scaffolds for semantic structures.
In Proceedings of the 2018 Conference on Em-
pirical Methods in Natural Language Processing,
pages 3772–3782, Brussels, Belgium. Association
for Computational Linguistics.
Bishan Yang and Tom Mitchell. 2017. A joint sequen-
tial and relational model for frame-semantic parsing.
In Proceedings of the 2017 Conference on Empiri-
cal Methods in Natural Language Processing, pages
1247–1256, Copenhagen, Denmark. Association for
Computational Linguistics.
A Qualitative Examples
Table 2provides lexicographic annotations with
sister LUs and their corresponding empty LUs, the
latter being populated by our approach.
Sister Lexical Units Empty Lexical Units
He
AGEN T
stamped his foot
BODY_PA RT
into his flying-
boot PERIPHERAL .
→
He
AGEN T
bended his foot
BODY_PA RT
into his flying-
boot PERIPHERAL .
John-William , who knew
that he would have been
a Chartist himself had
he remained a poor man
PERS ON
, felt sorry about
that death .
→
John-William , who knew
that he would have been
a Chartist himself had
he remained a rich man
PERS ON
, felt sorry about
that death .
Because of the censor-
ship , and the obvious
need to avoid danger-
ously
DEGR EE
critical com-
ments
EXPR ESS OR
about
the regime and the war
EVALUE E
, the correspon-
dence to and from the
front provides no easy
guide to political atti-
tudes .
→
Because of the censor-
ship , and the obvious
need to avoid danger-
ously
DEGR EE
commend-
able comments
EXPR ESS OR
about the regime and the
war
EVALUE E
, the corre-
spondence to and from
the front provides no easy
guide to political atti-
tudes .
When I did eventually tell
her she
EXPERIENCER
was re-
ally
DEGREE
embarrassed ,
and tried telling me that I
was making it up !
→
When I did eventually tell
her she
EXPERIENCER
was
really
DEGR EE
tormented ,
and tried telling me that I
was making it up !
This regulation prevented
US banks
THEM E
located
in the US
LOCATI ON
, but
not abroad, from paying
interest on deposits above
a given rate .
→
This regulation prevented
US banks
THEM E
situated
in the US
LOCATION
, but
not abroad, from paying
interest on deposits above
a given rate .
Table 2: Parallel lexicographic annotations of sister
LUs (expert-annotated) and emtpy LUs (augmented
by our approach). Frame Element (FE) names are
in SMALL CAPS and their instantiations are color-
coordinated; and LUs or targets are in italics.
