We describe our systems for SemEval-2024 Task 1: Semantic Textual Relatedness. We investigate the correlation between semantic relatedness and semantic similarity. Specifically, we test two hypotheses: (1) similarity is a special case of relatedness, and (2) semantic relatedness is preserved under translation. We experiment with a variety of approaches which are based on explicit semantics, downstream applications, contextual embeddings, large language models (LLMs), as well as ensembles of methods. We find empirical support for our theoretical insights. In addition, our best ensemble system yields highly competitive results in a number of diverse categories. Our code and data are available on GitHub.
Language-vision models like CLIP have made significant strides in vision tasks, such as zero-shot image classification (ZSIC). However, generating specific and expressive visual descriptions remains challenging; descriptions produced by current methods are often ambiguous and lacking in granularity. To tackle these issues, we propose V-GLOSS: Visual Glosses, a novel method built upon two key ideas. The first is Semantic Prompting, which conditions a language model on structured semantic knowledge. The second is a new contrastive algorithm that elicits fine-grained distinctions between similar concepts. With both ideas, we demonstrate that V-GLOSS improves visual descriptions and achieves strong results in the zero-shot setting on general and fine-grained image-classification datasets, including ImageNet, STL-10, FGVC Aircraft, and Flowers 102. Moreover, these descriptive capabilities contribute to enhancing image-generation performance. Finally, we introduce a quality-tested silver dataset with descriptions generated with V-GLOSS for all ImageNet classes.
Constructing lexicons with explicitly identified lexical gaps is a vital part of building multilingual lexical resources. Prior work has leveraged bilingual dictionaries and linguistic typologies for semi-automatic identification of lexical gaps. Instead, we propose a generally-applicable algorithmic method to automatically generate concept lexicalizations, which is based on machine translation and hypernymy relations between concepts. The absence of a lexicalization implies a lexical gap. We apply our method to kinship terms, which make a suitable case study because of their explicit definitions and regular structure. Empirical evaluations demonstrate that our approach yields higher accuracy than BabelNet and ChatGPT. Our error analysis indicates that enhancing the quality of translations can further improve the accuracy of our method.
Causal language models such as the GPT series have achieved significant success across various domains. However, their application to the lexical substitution task (LST) remains largely unexplored due to inherent limitations in autoregressive decoding. Our work is motivated by our observation that existing LST approaches tend to suffer from a misalignment between the pre-training objectives of the language models that they employ, and their subsequent fine-tuning and application for substitute generation. We introduce PromptSub, the first system to use causal language modeling (CLM) for LST. Through prompt-aware fine-tuning, PromptSub not only enriches the given context with additional knowledge, but also leverages the unidirectional nature of autoregressive decoding. PromptSub consistently outperforms GeneSis, the best previously published supervised LST method. Further analysis demonstrates the potential of PromptSub to further benefit from increased model capacity, expanded data resources, and retrieval of external knowledge. By framing LST within the paradigm of CLM, our approach indicates the versatility of general CLM-based systems, such as ChatGPT, in catering to specialized tasks, including LST.
Paraphrase identification (PI) and natural language inference (NLI) are two important tasks in natural language processing. Despite their distinct objectives, an underlying connection exists, which has been notably under-explored in empirical investigations. We formalize the relationship between these semantic tasks and introduce a method for solving PI using an NLI system, including the adaptation of PI datasets for fine-tuning NLI models. Through extensive evaluations on six PI benchmarks, across both zero-shot and fine-tuned settings, we showcase the efficacy of NLI models for PI through our proposed reduction. Remarkably, our fine-tuning procedure enables NLI models to outperform dedicated PI models on PI datasets. In addition, our findings provide insights into the limitations of current PI benchmarks.
Emotion identification and polarity classification seek to determine the sentiment expressed by a writer. Sentiment lexicons that provide classifications at the word level fail to distinguish between different senses of polysemous words. To address this problem, we propose a translation-based method for labeling each individual lexical concept and word sense. Specifically, we translate synsets into 20 different languages and verify the sentiment of these translations in multilingual sentiment lexicons. By applying our method to all WordNet synsets, we produce SentiSynset, a synset-level sentiment resource containing 12,429 emotional synsets and 15,567 polar synsets, which is significantly larger than previous resources. Experimental evaluation shows that our method outperforms prior automated methods that classify word senses, in addition to outperforming ChatGPT. We make the resulting resource publicly available on GitHub.
Acquiring large amounts of high-quality annotated data is an open issue in word sense disambiguation. This problem has become more critical recently with the advent of supervised models based on neural networks, which require large amounts of annotated data. We propose two algorithms for making selective corrections on a sense-annotated parallel corpus, based on cross-lingual synset mappings. We show that, when applied to bilingual parallel corpora, these algorithms can rectify noisy sense annotations, and thereby produce multilingual sense-annotated data of improved quality.
Existing definitions of lexical substitutes are often vague or inconsistent with the gold annotations. We propose a new definition which is grounded in the relation of entailment; namely, that the sentence that results from the substitution should be in the relation of mutual entailment with the original sentence. We argue that the new definition is well-founded and supported by previous work on lexical entailment. We empirically validate our definition by verifying that it covers the majority of gold substitutes in existing datasets. Based on this definition, we create a new dataset from existing semantic resources. Finally, we propose a novel context augmentation method motivated by the definition, which relates the substitutes to the sense of the target word by incorporating glosses and synonyms directly into the context. Experimental results demonstrate that our augmentation approach improves the performance of lexical substitution systems on the existing benchmarks.
Semantic tasks are rarely formally defined, and the exact relationship between them is an open question. We introduce a taxonomy that elucidates the connection between several problems in lexical semantics, including monolingual and cross-lingual variants. Our theoretical framework is based on the hypothesis of the equivalence of concept and meaning distinctions. Using algorithmic problem reductions, we demonstrate that all problems in the taxonomy can be reduced to word sense disambiguation (WSD), and that WSD itself can be reduced to some problems, making them theoretically equivalent. In addition, we carry out experiments that strongly support the soundness of the concept-meaning hypothesis, and the correctness of our reductions.
As the minimum semantic units of natural languages, sememes can provide precise representations of concepts. Despite the widespread utilization of lexical resources for semantic tasks, use of sememes is limited by a lack of available sememe knowledge bases. Recent efforts have been made to connect BabelNet with HowNet by automating sememe prediction. However, these methods depend on large manually annotated datasets. We propose to use sense alignment via a novel unsupervised and explainable method. Our method consists of four stages, each relaxing predefined constraints until a complete alignment of BabelNet synsets to HowNet senses is achieved. Experimental results demonstrate the superiority of our unsupervised method over previous supervised ones by an improvement of 12% overall F1 score, setting a new state of the art. Our work is grounded in an interpretable propagation of sememe information between lexical resources, and may benefit downstream applications which can incorporate sememe information.
We describe the systems of the University of Alberta team for the SemEval-2023 Visual Word Sense Disambiguation (V-WSD) Task. We present a novel algorithm that leverages glosses retrieved from BabelNet, in combination with text and image encoders. Furthermore, we compare language-specific encoders against the application of English encoders to translated texts. As the contexts given in the task datasets are extremely short, we also experiment with augmenting these contexts with descriptions generated by a language model. This yields substantial improvements in accuracy. We describe and evaluate additional V-WSD methods which use image generation and text-conditioned image segmentation. Some of our experimental results exceed those of our official submissions on the test set. Our code is publicly available at https://github.com/UAlberta-NLP/v-wsd.
Large language models (LLMs) have demonstrated exceptional natural language understanding abilities, and have excelled in a variety of natural language processing (NLP) tasks. Despite the fact that most LLMs are trained predominantly on English, multiple studies have demonstrated their capabilities in a variety of languages. However, fundamental questions persist regarding how LLMs acquire their multilingual abilities and how performance varies across different languages. These inquiries are crucial for the study of LLMs since users and researchers often come from diverse language backgrounds, potentially influencing how they use LLMs and interpret their output. In this work, we propose a systematic way of qualitatively and quantitatively evaluating the multilingual capabilities of LLMs. We investigate the phenomenon of cross-language generalization in LLMs, wherein limited multilingual training data leads to advanced multilingual capabilities. To accomplish this, we employ a novel prompt back-translation method. The results demonstrate that LLMs, such as GPT, can effectively transfer learned knowledge across different languages, yielding relatively consistent results in translation-equivariant tasks, in which the correct output does not depend on the language of the input. However, LLMs struggle to provide accurate results in translation-variant tasks, which lack this property, requiring careful user judgment to evaluate the answers.
Aligning lexical resources that associate words with concepts in multiple languages increases the total amount of semantic information that can be leveraged for various NLP tasks. We present a translation-based approach to mapping concepts across diverse resources. Our methods depend only on multilingual lexicalization information. When applied to align WordNet/BabelNet to CLICS and OmegaWiki, our methods achieve state-of-the-art accuracy, without any dependence on other sources of semantic knowledge. Since each word-concept pair corresponds to a unique sense of the word, we also demonstrate that the mapping task can be framed as word sense disambiguation. To facilitate future work, we release a set of high-precision WordNet-CLICS alignments, produced by combining three different mapping methods.
We describe the University of Alberta systems for the SemEval-2022 Task 2 on multilingual idiomaticity detection. Working under the assumption that idiomatic expressions are noncompositional, our first method integrates information on the meanings of the individual words of an expression into a binary classifier. Further hypothesizing that literal and idiomatic expressions translate differently, our second method translates an expression in context, and uses a lexical knowledge base to determine if the translation is literal. Our approaches are grounded in linguistic phenomena, and leverage existing sources of lexical knowledge. Our results offer support for both approaches, particularly the former.
Word sense disambiguation (WSD) is the task of identifying the intended sense of a word in context. While prior work on unsupervised WSD has leveraged lexical knowledge bases, such as WordNet and BabelNet, these resources have proven to be less effective for Chinese. Instead, the most widely used lexical knowledge base for Chinese is HowNet. Previous HowNet-based WSD methods have not exploited contextual translation information. In this paper, we present the first HowNet-based WSD system which combines monolingual contextual information from a pretrained neural language model with bilingual information obtained via machine translation and sense translation information from HowNet. The results of our evaluation experiment on a test set from prior work demonstrate that our new method achieves a new state of the art for unsupervised Chinese WSD.
The Word-in-Context (WiC) task has attracted considerable attention in the NLP community, as demonstrated by the popularity of the recent MCL-WiC SemEval shared task. Systems and lexical resources from word sense disambiguation (WSD) are often used for the WiC task and WiC dataset construction. In this paper, we establish the exact relationship between WiC and WSD, as well as the related task of target sense verification (TSV). Building upon a novel hypothesis on the equivalence of sense and meaning distinctions, we demonstrate through the application of tools from theoretical computer science that these three semantic classification problems can be pairwise reduced to each other, and therefore are equivalent. The results of experiments that involve systems and datasets for both WiC and WSD provide strong empirical evidence that our problem reductions work in practice.
We describe our two systems for the shared task on Lexical Semantic Change Discovery in Spanish. For binary change detection, we frame the task as a word sense disambiguation (WSD) problem. We derive sense frequency distributions for target words in both old and modern corpora. We assume that the word semantics have changed if a sense is observed in only one of the two corpora, or the relative change for any sense exceeds a tuned threshold. For graded change discovery, we follow the design of CIRCE (Pömsl and Lyapin, 2020) by combining both static and contextual embeddings. For contextual embeddings, we use XLM-RoBERTa instead of BERT, and train the model to predict a masked token instead of the time period. Our language-independent methods achieve results that are close to the best-performing systems in the shared task.
Colexification occurs when two distinct concepts are lexified by the same word. The term covers both polysemy and homonymy. We posit and investigate the hypothesis that no pair of concepts are colexified in every language. We test our hypothesis by analyzing colexification data from BabelNet, Open Multilingual WordNet, and CLICS. The results show that our hypothesis is supported by over 99.9% of colexified concept pairs in these three lexical resources.
Deciding whether a semantically ambiguous word is homonymous or polysemous is equivalent to establishing whether it has any pair of senses that are semantically unrelated. We present novel methods for this task that leverage information from multilingual lexical resources. We formally prove the theoretical properties that provide the foundation for our methods. In particular, we show how the One Homonym Per Translation hypothesis of Hauer and Kondrak (2020a) follows from the synset properties formulated by Hauer and Kondrak (2020b). Experimental evaluation shows that our approach sets a new state of the art for homonymy detection.
We describe the University of Alberta systems for the SemEval-2021 Word-in-Context (WiC) disambiguation task. We explore the use of translation information for deciding whether two different tokens of the same word correspond to the same sense of the word. Our focus is on developing principled theoretical approaches which are grounded in linguistic phenomena, leading to more explainable models. We show that translations from multiple languages can be leveraged to improve the accuracy on the WiC task.
Acquisition of multilingual training data continues to be a challenge in word sense disambiguation (WSD). To address this problem, unsupervised approaches have been proposed to automatically generate sense annotations for training supervised WSD systems. We present three new methods for creating sense-annotated corpora which leverage translations, parallel bitexts, lexical resources, as well as contextual and synset embeddings. Our semi-supervised method applies machine translation to transfer existing sense annotations to other languages. Our two unsupervised methods refine sense annotations produced by a knowledge-based WSD system via lexical translations in a parallel corpus. We obtain state-of-the-art results on standard WSD benchmarks.
The Dorabella cipher is an encrypted note of English composer Edward Elgar, which has defied decipherment attempts for more than a century. While most proposed solutions are English texts, we investigate the hypothe- sis that Dorabella represents enciphered music. We weigh the evidence in favor of and against the hypothesis, devise a simplified music nota- tion, and attempt to reconstruct a melody from the cipher. Our tools are n-gram models of mu- sic which we validate on existing music cor- pora enciphered using monoalphabetic substi- tution. By applying our methods to Dorabella, we produce a decipherment with musical qual- ities, which is then transformed via artful com- position into a listenable melody. Far from ar- guing that the end result represents the only true solution, we instead frame the process of decipherment as part of the composition pro- cess.
This paper presents the University of Alberta systems and results in the SIGMORPHON 2020 Task 1: Multilingual Grapheme-to-Phoneme Conversion. Following previous SIGMORPHON shared tasks, we define a low-resource setting with 100 training instances. We experiment with three transduction approaches in both standard and low-resource settings, as well as on the related task of phoneme-to-grapheme conversion. We propose a method for synthesizing training data using a combination of diverse models.
It has been conjectured that multilingual information can help monolingual word sense disambiguation (WSD). However, existing WSD systems rarely consider multilingual information, and no effective method has been proposed for improving WSD by generating translations. In this paper, we present a novel approach that improves the performance of a base WSD system using machine translation. Since our approach is language independent, we perform WSD experiments on several languages. The results demonstrate that our methods can consistently improve the performance of WSD systems, and obtain state-ofthe-art results in both English and multilingual WSD. To facilitate the use of lexical translation information, we also propose BABALIGN, an precise bitext alignment algorithm which is guided by multilingual lexical correspondences from BabelNet.
We investigate the hypothesis that translations can be used to identify cross-lingual lexical entailment. We propose novel methods that leverage parallel corpora, word embeddings, and multilingual lexical resources. Our results demonstrate that the implementation of these ideas leads to improvements in predicting entailment.
The conversion of romanized texts back to the native scripts is a challenging task because of the inconsistent romanization conventions and non-standard language use. This problem is compounded by code-mixing, i.e., using words from more than one language within the same discourse. In this paper, we propose a novel approach for handling these two problems together in a single system. Our approach combines three components: language identification, back-transliteration, and sequence prediction. The results of our experiments on Bengali and Hindi datasets establish the state of the art for the task of deromanization of code-mixed texts.
We propose cognate projection as a method of crosslingual transfer for inflection generation in the context of the SIGMORPHON 2019 Shared Task. The results on four language pairs show the method is effective when no low-resource training data is available.
We report the results of our experiments in the context of the NEWS 2018 Shared Task on Transliteration. We focus on the comparison of several diverse systems, including three neural MT models. A combination of discriminative, generative, and neural models obtains the best results on the development sets. We also put forward ideas for improving the shared task.
Many character-level tasks can be framed as sequence-to-sequence transduction, where the target is a word from a natural language. We show that leveraging target language models derived from unannotated target corpora, combined with a precise alignment of the training data, yields state-of-the art results on cognate projection, inflection generation, and phoneme-to-grapheme conversion.
We present a system for identifying cognate sets across dictionaries of related languages. The likelihood of a cognate relationship is calculated on the basis of a rich set of features that capture both phonetic and semantic similarity, as well as the presence of regular sound correspondences. The similarity scores are used to cluster words from different languages that may originate from a common proto-word. When tested on the Algonquian language family, our system detects 63% of cognate sets while maintaining cluster purity of 70%.
The task of morphological analysis is to produce a complete list of lemma+tag analyses for a given word-form. We propose a discriminative string transduction approach which exploits plain inflection tables and raw text corpora, thus obviating the need for expert annotation. Experiments on four languages demonstrate that our system has much higher coverage than a hand-engineered FST analyzer, and is more accurate than a state-of-the-art morphological tagger.
The task of unsupervised lexicon induction is to find translation pairs across monolingual corpora. We develop a novel method that creates seed lexicons by identifying cognates in the vocabularies of related languages on the basis of their frequency and lexical similarity. We apply bidirectional bootstrapping to a method which learns a linear mapping between context-based vector spaces. Experimental results on three language pairs show consistent improvement over prior work.
Algorithmic decipherment is a prime example of a truly unsupervised problem. The first step in the decipherment process is the identification of the encrypted language. We propose three methods for determining the source language of a document enciphered with a monoalphabetic substitution cipher. The best method achieves 97% accuracy on 380 languages. We then present an approach to decoding anagrammed substitution ciphers, in which the letters within words have been arbitrarily transposed. It obtains the average decryption word accuracy of 93% on a set of 50 ciphertexts in 5 languages. Finally, we report the results on the Voynich manuscript, an unsolved fifteenth century cipher, which suggest Hebrew as the language of the document.
We evaluate several orthographic word similarity measures in the context of bitext word alignment. We investigate the relationship between the length of the words and the length of their longest common subsequence. We present an alternative to the longest common subsequence ratio (LCSR), a widely-used orthographic word similarity measure. Experiments involving identification of cognates in bitexts suggest that the alternative method outperforms LCSR. Our results also indicate that alignment links can be used as a substitute for cognates for the purpose of evaluating word similarity measures.