Papers by Francis Bond
ChainNet: Structured Metaphor and Metonymy in WordNet (2024.lrec-main)
Copied to clipboard
| Challenge: | In a typical lexicon, word senses are encoded as a list, without inter-sense relations. |
| Approach: | They propose a lexical resource which explicitly identifies the senses of a word's senses by expressing how they are derived from one another. |
| Outcome: | The proposed resource expresses how senses in the Open English Wordnet are derived from one another. |
Commonsense inference in human-robot communication (D19-60)
Copied to clipboard
| Challenge: | a gap exists in natural language understanding of commands between humans and machines. |
| Approach: | They propose a method for commonsense inference to transform high-level commands into action commands for robotic systems to execute. |
| Outcome: | The proposed method allows to build a knowledge base that consists of a large set of commonsense inferences. |
Singlish Where Got Rules One? Constructing a Computational Grammar for Singlish (2022.lrec-1)
Copied to clipboard
| Challenge: | Singlish is a variety of English spoken in Singapore and has many non-standard features. |
| Approach: | They propose to use Singlish as a branch of English grammar to implement new rules and add new lexical types to it. |
| Outcome: | The proposed grammar is based on the existing rules and lexical types from the English resource grammar and compared with the standard English grammar. |
Toward An Epic Epigraph Graph (L18-1)
Copied to clipboard
| Challenge: | a database of epigraphs is being developed to reveal literary influence as a set of connections between authors over time. |
| Approach: | a database of epigraphs is created to map literary influence as a set of connections between authors . the database is being developed under an open license . |
| Outcome: | a database of epigraphs is being developed to reveal literary influence over time . the database includes epigraph quotations from over 12,000 literary works . authors use epigraph to set theme and link work to existing body of literature . |
More Aligned, Less Diverse? Analyzing the Grammar and Lexicon of Two Generations of LLMs (2026.acl-long)
Copied to clipboard
| Challenge: | a growing number of studies compare LLMs with human-authored text . diversity is unclear, but it is important to understand what makes human and machine writing distinct . |
| Approach: | They compare syntactic properties of AI-generated and human-authored English news texts . they use the Head-Driven Phrase Structure Grammar and the English Resource Grammar . |
| Outcome: | The proposed model differs from human-authored English news text in two years. |
Comparing LLM-generated and human-authored news text using formal syntactic theory (2025.acl-long)
Copied to clipboard
| Challenge: | a systematic comparison of LLM-generated and human-authored texts is a topic of growing interest in the field of natural language processing. |
| Approach: | They compare LLM-generated and human-authored New York Times texts using a formal syntactic theory . they use a broad-coverage English resource grammar to analyze the texts . |
| Outcome: | The proposed comparisons reveal systematic differences between human and LLM-generated texts . the authors hope the results will lead to further discoveries about grammatical properties of LLMs . |
Linking the TUFS Basic Vocabulary to the Open Multilingual Wordnet (2020.lrec-1)
Copied to clipboard
| Challenge: | The TUFS Basic Vocabulary Modules are hand created, using commonly occurring vocabulary. |
| Approach: | They propose to link the TUFS Basic Vocabulary Modules with the Open Multilingual Wordnet to create a multilingual lexicon. |
| Outcome: | The proposed lexicons can be used to evaluate existing wordnets, add data to wordnet synsets and create new open wordnet for Khmer, Korean, Lao, Mongolian, Russian, Tagalog, Urdua nd Vietnamese. |
Some Issues with Building a Multilingual Wordnet (2020.lrec-1)
Copied to clipboard
| Challenge: | Notable extensions include: confidence, corpus frequency, orthographic variants, lexicalized and non-lexicalised synsets and lemmas, new parts of speech, and more. |
| Approach: | They propose to integrate a new open multilingual wordnet format that tests the extensions introduced by the new format and integrates a set of tools to ensure the integrity of the Collaborative Interlingual Index. |
| Outcome: | The proposed format integrates a set of tools that test the extensions while ensuring the integrity of the Collaborative Interlingual Index (CILI). |
The Tembusu Treebank: An English Learner Treebank (2022.lrec-1)
Copied to clipboard
| Challenge: | a new treebank is created to help diagnose ungrammatical sentences using mal-rules . the Tembusu Learner Treebank is an open treebank created from the corpus of Learner English . |
| Approach: | They propose to use the Tembusu Learner Treebank to train a new parse-ranking model for the English Resource Grammar . the model incorporates mal-rules in the annotation of ungrammatical sentences . |
| Outcome: | The Tembusu Learner Treebank is an open treebank created from the NTU Corpus of Learner English . the treebank is unique for incorporating mal-rules in the annotation of ungrammatical sentences . |
Automated Writing Support Using Deep Linguistic Parsers (2020.lrec-1)
Copied to clipboard
Luís Morgado da Costa, Roger V P Winder, Shu Yun Li, Benedict Christopher Lin Tzer Liang, Joseph Mackinnon, Francis Bond
| Challenge: | Automated Grammar Error Detection (GED) and Grammar Erreor Correction (GEC) are tasks that have attracted some attention within the NLP community. |
| Approach: | They propose a web-based system that integrates English Grammatical Error Detection (GED) and course-specific stylistic guidelines to automatically review and provide feedback on student assignments. |
| Outcome: | The system integrates both general NLP methods and high precision parsers to check student assignments before they are submitted for grading. |
Sense and Sentiment (2022.lrec-1)
Copied to clipboard
| Challenge: | Existing sentiment lexicons and concept-based sentiment-tagged corpora are not accurate, and it is difficult to map sentiment scores accurately to different languages. |
| Approach: | They examine existing sentiment lexicons and sense-based sentiment-tagged corpora to find out how sense and concept-based semantic relations effect sentiment scores. |
| Outcome: | The proposed lexicon can be used to generate sentiment lexicos for English using the Open Multilingual Wordnet. |