Papers by Senyu Li
Warmup Generations: A Task-Agnostic Approach for Guiding Sequence-to-Sequence Learning with Unsupervised Initial State Generation (2025.acl-long)
Copied to clipboard
| Challenge: | Existing supervised fine-tuning (SFT) methods focus on directly generating the target output without leveraging the benefits of intermediate steps or initial guidance. |
| Approach: | They propose a task-agnostic framework that enables models to generate intermediate "warmup" sequences that are iteratively refined to maximize their contribution to the final output. |
| Outcome: | The proposed framework outperforms traditional supervised fine-tuning methods on translation, summarization, and multi-choice question answering tasks. |
Don’t Trust ChatGPT when your Question is not in English: A Study of Multilingual Abilities and Types of LLMs (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies have shown that large language models can perform a wide variety of language tasks when presented in English. |
| Approach: | They propose a method to evaluate the multilingual capabilities of large language models using a prompt back-translation method to find out how LLMs acquire their multilingual abilities. |
| Outcome: | The proposed method shows that large language models can transfer learned knowledge across different languages, but struggle to provide accurate results in translation-variant tasks. |
SSA-COMET: Do LLMs Outperform Learned Metrics in Evaluating MT for Under-Resourced African Languages? (2025.emnlp-main)
Copied to clipboard
Senyu Li, Jiayi Wang, Felermino D. M. A. Ali, Colin Cherry, Daniel Deutsch, Eleftheria Briakou, Rui Sousa-Silva, Henrique Lopes Cardoso, Pontus Stenetorp, David Ifeoluwa Adelani
| Challenge: | Existing metrics for machine translation quality for under-resourced African languages suffer from limited language coverage and poor performance in low-resource settings. |
| Approach: | They propose a large-scale human-annotated machine translation evaluation dataset . they use a reference-based and reference-free evaluation model to compare MT quality . |
| Outcome: | The proposed models outperform AfriCOMET and the strongest LLM on low-resource languages. |
Identifying Emotional and Polar Concepts via Synset Translation (2024.starsem-1)
Copied to clipboard
| Challenge: | Emotion identification and polarity classification seek to determine sentiment expressed by a writer. |
| Approach: | They propose a translation-based method for labeling each individual word sense and lexical concept into 20 different languages and translate them into multilingual sentiment lexicons. |
| Outcome: | The proposed method outperforms existing methods and is available on GitHub . it contains 12,429 emotional synsets and 15,567 polar synset. |
Translation-based Lexicalization Generation and Lexical Gap Detection: Application to Kinship Terms (2024.acl-long)
Copied to clipboard
| Challenge: | Existing methods for identifying lexical gaps have been limited . kinship terms are well-suited for investigations into lexicons and lexicals . |
| Approach: | They propose an algorithm to automatically generate concept lexicalizations based on machine translation and hypernymy relations between concepts. |
| Outcome: | Empirical evaluations show that the proposed method is more accurate than BabelNet and ChatGPT. |