Papers by YunSeok Choi
CodePrompt: Task-Agnostic Prefix Tuning for Program and Language Generation (2023.findings-acl)
Copied to clipboard
| Challenge: | Prompt-tuning methods have been used to solve inefficient parameter update and storage issues in Natural Language Generation tasks. |
| Approach: | They propose a task-agnostic prompt tuning method that reflects the traits of PLM for program language. |
| Outcome: | The proposed method is effective in three PLG tasks, not only in the full-data setting but also in the low-resource setting and cross-domain setting. |
ReFEree: Reference-Free and Fine-Grained Method for Evaluating Factual Consistency in Real-World Code Summarization (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for evaluating factual consistency are primarily designed for short summaries of isolated code snippets. |
| Approach: | They propose a reference-free and fine-grained method for evaluating factual consistency in real-world code summaries. |
| Outcome: | The proposed method achieves highest correlation with human judgment among 13 baselines, improving 15-18% over the previous state-of-the-art. |
EXPO-SQL: Execution-based Clause-level Policy Optimization for Text-to-SQL (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing RL methods assign query-level rewards to all clauses, treating correct and incorrect clauses equally. |
| Approach: | They propose a method which provides fine-grained supervision through clause-level rewards. |
| Outcome: | Experiments on widely-used Text-to-SQL benchmarks show that EXPO-SqL outperforms existing methods by fine-grained clause-level learning. |
DeCAP: Context-Adaptive Prompt Generation for Debiasing Zero-shot Question Answering in Large Language Models (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing zero-shot methods for Question Answering (QA) are efficient but fail to consider context and prevent bias propagation in the answers. |
| Approach: | They propose a method for debiasing Large Language Models using context-adaptive prompt generation that takes appropriate debiased actions based on the context and aNeutral Answer Guidance Generation to suppress the LLMs make objective judgments about the context. |
| Outcome: | The proposed method achieves state-of-the-art zero-shot debiased QA performance across eight LLMs. |
TABS: Efficient Textual Adversarial Attack for Pre-trained NL Code Model Using Semantic Beam Search (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing black-box adversarial attacks on pre-trained models generate adversarials with greedy search. |
| Approach: | They propose an efficient beam search black-box adversarial attack method . they use contextual semantic filtering to effectively reduce the search space . |
| Outcome: | The proposed method shows good performance in terms of attack success rate, number of queries, and semantic similarity for two tasks: NL code search classification and retrieval tasks. |
DIP: Dead code Insertion based Black-box Attack for Programming Language Model (2023.acl-long)
Copied to clipboard
| Challenge: | Existing methods to attack natural language models are difficult to apply due to the requirements. |
| Approach: | They propose a black-box attack method that generates adversarial examples using dead code insertion. |
| Outcome: | The proposed method outperforms the state-of-the-art black-box attack in both attack efficiency and attack quality on 9 victim downstream-task large code models. |
BLOCSUM: Block Scope-based Source Code Summarization via Shared Block Representation (2023.findings-acl)
Copied to clipboard
| Challenge: | Abstract Syntax Tree (AST) and sequence of code tokens are useful for code summarization. |
| Approach: | They propose a shared block position embedding to represent various code blocks . they also develop variant ASTs to learn rich information such as block and global dependencies . |
| Outcome: | The proposed method improves on two real-world datasets, including ablation studies and a human evaluation. |
Learning Sequential and Structural Information for Source Code Summarization (2021.findings-acl)
Copied to clipboard
| Challenge: | Existing models that learn both sequential and structural features of source code are limited in their use. |
| Approach: | They propose a model that learns both the sequential and the structural features of code for source code summarization. |
| Outcome: | The proposed model outperforms the state-of-the-art for source code summarization by experiments and human evaluations. |
Code Defect Detection Using Pre-trained Language Models with Encoder-Decoder via Line-Level Defect Localization (2024.lrec-main)
Copied to clipboard
| Challenge: | Recent code Pre-trained Language Models (PLMs) have shown great success in code defect detection tasks. |
| Approach: | They propose a method that integrates line-level defect localization into a unified training process to identify which lines contain defects. |
| Outcome: | The proposed method significantly improves performance on four benchmark datasets for code defect detection. |
UCGRec: User-Centric Graph Learning for LLM-based Sequential Recommendation (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for sequential recommendation rely primarily on item descriptions or utilize user preferences independently. |
| Approach: | They propose a method that integrates diverse user-relevant preference signals into a unified user-centric graph and injects the graph-based knowledge into the LLM through end-to-end training with graph neural networks. |
| Outcome: | The proposed method outperforms conventional and state-of-the-art methods on four widely used sequential real-world recommendation datasets. |
Unsupervised Detection of LLM-Generated Text in Korean Using Syntactic and Semantic Cues (2026.findings-eacl)
Copied to clipboard
| Challenge: | Prior work focused on English, leaving low-resource languages such as Korean underexplored. |
| Approach: | They propose an unsupervised framework that integrates syntactic token cohesiveness and semantic regeneration similarity to detect Korean text. |
| Outcome: | The proposed framework outperforms baselines in Korean and other low-resource languages without training. |
SALAD: Improving Robustness and Generalization through Contrastive Learning with Structure-Aware and LLM-Driven Augmented Data (2025.naacl-long)
Copied to clipboard
| Challenge: | In many natural language processing tasks, model training often leads to spurious correlations . shortcuts allow models to rely on irrelevant patterns in the data, leading to biased predictions. |
| Approach: | They propose a method to generate structure-aware positive and negative sentences using tagging. |
| Outcome: | The proposed method improves model robustness and generalization across different environments while minimizing spurious correlations. |
DCG-SQL: Enhancing In-Context Learning for Text-to-SQL with Deep Contextual Schema Link Graph (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for Text-to-SQL show little improvement compared to random selections . Existing approaches rely on intrinsic capabilities of hyper-scaled LLMs, not useful demonstrations. |
| Approach: | They propose a novel approach to effectively retrieving demonstrations and generating SQL queries by linking a question and its database schema items. |
| Outcome: | The proposed method shows consistent improvements in performance and efficiency across hyper-scaled LLMs and small LLM. |
CoRAC: Integrating Selective API Document Retrieval with Question Semantic Intent for Code Question Answering (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing automated code question answering methods provide accurate and relevant answers to questions about code. |
| Approach: | They propose a knowledge-based framework that generates precise code question answers by analyzing code snippets. |
| Outcome: | The proposed framework generates high-quality answers compared to large language models, such as ChatGPT. |
Q-FAKER: Query-free Hard Black-box Attack via Controlled Generation (2025.findings-naacl)
Copied to clipboard
| Challenge: | Existing adversarial approaches to verify the vulnerability of language models require numerous queries and the information on the target model. |
| Approach: | They propose a method that generates adversarial examples without accessing the target model and uses controlled generation techniques to generate adversarials. |
| Outcome: | The proposed method generates adversarial examples without accessing the target model and is highly transferable and high quality. |