Papers by Wonseok Hwang
SymBa: Symbolic Backward Chaining for Structured Natural Language Reasoning (2025.naacl-long)
Copied to clipboard
| Challenge: | Among different methods for structured reasoning, we focus on backward chaining, where the goal is recursively decomposed into subgoals by searching and applying rules. |
| Approach: | They propose a backward chaining system that integrates a symbolic solver and an LLM to improve the performance of LLM-based reasoning. |
| Outcome: | The proposed system improves deductive, relational, and arithmetic reasoning benchmarks compared to baselines. |
Layer-wise Swapping for Generalizable Multilingual Safety (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing safety datasets are predominantly English-centric, limiting progress in multilingual safety alignment. |
| Approach: | They propose a safety-aware layer swapping method that transfers alignment from an English safety expert to low-resource language experts without additional training. |
| Outcome: | The proposed method preserves performance on general language understanding tasks while enhancing safety in the target languages. |
Taxation Perspectives from Large Language Models: A Case Study on Additional Tax Penalties (2026.eacl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) have demonstrated promising results across various domains, including the legal domain. |
| Approach: | They propose a benchmark to assess the ability of large language models to predict the legitimacy of additional tax penalties. |
| Outcome: | The proposed model is based on 100 Korean court precedents and 100 binary-choice questions. |
Cost-effective End-to-end Information Extraction for Semi-structured Document Images (2021.emnlp-main)
Copied to clipboard
| Challenge: | a real-world information extraction system for semi-structured document images often involves a long pipeline of multiple modules, which can lead to unstable performance if not designed carefully. |
| Approach: | They propose to use a sequence generation task to build an end-to-end IE system . they propose to combine three manually engineered modules with one data-driven module . |
| Outcome: | The proposed system can be easily replaced and deployed in large-scale production. |
Spatial Dependency Parsing for Semi-Structured Document Information Extraction (2021.findings-acl)
Copied to clipboard
| Challenge: | Information extraction (IE) for semistructured document images is often considered as a sequence tagging problem . however, such a setup cannot handle complex spatial relationships and is not suitable for highly structured information. |
| Approach: | They propose a spatial dependency parsing problem that models complex spatial relationships . they evaluate it on receipts, name cards, forms, and invoices and compare it to other methods . |
| Outcome: | The proposed parser achieves similar or better performance on various kinds of documents compared to baselines including BERT-based IOB taggger. |
LegalSearchLM: Rethinking Legal Case Retrieval as Legal Elements Generation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies on legal case retrieval have limited results . limited representations and legally irrelevant matches are often used . |
| Approach: | They propose a large-scale Korean LCR benchmark and a retrieval model that performs legal element reasoning over the query case. |
| Outcome: | a new model outperforms baseline models on a Korean LCR benchmark . it performs state-of-the-art on 411 diverse crime types in queries over 1.2M candidate cases . previous studies have shown that the model can generalize to out-of domain cases if it is trained on in-domain data . |
Developing a Pragmatic Benchmark for Assessing Korean Legal Language Understanding in Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models (LLMs) have demonstrated remarkable performance in the legal domain, with GPT-4 even passing the Uniform Bar Exam in the U.S. However their efficacy remains limited for non-standardized tasks and tasks in languages other than English. |
| Approach: | They propose a benchmark for assessing the Korean legal language understanding of LLMs consisting of 7 legal knowledge tasks and 4 legal reasoning tasks. |
| Outcome: | The proposed model passes the Uniform Bar Exam in the U.S. but its performance is limited for non-standardized tasks and tasks in languages other than English. |
NESTLE: a No-Code Tool for Statistical Analysis of Legal Corpus (2024.eacl-demo)
Copied to clipboard
| Challenge: | a comprehensive statistical analysis of legal corpus requires specialized tools or programming skills. |
| Approach: | They propose a no-code tool for large-scale statistical analysis of legal corpus . NESTLE can extract any type of information that has not been predefined in the IE system . |
| Outcome: | The proposed tool can perform comparable to LexGLUE on 15 Korean precedent IE tasks and 3 legal text classification tasks. |
Korean Canonical Legal Benchmark: Toward Knowledge-Independent Evaluation of LLMs’ Legal Reasoning Capabilities (2026.eacl-short)
Copied to clipboard
| Challenge: | Large reasoning models trained to reason explicitly in the verbal space have shown superior performance over general large language models (Guo et al., 2025). |
| Approach: | They propose to use Korean Canonical Legal Benchmark to assess language models' legal reasoning capabilities independently of domain-specific knowledge. |
| Outcome: | The proposed benchmark outperforms general-purpose models in a systematic evaluation of 30+ models. |