Wonseok Hwang

Wonseok Hwang, Dongjun Lee, Kyoungyeon Cho et al.

The recent advances of deep learning have dramatically changed how machine learning, especially in the domain of natural language processing, can be applied to legal domain. However, this shift to the data-driven approaches calls for larger and more diverse datasets, which are nevertheless still small in number, especially in non-English languages. Here we present the first large-scale benchmark of Korean legal AI datasets, LBOX OPEN, that consists of one legal corpus, two classification tasks, two legal judgement prediction (LJP) tasks, and one summarization task. The legal corpus consists of 147k Korean precedents (259M tokens), of which 63k are sentenced in last 4 years and 96k are from the first and the second level courts in which factual issues are reviewed. The two classification tasks are case names (11.3k) and statutes (2.8k) prediction from the factual description of individual cases. The LJP tasks consist of (1) 10.5k criminal examples where the model is asked to predict fine amount, imprisonment with labor, and imprisonment without labor ranges for the given facts, and (2) 4.7k civil examples where the inputs are facts and claim for relief and outputs are the degrees of claim acceptance. The summarization task consists of the Supreme Court precedents and the corresponding summaries (20k). We also release realistic variants of the datasets by extending the domain (1) to infrequent case categories in case name (31k examples) and statute (17.7k) classification tasks, and (2) to long input sequences in the summarization task (51k). Finally, we release LCUBE, the first Korean legal language model trained on the legal corpus from this study. Given the uniqueness of the Law of South Korea and the diversity of the legal tasks covered in this work, we believe that LBOX OPEN contributes to the multilinguality of global legal research. LBOX OPEN and LCUBE will be publicly available.

Wonseok Hwang, Saehee Eom, Hanuhl Lee et al.

Legal practitioners often face a vast amount of documents. Lawyers, for instance, search for appropriate precedents favorable to their clients, while the number of legal precedents is ever-growing. Although legal search engines can assist finding individual target documents and narrowing down the number of candidates, retrieved information is often presented as unstructured text and users have to examine each document thoroughly which could lead to information overloading. This also makes their statistical analysis challenging. Here, we present an end-to-end information extraction (IE) system for legal documents. By formulating IE as a generation task, our system can be easily applied to various tasks without domain-specific engineering effort. The experimental results of four IE tasks on Korean precedents shows that our IE system can achieve competent scores (-2.3 on average) compared to the rule-based baseline with as few as 50 training examples per task and higher score (+5.4 on average) with 200 examples. Finally, our statistical analysis on two case categories--drunk driving and fraud--with 35k precedents reveals the resulting structured information from our IE system faithfully reflects the macroscopic features of Korean legal system.

Hongseok Oh, Wonseok Hwang, Kyoung-Woon On

We introduce the Korean Canonical Legal Benchmark (KCL), a benchmark designed to assess language models' legal reasoning capabilities independently of domain-specific knowledge. KCL provides question-level supporting precedents, enabling a more faithful disentanglement of reasoning ability from parameterized knowledge. KCL consists of two components: (1) KCL-MCQA, multiple-choice problems of 283 questions with 1,103 aligned precedents, and (2) KCL-Essay, open-ended generation problems of 169 questions with 550 aligned precedents and 2,739 instance-level rubrics for automated evaluation. Our systematic evaluation of 30+ models shows large remaining gaps, particularly in KCL-Essay, and that reasoning-specialized models consistently outperform their general-purpose counterparts. We release all resources, including the benchmark dataset and evaluation code, at https://github.com/lbox-kr/kcl.

Yeeun Kim, Hyunseo Shin, Eunkyung Choi et al.

Open source is a driving force behind scientific advancement.However, this openness is also a double-edged sword, with the inherent risk that innovative technologies can be misused for purposes harmful to society. What is the likelihood that an open source AI model or dataset will be used to commit a real-world crime, and if a criminal does exploit it, will the people behind the technology be able to escape legal liability? To address these questions, we explore a legal domain where individual choices can have a significant impact on society. Specifically, we build the EVE-V1 dataset that comprises 200 question-answer pairs related to criminal offenses based on 200 Korean precedents first to explore the possibility of malicious models emerging. We further developed EVE-V2 using 600 fraud-related precedents to confirm the existence of malicious models that can provide harmful advice on a wide range of criminal topics to test the domain generalization ability. Remarkably, widely used open-source large-scale language models (LLMs) provide unethical and detailed information about criminal activities when fine-tuned with EVE. We also take an in-depth look at the legal issues that malicious language models and their builders could realistically face. Our findings highlight the paradoxical dilemma that open source accelerates scientific progress, but requires great care to minimize the potential for misuse. Warning: This paper contains content that some may find unethical.

Minhu Park, Hongseok Oh, Eunkyung Choi et al.

Recently, building retrieval-augmented generation (RAG) systems to enhance the capability of large language models (LLMs) has become a common practice. Especially in the legal domain, previous judicial decisions play a significant role under the doctrine of stare decisis which emphasizes the importance of making decisions based on (retrieved) prior documents. However, the overall performance of RAG system depends on many components: (1) retrieval corpora, (2) retrieval algorithms, (3) rerankers, (4) LLM backbones, and (5) evaluation metrics. Here we propose LRAGE, an open-source tool for holistic evaluation of RAG systems focusing on the legal domain. LRAGE provides GUI and CLI interfaces to facilitate seamless experiments and investigate how changes in the aforementioned five components affect the overall accuracy. We validated LRAGE using multilingual legal benches including Korean (KBL), English (LegalBench), and Chinese (LawBench) by demonstrating how the overall accuracy changes when varying the five components mentioned above. The source code is available at https://github.com/hoorangyee/LRAGE.

Eunkyung Choi, Young Jin Suh, Hun Park et al.

How capable are large language models (LLMs) in the domain of taxation? Although numerous studies have explored the legal domain in general, research dedicated to taxation remain scarce. Moreover, the datasets used in these studies are either simplified, failing to reflect the real-world complexities, or unavailable as open source. To address this gap, we introduce PLAT, a new benchmark designed to assess the ability of LLMs to predict the legitimacy of additional tax penalties. PLAT is constructed to evaluate LLMs' understanding of tax law, particularly in cases where resolving the issue requires more than just applying related statutes. Our experiments with six LLMs reveal that their baseline capabilities are limited, especially when dealing with conflicting issues that demand a comprehensive understanding. However, we found that enabling retrieval, self-reasoning, and discussion among multiple agents with specific role assignments, this limitation can be mitigated.

Geewook Kim, Teakgyu Hong, Moonbin Yim et al.

Understanding document images (e.g., invoices) is a core but challenging task since it requires complex functions such as reading text and a holistic understanding of the document. Current Visual Document Understanding (VDU) methods outsource the task of reading text to off-the-shelf Optical Character Recognition (OCR) engines and focus on the understanding task with the OCR outputs. Although such OCR-based approaches have shown promising performance, they suffer from 1) high computational costs for using OCR; 2) inflexibility of OCR models on languages or types of document; 3) OCR error propagation to the subsequent process. To address these issues, in this paper, we introduce a novel OCR-free VDU model named Donut, which stands for Document understanding transformer. As the first step in OCR-free VDU research, we propose a simple architecture (i.e., Transformer) with a pre-training objective (i.e., cross-entropy loss). Donut is conceptually simple yet effective. Through extensive experiments and analyses, we show a simple OCR-free VDU model, Donut, achieves state-of-the-art performances on various VDU tasks in terms of both speed and accuracy. In addition, we offer a synthetic data generator that helps the model pre-training to be flexible in various languages and domains. The code, trained model and synthetic data are available at https://github.com/clovaai/donut.

Teakgyu Hong, Donghyun Kim, Mingi Ji et al.

Key information extraction (KIE) from document images requires understanding the contextual and spatial semantics of texts in two-dimensional (2D) space. Many recent studies try to solve the task by developing pre-trained language models focusing on combining visual features from document images with texts and their layout. On the other hand, this paper tackles the problem by going back to the basic: effective combination of text and layout. Specifically, we propose a pre-trained language model, named BROS (BERT Relying On Spatiality), that encodes relative positions of texts in 2D space and learns from unlabeled documents with area-masking strategy. With this optimized training scheme for understanding texts in 2D space, BROS shows comparable or better performance compared to previous methods on four KIE benchmarks (FUNSD, SROIE*, CORD, and SciTSR) without relying on visual features. This paper also reveals two real-world challenges in KIE tasks-(1) minimizing the error from incorrect text ordering and (2) efficient learning from fewer downstream examples-and demonstrates the superiority of BROS over previous methods. Code is available at https://github.com/clovaai/bros.

Hongseok Oh, Wonseok Hwang

Large Language Models (LLMs) have shown remarkable progress, but their real-world application necessitates reliable calibration. This study conducts a comprehensive analysis of calibration degradation of LLMs across four dimensions: models, calibration metrics, tasks, and confidence extraction methods. Initial analysis showed that the relationship between alignment and calibration is not always a trade-off, but under stricter analysis conditions, we found the alignment process consistently harms calibration. This highlights the need for (1) a careful approach when measuring model confidences and calibration errors and (2) future research into algorithms that can help LLMs to achieve both instruction-following and calibration without sacrificing either.

Kyoungyeon Cho, Seungkum Han, Young Rok Choi et al.

The statistical analysis of large scale legal corpus can provide valuable legal insights. For such analysis one needs to (1) select a subset of the corpus using document retrieval tools, (2) structure text using information extraction (IE) systems, and (3) visualize the data for the statistical analysis. Each process demands either specialized tools or programming skills whereas no comprehensive unified "no-code" tools have been available. Here we provide NESTLE, a no-code tool for large-scale statistical analysis of legal corpus. Powered by a Large Language Model (LLM) and the internal custom end-to-end IE system, NESTLE can extract any type of information that has not been predefined in the IE system opening up the possibility of unlimited customizable statistical analysis of the corpus without writing a single line of code. We validate our system on 15 Korean precedent IE tasks and 3 legal text classification tasks from LexGLUE. The comprehensive experiments reveal NESTLE can achieve GPT-4 comparable performance by training the internal IE module with 4 human-labeled, and 192 LLM-labeled examples.

Hyunseo Shin, Wonseok Hwang

Despite the rapid advancements of Large Language Models (LLMs), safety risks remain a critical challenge for low-resource languages. Existing safety datasets are predominantly English centric, limiting progress in multilingual safety alignment. As a result, low resource expert models, finetuned on their respective instruction datasets, tend to exhibit higher unsafety rates compared to their high resource counterparts. In this work, we propose a safety aware layer swapping method that transfers safety alignment from an English safety expert to low resource language experts without additional training. To further enhance transfer ability, our method adaptively selects or blends modules based on their degree of specialization. Our approach preserves performance on general language understanding tasks while enhancing safety in the target languages. Experimental results show that the proposed method achieves comparable performance to the language expert on general benchmarks such as MMMLU, BELEBELE, and MGSM, while producing more aligned and less harmful responses on the MultiJail safety benchmark.

Jinu Lee, Wonseok Hwang

To improve the performance and explainability of LLM-based natural language reasoning, structured reasoning can be applied to generate explicitly structured proofs. Among different methods for structured reasoning, we specifically focus on backward chaining, where the proof goal is recursively decomposed to subgoals by searching and applying rules. We argue that current LLM-based backward chaining systems (e.g. Least-to-most prompting and LAMBADA) are incomplete, as they omit crucial algorithmic components identified from the classic backward chaining algorithm in computational logic (SLD Resolution). To this end, we propose a novel backward chaining system, SymBa (Symbolic Backward Chaining), which integrates a symbolic solver and an LLM. In SymBa, the solver controls the proof process, and the LLM is only called when the solver requires new information to complete the proof. Empowered by completeness, SymBa achieves a significant improvement in seven deductive, relational, and arithmetic reasoning benchmarks compared to the baselines.

Hongseok Oh, Wonseok Hwang

Recently, Large Vision-Language Models (LVLMs) show remarkable performance across various domains. However, these models suffer from object hallucination. This study revisits the previous claim that the cause of such hallucinations lies in the limited representational capacity of the vision encoder. Our analysis implies that the capacity of the vision encoder is not necessarily a major limiting factor in detecting object hallucination. Based on this insight, we propose Fine-grained CLIPScore (F-CLIPScore), a simple yet effective evaluation metric that enhances object-level granularity by incorporating text embeddings at the noun level. Evaluations on the OHD-Caps benchmark show that F-CLIPScore significantly outperforms conventional CLIPScore in accuracy by a large margin of \textbf{39.6\%} without additional training. We further demonstrate that F-CLIPScore-based data filtering reduces object hallucination in LVLM (4.9\% in POPE).

Yeeun Kim, Young Rok Choi, Eunkyung Choi et al.

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. This underscores the need for careful evaluation of LLMs within each legal system before application. Here, we introduce KBL, a benchmark for assessing the Korean legal language understanding of LLMs, consisting of (1) 7 legal knowledge tasks (510 examples), (2) 4 legal reasoning tasks (288 examples), and (3) the Korean bar exam (4 domains, 53 tasks, 2,510 examples). First two datasets were developed in close collaboration with lawyers to evaluate LLMs in practical scenarios in a certified manner. Furthermore, considering legal practitioners' frequent use of extensive legal documents for research, we assess LLMs in both a closed book setting, where they rely solely on internal knowledge, and a retrieval-augmented generation (RAG) setting, using a corpus of Korean statutes and precedents. The results indicate substantial room and opportunities for improvement.

Chaeeun Kim, Jinu Lee, Wonseok Hwang

Legal Case Retrieval (LCR), which retrieves relevant cases from a query case, is a fundamental task for legal professionals in research and decision-making. However, existing studies on LCR face two major limitations. First, they are evaluated on relatively small-scale retrieval corpora (e.g., 100-55K cases) and use a narrow range of criminal query types, which cannot sufficiently reflect the complexity of real-world legal retrieval scenarios. Second, their reliance on embedding-based or lexical matching methods often results in limited representations and legally irrelevant matches. To address these issues, we present: (1) LEGAR BENCH, the first large-scale Korean LCR benchmark, covering 411 diverse crime types in queries over 1.2M candidate cases; and (2) LegalSearchLM, a retrieval model that performs legal element reasoning over the query case and directly generates content containing those elements, grounded in the target cases through constrained decoding. Experimental results show that LegalSearchLM outperforms baselines by 6-20% on LEGAR BENCH, achieving state-of-the-art performance. It also demonstrates strong generalization to out-of-domain cases, outperforming naive generative models trained on in-domain data by 15%.

Hyunseo Shin, Wonseok Hwang

Recent advancements in large language models (LLMs) have significantly enhanced the performance of conversational AI systems. To extend their capabilities to knowledge-intensive domains such as biomedical and legal fields, where the accuracy is critical, LLMs are often combined with information retrieval (IR) systems to generate responses based on retrieved documents. However, for IR systems to effectively support such applications, they must go beyond simple semantic matching and accurately capture diverse query intents, including causal relationships. Existing IR models primarily focus on retrieving documents based on surface-level semantic similarity, overlooking deeper relational structures such as causality. To address this, we propose CAWAI, a retrieval model that is trained with dual objectives: semantic and causal relations. Our extensive experiments demonstrate that CAWAI outperforms various models on diverse causal retrieval tasks especially under large-scale retrieval settings. We also show that CAWAI exhibits strong zero-shot generalization across scientific domain QA tasks.

Geewook Kim, Wonseok Hwang, Minjoon Seo et al.

Semi-structured query systems for document-oriented databases have many real applications. One particular application that we are interested in is matching each financial receipt image with its corresponding place of interest (POI, e.g., restaurant) in the nationwide database. The problem is especially challenging in the real production environment where many similar or incomplete entries exist in the database and queries are noisy (e.g., errors in optical character recognition). In this work, we aim to address practical challenges when using embedding-based retrieval for the query grounding problem in semi-structured data. Leveraging recent advancements in deep language encoding for retrieval, we conduct extensive experiments to find the most effective combination of modules for the embedding and retrieval of both query and database entries without any manually engineered component. The proposed model significantly outperforms the conventional manual pattern-based model while requiring much less development and maintenance cost. We also discuss some core observations in our experiments, which could be helpful for practitioners working on a similar problem in other domains.

Wonseok Hwang, Hyunji Lee, Jinyeong Yim et al.

A real-world information extraction (IE) system for semi-structured document images often involves a long pipeline of multiple modules, whose complexity dramatically increases its development and maintenance cost. One can instead consider an end-to-end model that directly maps the input to the target output and simplify the entire process. However, such generation approach is known to lead to unstable performance if not designed carefully. Here we present our recent effort on transitioning from our existing pipeline-based IE system to an end-to-end system focusing on practical challenges that are associated with replacing and deploying the system in real, large-scale production. By carefully formulating document IE as a sequence generation task, we show that a single end-to-end IE system can be built and still achieve competent performance.

Juno Hwang, Wonseok Hwang, Junghyo Jo

The restricted Boltzmann machine (RBM) is a representative generative model based on the concept of statistical mechanics. In spite of the strong merit of interpretability, unavailability of backpropagation makes it less competitive than other generative models. Here we derive differentiable loss functions for both binary and multinary RBMs. Then we demonstrate their learnability and performance by generating colored face images.

Wonseok Hwang, Jinyeong Yim, Seunghyun Park et al.

Deep learning approaches to semantic parsing require a large amount of labeled data, but annotating complex logical forms is costly. Here, we propose Syntactic Question Abstraction and Retrieval (SQAR), a method to build a neural semantic parser that translates a natural language (NL) query to a SQL logical form (LF) with less than 1,000 annotated examples. SQAR first retrieves a logical pattern from the train data by computing the similarity between NL queries and then grounds a lexical information on the retrieved pattern in order to generate the final LF. We validate SQAR by training models using various small subsets of WikiSQL train data achieving up to 4.9% higher LF accuracy compared to the previous state-of-the-art models on WikiSQL test set. We also show that by using query-similarity to retrieve logical pattern, SQAR can leverage a paraphrasing dataset achieving up to 5.9% higher LF accuracy compared to the case where SQAR is trained by using only WikiSQL data. In contrast to a simple pattern classification approach, SQAR can generate unseen logical patterns upon the addition of new examples without re-training the model. We also discuss an ideal way to create cost efficient and robust train datasets when the data distribution can be approximated under a data-hungry setting.

Wonseok Hwang, Jinyeong Yim, Seunghyun Park et al.

Information Extraction (IE) for semi-structured document images is often approached as a sequence tagging problem by classifying each recognized input token into one of the IOB (Inside, Outside, and Beginning) categories. However, such problem setup has two inherent limitations that (1) it cannot easily handle complex spatial relationships and (2) it is not suitable for highly structured information, which are nevertheless frequently observed in real-world document images. To tackle these issues, we first formulate the IE task as spatial dependency parsing problem that focuses on the relationship among text tokens in the documents. Under this setup, we then propose SPADE (SPAtial DEpendency parser) that models highly complex spatial relationships and an arbitrary number of information layers in the documents in an end-to-end manner. We evaluate it on various kinds of documents such as receipts, name cards, forms, and invoices, and show that it achieves a similar or better performance compared to strong baselines including BERT-based IOB taggger.

Wonseok Hwang, Jinyeong Yim, Seunghyun Park et al.

We present SQLova, the first Natural-language-to-SQL (NL2SQL) model to achieve human performance in WikiSQL dataset. We revisit and discuss diverse popular methods in NL2SQL literature, take a full advantage of BERT {Devlin et al., 2018) through an effective table contextualization method, and coherently combine them, outperforming the previous state of the art by 8.2% and 2.5% in logical form and execution accuracy, respectively. We particularly note that BERT with a seq2seq decoder leads to a poor performance in the task, indicating the importance of a careful design when using such large pretrained models. We also provide a comprehensive analysis on the dataset and our model, which can be helpful for designing future NL2SQL datsets and models. We especially show that our model's performance is near the upper bound in WikiSQL, where we observe that a large portion of the evaluation errors are due to wrong annotations, and our model is already exceeding human performance by 1.3% in execution accuracy.