Jong-hun Shin

Hongjin Kim, Jaewook Lee, Kiyoung Lee et al.

Large Language Models (LLMs) demonstrate strong reasoning and self-correction abilities in high-resource languages like English, but their performance remains limited in low-resource languages such as Korean. In this study, we investigate whether reinforcement learning (RL) can enhance Korean reasoning abilities to a degree comparable to English. Our findings reveal that RL alone yields limited improvements when applied to models lacking inherent Korean reasoning capabilities. To address this, we explore several fine-tuning strategies and show that aligning the model's internal reasoning processes with Korean inputs-particularly by tuning Korean-specific neurons in early layers-is key to unlocking RL's effectiveness. We introduce a self-correction code-switching dataset to facilitate this alignment and observe significant performance gains in both mathematical reasoning and self-correction tasks. Ultimately, we conclude that the crucial factor in multilingual reasoning enhancement is not injecting new linguistic knowledge, but effectively eliciting and aligning existing reasoning capabilities. Our study provides a new perspective on how internal translation and neuron-level tuning contribute to multilingual reasoning alignment in LLMs.

Jin Seong, Wencke Liermann, Minho Kim et al.

Accurate transcription of handwritten mathematics is crucial for educational AI systems, yet current benchmarks fail to evaluate this capability properly. Most prior studies focus on single-line expressions and rely on lexical metrics such as BLEU, which fail to assess the semantic reasoning across multi-line student solutions. In this paper, we present the first systematic study of multi-line handwritten math Optical Character Recognition (OCR), revealing a critical failure mode of Vision-Language Models (VLMs): over-correction. Instead of faithfully transcribing a student's work, these models often "fix" errors, thereby hiding the very mistakes an educational assessment aims to detect. To address this, we propose PINK (Penalized INK-based score), a semantic evaluation metric that leverages a Large Language Model (LLM) for rubric-based grading and explicitly penalizes over-correction. Our comprehensive evaluation of 15 state-of-the-art VLMs on the FERMAT dataset reveals substantial ranking reversals compared to BLEU: models like GPT-4o are heavily penalized for aggressive over-correction, whereas Gemini 2.5 Flash emerges as the most faithful transcriber. Furthermore, human expert studies show that PINK aligns significantly better with human judgment (55.0% preference over BLEU's 39.5%), providing a more reliable evaluation framework for handwritten math OCR in educational settings.

HeeSeung Jung, Kangil Kim, Hoyong Kim et al.

The flexibility of decision boundaries in neural networks that are unguided by training data is a well-known problem typically resolved with generalization methods. A surprising result from recent knowledge distillation (KD) literature is that random, untrained, and equally structured teacher networks can also vastly improve generalization performance. It raises the possibility of existence of undiscovered assumptions useful for generalization on an uncertain region. In this paper, we shed light on the assumptions by analyzing decision boundaries and confidence distributions of both simple and KD-based generalization methods. Assuming that a decision boundary exists to represent the most general tendency of distinction on an input sample space (i.e., the simplest hypothesis), we show the various limitations of methods when using the hypothesis. To resolve these limitations, we propose matured dumb teacher based KD, conservatively transferring the hypothesis for generalization of the student without massive destruction of trained information. In practical experiments on feed-forward and convolution neural networks for image classification tasks on MNIST, CIFAR-10, and CIFAR-100 datasets, the proposed method shows stable improvement to the best test performance in the grid search of hyperparameters. The analysis and results imply that the proposed method can provide finer generalization than existing methods.

Gyu-Hyeon Choi, Jong-Hun Shin, Young-Kil Kim

In machine translation, we often try to collect resources to improve performance. However, most of the language pairs, such as Korean-Arabic and Korean-Vietnamese, do not have enough resources to train machine translation systems. In this paper, we propose the use of synthetic methods for extending a low-resource corpus and apply it to a multi-source neural machine translation model. We showed the improvement of machine translation performance through corpus extension using the synthetic method. We specifically focused on how to create source sentences that can make better target sentences, including the use of synthetic methods. We found that the corpus extension could also improve the performance of multi-source neural machine translation. We showed the corpus extension and multi-source model to be efficient methods for a low-resource language pair. Furthermore, when both methods were used together, we found better machine translation performance.