Junseung Lee

Jinho Choo, JunSeung Lee, Jimyeong Kim et al.

Large language models (LLMs) demonstrate strong multilingual capabilities, yet often fail to consistently generate responses in the intended language, exhibiting a phenomenon known as language confusion. Prior mitigation approaches based on sequence-level fine-tuning, such as DPO, ORPO, and GRPO, operate at the level of entire responses and can lead to unintended degradation of general model capabilities, motivating the need for more fine-grained alternatives. To address this, we introduce Token-Level Policy Optimization (TLPO), a fine-tuning framework designed to mitigate language confusion through localized, token-level updates. TLPO identifies error-prone positions, explores alternative candidate tokens, and updates the policy using a tailored objective to suppress error-inducing outputs at a granular level. This selective intervention enables effective mitigation of language confusion without compromising the model's general abilities. Experiments on multiple multilingual LLMs across diverse languages demonstrate that TLPO significantly outperforms baselines in improving language consistency while preserving downstream task accuracy.

Hyungu Kahng, Yonghyun Jeong, Yoon Sang Cho et al.

StarCraft, one of the most popular real-time strategy games, is a compelling environment for artificial intelligence research for both micro-level unit control and macro-level strategic decision making. In this study, we address an eminent problem concerning macro-level decision making, known as the 'fog-of-war', which rises naturally from the fact that information regarding the opponent's state is always provided in the incomplete form. For intelligent agents to play like human players, it is obvious that making accurate predictions of the opponent's status under incomplete information will increase its chance of winning. To reflect this fact, we propose a convolutional encoder-decoder architecture that predicts potential counts and locations of the opponent's units based on only partially visible and noisy information. To evaluate the performance of our proposed method, we train an additional classifier on the encoder-decoder output to predict the game outcome (win or lose). Finally, we designed an agent incorporating the proposed method and conducted simulation games against rule-based agents to demonstrate both effectiveness and practicality. All experiments were conducted on actual game replay data acquired from professional players.