Hojin Lee

Sangwon Yu, Changmin Lee, Hojin Lee et al.

Controlled text generation is very important for the practical use of language models because it ensures that the produced text includes only the desired attributes from a specific domain or dataset. Existing methods, however, are inapplicable to black-box models or suffer a significant trade-off between controlling the generated text and maintaining its fluency. This paper introduces the Score-based Progressive Editor (ScoPE), a novel approach designed to overcome these issues. ScoPE modifies the context at the token level during the generation process of a backbone language model. This modification guides the subsequent text to naturally include the target attributes. To facilitate this process, ScoPE employs a training objective that maximizes a target score, thoroughly considering both the ability to guide the text and its fluency. Experimental results on diverse controlled generation tasks demonstrate that ScoPE can effectively regulate the attributes of the generated text while fully utilizing the capability of the backbone large language models. Our codes are available at \url{https://github.com/ysw1021/ScoPE}.

Taekyung Kim, Hojin Lee, Wonsuk Lee

Non-holonomic vehicle motion has been studied extensively using physics-based models. Common approaches when using these models interpret the wheel/ground interactions using a linear tire model and thus may not fully capture the nonlinear and complex dynamics under various environments. On the other hand, neural network models have been widely employed in this domain, demonstrating powerful function approximation capabilities. However, these black-box learning strategies completely abandon the existing knowledge of well-known physics. In this paper, we seamlessly combine deep learning with a fully differentiable physics model to endow the neural network with available prior knowledge. The proposed model shows better generalization performance than the vanilla neural network model by a large margin. We also show that the latent features of our model can accurately represent lateral tire forces without the need for any additional training. Lastly, We develop a risk-aware model predictive controller using proprioceptive information derived from the latent features. We validate our idea in two autonomous driving tasks under unknown friction, outperforming the baseline control framework.

Kanana LLM Team, Yunju Bak, Hojin Lee et al.

We introduce Kanana, a series of bilingual language models that demonstrate exceeding performance in Korean and competitive performance in English. The computational cost of Kanana is significantly lower than that of state-of-the-art models of similar size. The report details the techniques employed during pre-training to achieve compute-efficient yet competitive models, including high quality data filtering, staged pre-training, depth up-scaling, and pruning and distillation. Furthermore, the report outlines the methodologies utilized during the post-training of the Kanana models, encompassing supervised fine-tuning and preference optimization, aimed at enhancing their capability for seamless interaction with users. Lastly, the report elaborates on plausible approaches used for language model adaptation to specific scenarios, such as embedding, retrieval augmented generation, and function calling. The Kanana model series spans from 2.1B to 32.5B parameters with 2.1B models (base, instruct, embedding) publicly released to promote research on Korean language models.

Akash Vishwakarma, Hojin Lee, Mohith Suresh et al.

The emergence of capable large language model (LLM) based agents necessitates memory architectures that transcend mere data storage, enabling continuous learning, nuanced reasoning, and dynamic adaptation. Current memory systems often grapple with fundamental limitations in structural flexibility, temporal awareness, and the ability to synthesize higher-level insights from raw interaction data. This paper introduces Cognitive Weave, a novel memory framework centered around a multi-layered spatio-temporal resonance graph (STRG). This graph manages information as semantically rich insight particles (IPs), which are dynamically enriched with resonance keys, signifiers, and situational imprints via a dedicated semantic oracle interface (SOI). These IPs are interconnected through typed relational strands, forming an evolving knowledge tapestry. A key component of Cognitive Weave is the cognitive refinement process, an autonomous mechanism that includes the synthesis of insight aggregates (IAs) condensed, higher-level knowledge structures derived from identified clusters of related IPs. We present comprehensive experimental results demonstrating Cognitive Weave's marked enhancement over existing approaches in long-horizon planning tasks, evolving question-answering scenarios, and multi-session dialogue coherence. The system achieves a notable 34% average improvement in task completion rates and a 42% reduction in mean query latency when compared to state-of-the-art baselines. Furthermore, this paper explores the ethical considerations inherent in such advanced memory systems, discusses the implications for long-term memory in LLMs, and outlines promising future research trajectories.

Hojin Lee, Taekyung Kim, Jungwi Mun et al.

High-speed autonomous driving in off-road environments has immense potential for various applications, but it also presents challenges due to the complexity of vehicle-terrain interactions. In such environments, it is crucial for the vehicle to predict its motion and adjust its controls proactively in response to environmental changes, such as variations in terrain elevation. To this end, we propose a method for learning terrain-aware kinodynamic model which is conditioned on both proprioceptive and exteroceptive information. The proposed model generates reliable predictions of 6-degree-of-freedom motion and can even estimate contact interactions without requiring ground truth force data during training. This enables the design of a safe and robust model predictive controller through appropriate cost function design which penalizes sampled trajectories with unstable motion, unsafe interactions, and high levels of uncertainty derived from the model. We demonstrate the effectiveness of our approach through experiments on a simulated off-road track, showing that our proposed model-controller pair outperforms the baseline and ensures robust high-speed driving performance without control failure.

Taekyung Kim, Hojin Lee, Seongil Hong et al.

Neural networks have been increasingly employed in Model Predictive Controller (MPC) to control nonlinear dynamic systems. However, MPC still poses a problem that an achievable update rate is insufficient to cope with model uncertainty and external disturbances. In this paper, we present a novel control scheme that can design an optimal tracking controller using the neural network dynamics of the MPC, making it possible to be applied as a plug-and-play extension for any existing model-based feedforward controller. We also describe how our method handles a neural network containing history information, which does not follow a general form of dynamics. The proposed method is evaluated by its performance in classical control benchmarks with external disturbances. We also extend our control framework to be applied in an aggressive autonomous driving task with unknown friction. In all experiments, our method outperformed the compared methods by a large margin. Our controller also showed low control chattering levels, demonstrating that our feedback controller does not interfere with the optimal command of MPC.

Kyudong Park, Sung H. Han, Hojin Lee

The lane keeping assistance system (LKAS), a representative of the advanced driver assistance system (ADAS), comprises a shared control that cooperates with the driver to achieve a common goal. The experience of the driver through the steering wheel may vary significantly depending on the steering control strategy of the system. In this study, we examine how driving experience changes according to various steering control strategies. Based on the preliminary study and typical LKAS parameters, nine control strategies (3 torque amounts (TOR) x 3 deviations to start control (DEV)) were designed as a prototype. Eighteen participants participated in evaluating each strategy in a highway environment provided by a driving simulator. Two-way repeated measure ANOVA was used to assess the effects of the system. Both the objective measures (standard deviation of lane position, steering reversal rate, and root mean square of lateral speed) and subjective measures (pleasure and arousal of emotion, trust, disturbance, and satisfaction) are analyzed. The experimental results demonstrate that all dependent measures are significant. As the TOR increased, SDLP decreased. However, no difference is observed between the 2-Nm and 3-Nm TOR in terms of trust and satisfaction. The high disturbance and negative emotion in 3 Nm appear to be the cause. In terms of the DEV, the high level of the root mean square of the lateral speed is observed at 0.8 m. Further, negative effects are found in pleasure, trust, and satisfaction. There is little difference at all dependent measures between 0.0-m and 0.4-m DEV. In the regression model analyzed from the aspect of satisfaction, the 2.32-Nm TOR and 0.27-m DEV are the optimal values. We expect our research on shared steering control with an assistance system to be applied to the experience design of a lateral semi-autonomous vehicle.

Hojin Lee, Hyoungkyun Kim, Seungmoon Choi

This paper addresses a data-driven framework, modeling expert driving skills for performance-based haptic assistance using neural networks (NNs). We have built a haptic driving training simulator to collect expert driving data and to provide proper haptic feedback. We establish an expert driving skill model by training NNs with the collected data. Then, the skill model is applied to performance-based haptic assistance to provide optimized references of the steering/pedaling movements. We evaluate the skill model and its application to performance-based haptic assistance in two user experiments. The results of the first experiment demonstrate that our skill model has appropriately captured experts' steering/pedaling skills. The results of the second experiment show that our performance-based haptic assistance can help novice drivers perform steering as expert drivers, but cannot assist their pedaling performance.