YLYonghyeon Lee

Yonghyeon Lee

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Yonghyeon Lee has a total of 22 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Rotation-Invariant Multi-IMU Activity Recognition under Independent Per-Location Orientation Shifts

    by Seungyeol Baek, Yoonbyung Chai, Yonghyeon Lee, Sungjoon Choi, Sungho Suh

  • INGRID: Interactive geometry and instance identification for occluded scenes

    by Junho Lee, Sang Min Kim, Yonghyeon Lee, Young Min Kim

  • Open Access

    INGRID:Interactive Geometry and Instance Identification for Occluded Scenes

    by Junho Lee, Sang Min Kim, Yonghyeon Lee, Young Min Kim

    While identifying object instances and accurately reconstructing their geometry are fundamental to manipulation, passive visual observation in static scenes is inherently constrained by occlusions, which gives rise to both geometric and semantic ambiguities, particularly in complex multi-object settings. We present INGRID (INteractive GeometRy and instance IDentification), a system in which a robotic manipulator actively interacts with the scene to reveal previously hidden surfaces and disambiguate object instances and geometries. The key challenge lies first in deciding where to interact in order to induce the most informative scene changes. A subsequent challenge is then to efficiently update both the object instances and their geometries once new visual observations become available. To address these challenges, INGRID follows a four-step pipeline: (i) constructing an instance candidate tree, (ii) optimizing informative actions, (iii) identifying instances under rigid-body assumptions, and (iv) selectively finetuning geometry. Together, these steps enable efficient, interaction-driven geometry and instance identification. Our method proves robust and practical in complex multi-object scenes with severe occlusions, both in simulation and in the real world.

  • Behavior-Controllable Stable Dynamics Models on Riemannian Configuration Manifolds

    by Byeongho Lee, Yonghyeon Lee, Junsu Ha, Frank C. Park

    Due to their stability and robustness properties, Stable Dynamical Systems (SDS) have received considerable attention as a means of representing motions in learning from demonstration tasks. Designing vector fields that fit complex trajectories while ensuring stability still remains a key challenge; although recent deep learning-based methods have shown substantial progress in this direction, their tendency to overfit to demonstration trajectories often leads to undesirable behaviors, particularly as tasks deviate from demonstrations. At a fundamental level, the only reliable way to address this lack of generalization is to provide supervision in out-of-demonstration regions. Focusing on two types of general behaviors, mimicking and contracting, we propose a Behavior-Controllable Stable Dynamics Model (BCSDM), a one-parameter family of SDS that allows users to adjust the system's overall behavior depending on user intent. We show how to extend BCSDM to accommodate demonstrations of multiple tasks, and also propose a Deep Operator Vector Field (DeepOVec) for memory-efficient encoding of multiple dynamical systems. Extensive experiments on tasks that involve mimicking or contracting behaviors demonstrate the advantages of BCSDMs over existing state-of-the-art SDS learning methods.


  • Motion Manifold Flow Primitives for Task-Conditioned Trajectory Generation Under Complex Task-Motion Dependencies

    by Yonghyeon Lee, Byeongho Lee, Seungyeon Kim, Frank C. Park

    Effective movement primitives should be capable of encoding and generating a rich repertoire of trajectories conditioned on task-defining parameters such as vision or language inputs. While recent methods based on the motion manifold hypothesis, which assumes that a set of trajectories lies on a lower-dimensional nonlinear subspace, address challenges such as limited dataset size and the high dimensionality of trajectory data, they often struggle to capture complex task-motion dependencies, i.e., when motion distributions shift drastically with task variations. To address this, we introduce Motion Manifold Flow Primitives (MMFP), a framework that decouples the training of the motion manifold from task-conditioned distributions. Specifically, we employ flow matching models, state-of-the-art conditional deep generative models, to learn task-conditioned distributions in the latent coordinate space of the learned motion manifold. Experiments are conducted on language-guided trajectory generation tasks, where many-to-many text-motion correspondences introduce complex task-motion dependencies, highlighting MMFP's superiority over existing methods.


  • Open Access

    Diverse Policy Learning via Random Obstacle Deployment for Zero-Shot Adaptation

    by Seokjin Choi, Yonghyeon Lee, Seungyeon Kim, Che-Sang Park, Himchan Hwang, Frank C. Park

    In this letter, we propose a novel reinforcement learning framework that enables zero-shot policy adaptation in environments with unseen, dynamically changing obstacles. Adopting the idea that learning a policy capable of generating diverse actions is key to achieving such adaptability, our primary contribution is a novel learning algorithm that incorporates random obstacle deployment, enabling the policy to explore and learn diverse actions. This method overcomes the limitations of existing diverse policy learning approaches, which primarily rely on mutual information maximization to increase diversity. To enable zero-shot dynamic adaptation, our method further involves two key components: a state-dependent latent skill sampler and a motion predictor. We sample multiple skill variables at each state using a skill sampler, then filter out unsafe skills using a motion predictor, and consequently execute actions corresponding to safe skills. Compared to existing methods, our experiments demonstrate that proposed method generates significantly more diverse actions and adapts better to dynamically changing environments, making it highly effective for tasks with varying constraints such as moving obstacles.


  • Open Access

    MMP++: Motion Manifold Primitives With Parametric Curve Models

    by Yonghyeon Lee

    Motion Manifold Primitives (MMP), a manifold-based approach for encoding basic motion skills, can produce diverse trajectories, enabling the system to adapt to unseen constraints. Nonetheless, we argue that current MMP models lack crucial functionalities of movement primitives, such as temporal and via-points modulation, found in traditional approaches. This shortfall primarily stems from MMP's reliance on discrete-time trajectories. To overcome these limitations, we introduce Motion Manifold Primitives++ (MMP++), a new model that integrates the strengths of both MMP and traditional methods by incorporating parametric curve representations into the MMP framework. Furthermore, we identify a significant challenge with MMP++: performance degradation due to geometric distortions in the latent space, meaning that similar motions are not closely positioned. To address this, Isometric Motion Manifold Primitives++ (IMMP++) is proposed to ensure the latent space accurately preserves the manifold's geometry. Our experimental results across various applications, including 2-DoF planar motions, 7-DoF robot arm motions, and SE(3) trajectory planning, show that MMP++ and IMMP++ outperform existing methods in trajectory generation tasks, achieving substantial improvements in some cases. Moreover, they enable the modulation of latent coordinates and via-points, thereby allowing efficient online adaptation to dynamic environments.


  • NFL: Normal Field Learning for 6-DoF Grasping of Transparent Objects

    by Junho Lee, Sang Min Kim, Yonghyeon Lee, Young Min Kim

    We present Normal Field Learning (NFL), a robust yet practical solution to perceive 3D layouts of transparent objects and grasp them quickly. Conventional input modalities for vision-based grasping do not provide sufficient information for transparent objects. However, with the recent advance on datasets and algorithms for transparent objects, we can at least obtain noisy estimates of normals and masks for various real-world conditions. Instead of directly using the RGB images, we propose to use the estimates to train a neural volume, which serves as an intermediate representation ignorant of challenging appearance variations. We formulate the training objective to account for inherent uncertainty in individual estimation, and together with the volumetric aggregation, we can reliably extract useful geometric information for grasping. Our neural volume deploys a voxel-grid based representation, motivated by acceleration techniques of neural radiance fields. However, we directly store the normal and density values in the grid cells instead of latent features. Our modification allows direct access to the geometric values without additional inference or volume rendering, further enhancing the efficiency. Our results show over 85% success rates in grasping in cluttered scenes with only 40 seconds of training time.


  • Open Access

    A Geometric Perspective on Autoencoders

    by Yonghyeon Lee

    This paper presents the geometric aspect of the autoencoder framework, which, despite its importance, has been relatively less recognized. Given a set of high-dimensional data points that approximately lie on some lower-dimensional manifold, an autoencoder learns the \textit{manifold} and its \textit{coordinate chart}, simultaneously. This geometric perspective naturally raises inquiries like "Does a finite set of data points correspond to a single manifold?" or "Is there only one coordinate chart that can represent the manifold?". The responses to these questions are negative, implying that there are multiple solution autoencoders given a dataset. Consequently, they sometimes produce incorrect manifolds with severely distorted latent space representations. In this paper, we introduce recent geometric approaches that address these issues.