Seungjun Lee

TCS Journal 2026 Journal Article

Largest similar copies of convex polygons in polygonal domains

• Taekang Eom
• Seungjun Lee
• Hee-Kap Ahn

NeurIPS Conference 2025 Conference Paper

Dynam3D: Dynamic Layered 3D Tokens Empower VLM for Vision-and-Language Navigation

• Zihan Wang
• Seungjun Lee
• Gim Hee Lee

Vision-and-Language Navigation (VLN) is a core task where embodied agents leverage their spatial mobility to navigate in 3D environments toward designated destinations based on natural language instructions. Recently, video-language large models (Video-VLMs) with strong generalization capabilities and rich commonsense knowledge have shown remarkable performance when applied to VLN tasks. However, these models still encounter the following challenges when applied to real-world 3D navigation: 1) Insufficient understanding of 3D geometry and spatial semantics; 2) Limited capacity for large-scale exploration and long-term environmental memory; 3) Poor adaptability to dynamic and changing environments. To address these limitations, we propose Dynam3D, a dynamic layered 3D representation model that leverages language-aligned, generalizable, and hierarchical 3D representations as visual input to train 3D-VLM in navigation action prediction. Given posed RGB-D images, our Dynam3D projects 2D CLIP features into 3D space and constructs multi-level 3D patch-instance-zone representations for 3D geometric and semantic understanding with a dynamic and layer-wise update strategy. Our Dynam3D is capable of online encoding and localization of 3D instances, and dynamically updates them in changing environments to provide large-scale exploration and long-term memory capabilities for navigation. By leveraging large-scale 3D-language pretraining and task-specific adaptation, our Dynam3D sets new state-of-the-art performance on VLN benchmarks including R2R-CE, REVERIE-CE and NavRAG-CE under monocular settings. Furthermore, experiments for pre-exploration, lifelong memory, and real-world robot validate the effectiveness of practical deployment.

AAAI Conference 2025 Conference Paper

ImagePiece: Content-aware Re-tokenization for Efficient Image Recognition

• Seungdong Yoa
• Seungjun Lee
• Hye-Seung Cho
• Bumsoo Kim
• Woohyung Lim

Vision Transformers (ViTs) have achieved remarkable success in various computer vision tasks. However, ViTs have a huge computational cost due to their inherent reliance on multi-head self-attention (MHSA), prompting efforts to accelerate ViTs for practical applications. To this end, recent works aim to reduce the number of tokens, mainly focusing on how to effectively prune or merge them. Nevertheless, since ViT tokens are generated from non-overlapping grid patches, they usually do not convey sufficient semantics, making it incompatible with efficient ViTs. To address this, we propose ImagePiece, a novel re-tokenization strategy for Vision Transformers. Following the MaxMatch strategy of NLP tokenization, ImagePiece groups semantically insufficient yet locally coherent tokens until they convey meaning. This simple retokenization is highly compatible with previous token reduction methods, being able to drastically narrow down relevant tokens, enhancing the inference speed of DeiT-S by 54% (nearly 1.5x faster) while achieving a 0.39% improvement in ImageNet classification accuracy. For hyper-speed inference scenarios (with 251% acceleration), our approach surpasses other baselines by an accuracy over 8%.

NeurIPS Conference 2025 Conference Paper

PDEfuncta: Spectrally-Aware Neural Representation for PDE Solution Modeling

• Minju Jo
• Woojin Cho
• Uvini Balasuriya Mudiyanselage
• Seungjun Lee
• Noseong Park
• Kookjin Lee

Scientific machine learning often involves representing complex solution fields that exhibit high-frequency features such as sharp transitions, fine-scale oscillations, and localized structures. While implicit neural representations (INRs) have shown promise for continuous function modeling, capturing such high-frequency behavior remains a challenge—especially when modeling multiple solution fields with a shared network. Prior work addressing spectral bias in INRs has primarily focused on single-instance settings, limiting scalability and generalization. In this work, we propose Global Fourier Modulation (GFM), a novel modulation technique that injects high-frequency information at each layer of the INR through Fourier-based reparameterization. This enables compact and accurate representation of multiple solution fields using low-dimensional latent vectors. Building upon GFM, we introduce PDEfuncta, a meta-learning framework designed to learn multi-modal solution fields and support generalization to new tasks. Through empirical studies on diverse scientific problems, we demonstrate that our method not only improves representational quality but also shows potential for forward and inverse inference tasks without the need for retraining.

ICML Conference 2025 Conference Paper

SCENT: Robust Spatiotemporal Learning for Continuous Scientific Data via Scalable Conditioned Neural Fields

• David Keetae Park
• Xihaier Luo
• Guang Zhao
• Seungjun Lee
• Miruna Oprescu
• Shinjae Yoo

Spatiotemporal learning is challenging due to the intricate interplay between spatial and temporal dependencies, the high dimensionality of the data, and scalability constraints. These challenges are further amplified in scientific domains, where data is often irregularly distributed (e. g. , missing values from sensor failures) and high-volume (e. g. , high-fidelity simulations), posing additional computational and modeling difficulties. In this paper, we present SCENT, a novel framework for scalable and continuity-informed spatiotemporal representation learning. SCENT unifies interpolation, reconstruction, and forecasting within a single architecture. Built on a transformer-based encoder-processor-decoder backbone, SCENT introduces learnable queries to enhance generalization and a query-wise cross-attention mechanism to effectively capture multi-scale dependencies. To ensure scalability in both data size and model complexity, we incorporate a sparse attention mechanism, enabling flexible output representations and efficient evaluation at arbitrary resolutions. We validate SCENT through extensive simulations and real-world experiments, demonstrating state-of-the-art performance across multiple challenging tasks while achieving superior scalability.

ICLR Conference 2025 Conference Paper

Segment Any 3D Object with Language

• Seungjun Lee
• Yuyang Zhao
• Gim Hee Lee

In this paper, we investigate Open-Vocabulary 3D Instance Segmentation (OV-3DIS) with free-form language instructions. Earlier works mainly rely on annotated base categories for training which leads to limited generalization to unseen novel categories. To mitigate the poor generalizability to novel categories, recent works generate class-agnostic masks or projecting generalized masks from 2D to 3D, subsequently classifying them with the assistance of 2D foundation model. However, these works often disregard semantic information in the mask generation, leading to sub-optimal performance. Instead, generating generalizable but semantic-aware masks directly from 3D point clouds would result in superior outcomes. To the end, we introduce Segment any 3D Object with LanguagE ($\textbf{SOLE}$), which is a semantic and geometric-aware visual-language learning framework with strong generalizability by generating semantic-related masks directly from 3D point clouds. Specifically, we propose a multimodal fusion network to incorporate multimodal semantics in both backbone and decoder. In addition, to align the 3D segmentation model with various language instructions and enhance the mask quality, we introduce three types of multimodal associations as supervision. Our SOLE outperforms previous methods by a large margin on ScanNetv2, ScanNet200, and Replica benchmarks, and the results are even closed to the fully-supervised counterpart despite the absence of class annotations in the training. Furthermore, extensive qualitative results demonstrate the versatility of our SOLE to language instructions. The code will be made publicly available.

AAAI Conference 2024 Conference Paper

Inducing Point Operator Transformer: A Flexible and Scalable Architecture for Solving PDEs

• Seungjun Lee
• TaeiL Oh

Solving partial differential equations (PDEs) by learning the solution operators has emerged as an attractive alternative to traditional numerical methods. However, implementing such architectures presents two main challenges: flexibility in handling irregular and arbitrary input and output formats and scalability to large discretizations. Most existing architectures are limited by their desired structure or infeasible to scale large inputs and outputs. To address these issues, we introduce an attention-based model called an inducing point operator transformer (IPOT). Inspired by inducing points methods, IPOT is designed to handle any input function and output query while capturing global interactions in a computationally efficient way. By detaching the inputs/outputs discretizations from the processor with a smaller latent bottleneck, IPOT offers flexibility in processing arbitrary discretizations and scales linearly with the size of inputs/outputs. Our experimental results demonstrate that IPOT achieves strong performances with manageable computational complexity on an extensive range of PDE benchmarks and real-world weather forecasting scenarios, compared to state-of-the-art methods. Our code is publicly available at https://github.com/7tl7qns7ch/IPOT.

IROS Conference 2022 Conference Paper

Development of a cable-driven Growing Sling to assist patient transfer

• MyungJoong Lee
• Yonghwan Moon
• Jeongryul Kim
• Seungjun Lee
• Keri Kim
• Hyunki In

As the aging of society continues to accelerate, the number of elderly patients is increasing, as is the demand for manpower to care for them. In particular, there is an urgent need for bedridden patient care. However, limitations in the supply of human resources have caused an increase in the burden for care. In particular, nursing personnel often experience inconvenience and difficulties owing to the great deal of effort required to transfer a patient from bed to wheelchair, or vice versa. The most difficult process during the patient transfer is inserting the sling under the patient. Aiming to solve this problem, a mechanical Growing Sling was devised. The proposed sling adapts a growing mechanism comprising a low-friction fabric and steel shafts, and the sling is inserted under the patient by towing the steel shafts with cables connected to a motor. For the comfort and safety of the sling insertion, the required towing force was analyzed to find the minimum diameter of the shaft. The results from experimental evaluations using the proposed sling verified that it can be inserted under the patient without moving the patient, and with an acceptable level of pressure being applied to the patient.

ICLR Conference 2021 Conference Paper

Identifying Physical Law of Hamiltonian Systems via Meta-Learning

• Seungjun Lee
• Haesang Yang
• Woojae Seong

Hamiltonian mechanics is an effective tool to represent many physical processes with concise yet well-generalized mathematical expressions. A well-modeled Hamiltonian makes it easy for researchers to analyze and forecast many related phenomena that are governed by the same physical law. However, in general, identifying a functional or shared expression of the Hamiltonian is very difficult. It requires carefully designed experiments and the researcher's insight that comes from years of experience. We propose that meta-learning algorithms can be potentially powerful data-driven tools for identifying the physical law governing Hamiltonian systems without any mathematical assumptions on the representation, but with observations from a set of systems governed by the same physical law. We show that a well meta-trained learner can identify the shared representation of the Hamiltonian by evaluating our method on several types of physical systems with various experimental settings.

IROS Conference 2020 Conference Paper

Development of a pneumatically-driven Growing Sling to assist patient transfer

• Jonggyu Choi
• Seungjun Lee
• Jeongryul Kim
• MyungJoong Lee
• Keri Kim
• Hyunki In

In this study, a new type of sling for assisting bedridden patients is developed using a pneumatic growing mechanism. Growing Sling focuses on minimizing the labor input of the caregivers by automating the sling insertion and retraction process while maintaining safety and comfort. Improvements over the typical growing mechanism were made by reinforcing the sling with shafts and filament tape for restricting the height of the sling to ensure its design purpose. Analysis of forces exerted on the structure was made to interpret the driving power of the automated insertion process and to ensure the structural integrity of components. Experiments on materials and prototype devices were conducted to determine the quantitative load that the sling needs to endure and what type of material is suitable for fabrication. Further, we propose a fabrication process for the Growing Sling, including its dimensions, and validate the performance of the fabricated prototype.