16 papers
Multi-View Relational Distillation for Spatial Reasoning with Vision-Language Models
Kiet T. Nguyen, Hanbo Shim, Jinwoo Kim +1
Vision-language models (VLMs) have achieved strong image and video understanding, yet their visual-spatial representations remain geometrically fragile, leading to failures in spat…
Training-Free Refinement of Flow Matching with Divergence-based Sampling
Yeonwoo Cha, Jaehoon Yoo, Semin Kim +3
Flow-based models learn a target distribution by modeling a marginal velocity field, defined as the average of sample-wise velocities connecting each sample from a simple prior to…
When Sinks Help or Hurt: Unified Framework for Attention Sink in Large Vision-Language Models
Jiho Choi, Jaemin Kim, Sanghwan Kim +2
Attention sinks are defined as tokens that attract disproportionate attention. While these have been studied in single modality transformers, their cross-modal impact in Large Visi…
Inverting Data Transformations via Diffusion Sampling
Jinwoo Kim, Sékou-Oumar Kaba, Jiyun Park +2
We study the problem of transformation inversion on general Lie groups: a datum is transformed by an unknown group element, and the goal is to recover an inverse transformation tha…
FlowBind: Efficient Any-to-Any Generation with Bidirectional Flows
Yeonwoo Cha, Semin Kim, Jinhyeon Kwon +1
Any-to-any generation seeks to translate between arbitrary subsets of modalities, enabling flexible cross-modal synthesis. Despite recent success, existing flow-based approaches ar…
Disentangled Representation Learning via Modular Compositional Bias
Whie Jung, Dong Hoon Lee, Seunghoon Hong
Recent disentangled representation learning (DRL) methods heavily rely on factor specific strategies-either learning objectives for attributes or model architectures for objects-to…
Bridging the gap to real-world language-grounded visual concept learning
Whie Jung, Semin Kim, Junee Kim +1
Human intelligence effortlessly interprets visual scenes along a rich spectrum of semantic dimensions. However, existing approaches to language-grounded visual concept learning are…
Sequence Modeling with Spectral Mean Flows
Jinwoo Kim, Max Beier, Petar Bevanda +2
A key question in sequence modeling with neural networks is how to represent and learn highly nonlinear and probabilistic state dynamics. Operator theory views such dynamics as lin…
Flock: A Knowledge Graph Foundation Model via Learning on Random Walks
Jinwoo Kim, Xingyue Huang, Krzysztof Olejniczak +4
We study the problem of zero-shot link prediction on knowledge graphs (KGs), which requires models to generalize to novel entities and novel relations. Knowledge graph foundation m…
ReDi: Rectified Discrete Flow
Jaehoon Yoo, Wonjung Kim, Seunghoon Hong
Discrete Flow-based Models (DFMs) are powerful generative models for high-quality discrete data but typically suffer from slow sampling speeds due to their reliance on iterative de…
Reward-Agnostic Prompt Optimization for Text-to-Image Diffusion Models
Semin Kim, Yeonwoo Cha, Jaehoon Yoo +1
We investigate a general approach for improving user prompts in text-to-image (T2I) diffusion models by finding prompts that maximize a reward function specified at test-time. Alth…
Simulation-Free Training of Neural ODEs on Paired Data
Semin Kim, Jaehoon Yoo, Jinwoo Kim +3
In this work, we investigate a method for simulation-free training of Neural Ordinary Differential Equations (NODEs) for learning deterministic mappings between paired data. Despit…
Revisiting Random Walks for Learning on Graphs
Jinwoo Kim, Olga Zaghen, Ayhan Suleymanzade +2
We revisit a simple model class for machine learning on graphs, where a random walk on a graph produces a machine-readable record, and this record is processed by a deep neural net…
Learning to Compose: Improving Object Centric Learning by Injecting Compositionality
Whie Jung, Jaehoon Yoo, Sungjin Ahn +1
Learning compositional representation is a key aspect of object-centric learning as it enables flexible systematic generalization and supports complex visual reasoning. However, mo…
Chameleon: A Data-Efficient Generalist for Dense Visual Prediction in the Wild
Donggyun Kim, Seongwoong Cho, Semin Kim +2
Large language models have evolved data-efficient generalists, benefiting from the universal language interface and large-scale pre-training. However, constructing a data-efficient…
Learning Symmetrization for Equivariance with Orbit Distance Minimization
Tien Dat Nguyen, Jinwoo Kim, Hongseok Yang +1
We present a general framework for symmetrizing an arbitrary neural-network architecture and making it equivariant with respect to a given group. We build upon the proposals of Kim…