Searcharxiv⌕ Search

arXiv subjects

Junho Lee

Publications and source records attributed to Junho Lee.

At least 37 records · Page 2Linked to original sources
arXiv2024

Finding NeMo: Negative-mined Mosaic Augmentation for Referring Image Segmentation

Referring Image Segmentation is a comprehensive task to segment an object referred by a textual query from an image. In nature, the level of difficulty in this task is affected by the existence of similar objects and the complexity of the referring expression. Recent RIS models still show a significant performance gap between easy and hard scenarios. We pose that the bottleneck exists in the data, and propose a simple but powerful data augmentation method, Negative-mined Mosaic Augmentation (NeMo). This method augments a training image into a mosaic with three other negative images carefully curated by a pretrained multimodal alignment model, e.g., CLIP, to make the sample more challenging. We discover that it is critical to properly adjust the difficulty level, neither too ambiguous nor too trivial. The augmented training data encourages the RIS model to recognize subtle differences and relationships between similar visual entities and to concretely understand the whole expression to locate the right target better. Our approach shows consistent improvements on various datasets and models, verified by extensive experiments.

cs.CV↗
arXiv2024

Isometric Representation Learning for Disentangled Latent Space of Diffusion Models

The latent space of diffusion model mostly still remains unexplored, despite its great success and potential in the field of generative modeling. In fact, the latent space of existing diffusion models are entangled, with a distorted mapping from its latent space to image space. To tackle this problem, we present Isometric Diffusion, equipping a diffusion model with a geometric regularizer to guide the model to learn a geometrically sound latent space of the training data manifold. This approach allows diffusion models to learn a more disentangled latent space, which enables smoother interpolation, more accurate inversion, and more precise control over attributes directly in the latent space. Our extensive experiments consisting of image interpolations, image inversions, and linear editing show the effectiveness of our method.

cs.LG↗
arXiv2024

Semi-Supervised Domain Adaptation Using Target-Oriented Domain Augmentation for 3D Object Detection

3D object detection is crucial for applications like autonomous driving and robotics. However, in real-world environments, variations in sensor data distribution due to sensor upgrades, weather changes, and geographic differences can adversely affect detection performance. Semi-Supervised Domain Adaptation (SSDA) aims to mitigate these challenges by transferring knowledge from a source domain, abundant in labeled data, to a target domain where labels are scarce. This paper presents a new SSDA method referred to as Target-Oriented Domain Augmentation (TODA) specifically tailored for LiDAR-based 3D object detection. TODA efficiently utilizes all available data, including labeled data in the source domain, and both labeled data and unlabeled data in the target domain to enhance domain adaptation performance. TODA consists of two stages: TargetMix and AdvMix. TargetMix employs mixing augmentation accounting for LiDAR sensor characteristics to facilitate feature alignment between the source-domain and target-domain. AdvMix applies point-wise adversarial augmentation with mixing augmentation, which perturbs the unlabeled data to align the features within both labeled and unlabeled data in the target domain. Our experiments conducted on the challenging domain adaptation tasks demonstrate that TODA outperforms existing domain adaptation techniques designed for 3D object detection by significant margins. The code is available at: https://github.com/rasd3/TODA.

cs.CV↗
arXiv2024

Non-freeness of parabolic two-generator groups

A complex number $λ$ is said to be non-free if the subgroup of $SL(2,\bc)$ generated by $$X=\begin{pmatrix} 1& 1\\ 0 & 1 \end{pmatrix} \,\, \text{and}\,\,\,Y_λ=\begin{pmatrix} 1& 0\\ λ& 1 \end{pmatrix}$$ is not a free group of rank 2. In this case the number $λ$ is called a relation number, and it has been a long standing problem to determine the relation numbers. In this paper, we characterize the relation numbers by establishing the equivalence between $λ$ being a relation number and $u:=\sqrt{- λ}$ being a root of a `generalized Chebyshev polynomial'. The generalized Chebyshev polynomials of degree $k$ are given by a sequence of $k$ integers $(n_1, n_2,\cdots, n_k)$ using the usual recursive formula, and thereby can be studied systematically using continuants and continued fractions. Such formulation, then, enables us to prove that, the question whether a given number $λ$ is a relation number of $u$-degree $k$ can be answered by checking only finitely many generalized Chebyshev polynomials. Based on these theorems, we design an algorithm deciding any given number is a relation number with minimal degree $k$. With its computer implementation we provide a few sample examples, with a particular emphasis on the well known conjecture that every rational number in the interval $(-4, 4)$ is a relation number.

math.GR↗
arXiv2024

Enhancing Effectiveness and Robustness in a Low-Resource Regime via Decision-Boundary-aware Data Augmentation

Efforts to leverage deep learning models in low-resource regimes have led to numerous augmentation studies. However, the direct application of methods such as mixup and cutout to text data, is limited due to their discrete characteristics. While methods using pretrained language models have exhibited efficiency, they require additional considerations for robustness. Inspired by recent studies on decision boundaries, this paper proposes a decision-boundary-aware data augmentation strategy to enhance robustness using pretrained language models. The proposed technique first focuses on shifting the latent features closer to the decision boundary, followed by reconstruction to generate an ambiguous version with a soft label. Additionally, mid-K sampling is suggested to enhance the diversity of the generated sentences. This paper demonstrates the performance of the proposed augmentation strategy compared to other methods through extensive experiments. Furthermore, the ablation study reveals the effect of soft labels and mid-K sampling and the extensibility of the method with curriculum data augmentation.

cs.CL↗
arXiv2024

AutoAugment Is What You Need: Enhancing Rule-based Augmentation Methods in Low-resource Regimes

Text data augmentation is a complex problem due to the discrete nature of sentences. Although rule-based augmentation methods are widely adopted in real-world applications because of their simplicity, they suffer from potential semantic damage. Previous researchers have suggested easy data augmentation with soft labels (softEDA), employing label smoothing to mitigate this problem. However, finding the best factor for each model and dataset is challenging; therefore, using softEDA in real-world applications is still difficult. In this paper, we propose adapting AutoAugment to solve this problem. The experimental results suggest that the proposed method can boost existing augmentation methods and that rule-based methods can enhance cutting-edge pre-trained language models. We offer the source code.

cs.CL↗
arXiv2024

SoftEDA: Rethinking Rule-Based Data Augmentation with Soft Labels

Rule-based text data augmentation is widely used for NLP tasks due to its simplicity. However, this method can potentially damage the original meaning of the text, ultimately hurting the performance of the model. To overcome this limitation, we propose a straightforward technique for applying soft labels to augmented data. We conducted experiments across seven different classification tasks and empirically demonstrated the effectiveness of our proposed approach. We have publicly opened our source code for reproducibility.

cs.CL↗
arXiv2022

Do Not Escape From the Manifold: Discovering the Local Coordinates on the Latent Space of GANs

The discovery of the disentanglement properties of the latent space in GANs motivated a lot of research to find the semantically meaningful directions on it. In this paper, we suggest that the disentanglement property is closely related to the geometry of the latent space. In this regard, we propose an unsupervised method for finding the semantic-factorizing directions on the intermediate latent space of GANs based on the local geometry. Intuitively, our proposed method, called Local Basis, finds the principal variation of the latent space in the neighborhood of the base latent variable. Experimental results show that the local principal variation corresponds to the semantic factorization and traversing along it provides strong robustness to image traversal. Moreover, we suggest an explanation for the limited success in finding the global traversal directions in the latent space, especially W-space of StyleGAN2. We show that W-space is warped globally by comparing the local geometry, discovered from Local Basis, through the metric on Grassmannian Manifold. The global warpage implies that the latent space is not well-aligned globally and therefore the global traversal directions are bound to show limited success on it.

cs.CV↗
arXiv2022

User Guide for KOTE: Korean Online Comments Emotions Dataset

Sentiment analysis that classifies data into positive or negative has been dominantly used to recognize emotional aspects of texts, despite the deficit of thorough examination of emotional meanings. Recently, corpora labeled with more than just valence are built to exceed this limit. However, most Korean emotion corpora are small in the number of instances and cover a limited range of emotions. We introduce KOTE dataset. KOTE contains 50k (250k cases) Korean online comments, each of which is manually labeled for 43 emotion labels or one special label (NO EMOTION) by crowdsourcing (Ps = 3,048). The emotion taxonomy of the 43 emotions is systematically established by cluster analysis of Korean emotion concepts expressed on word embedding space. After explaining how KOTE is developed, we also discuss the results of finetuning and analysis for social discrimination in the corpus.

cs.CL↗
arXiv2022

MCW-Net: Single Image Deraining with Multi-level Connections and Wide Regional Non-local Blocks

A recent line of convolutional neural network-based works has succeeded in capturing rain streaks. However, difficulties in detailed recovery still remain. In this paper, we present a multi-level connection and wide regional non-local block network (MCW-Net) to properly restore the original background textures in rainy images. Unlike existing encoder-decoder-based image deraining models that improve performance with additional branches, MCW-Net improves performance by maximizing information utilization without additional branches through the following two proposed methods. The first method is a multi-level connection that repeatedly connects multi-level features of the encoder network to the decoder network. Multi-level connection encourages the decoding process to use the feature information of all levels. In multi-level connection, channel-wise attention is considered to learn which level of features is important in the decoding process of the current level. The second method is a wide regional non-local block. As rain streaks primarily exhibit a vertical distribution, we divide the grid of the image into horizontally-wide patches and apply a non-local operation to each region to explore the rich rain-free background information. Experimental results on both synthetic and real-world rainy datasets demonstrate that the proposed model significantly outperforms existing state-of-the-art models. Furthermore, the results of the joint deraining and segmentation experiment prove that our model contributes effectively to other vision tasks.

cs.CV↗
arXiv2021

GATSBI: Generative Agent-centric Spatio-temporal Object Interaction

We present GATSBI, a generative model that can transform a sequence of raw observations into a structured latent representation that fully captures the spatio-temporal context of the agent's actions. In vision-based decision-making scenarios, an agent faces complex high-dimensional observations where multiple entities interact with each other. The agent requires a good scene representation of the visual observation that discerns essential components and consistently propagates along the time horizon. Our method, GATSBI, utilizes unsupervised object-centric scene representation learning to separate an active agent, static background, and passive objects. GATSBI then models the interactions reflecting the causal relationships among decomposed entities and predicts physically plausible future states. Our model generalizes to a variety of environments where different types of robots and objects dynamically interact with each other. We show GATSBI achieves superior performance on scene decomposition and video prediction compared to its state-of-the-art counterparts.

cs.CV↗
arXiv2019

Polarization fraction measurement in ZZ scattering using deep learning

Measuring longitudinally polarized vector boson scattering in the ZZ channel is a promising way to investigate unitarity restoration with the Higgs mechanism and to search for possible new physics. We investigated several deep neural network structures and compared their ability to improve the measurement of the longitudinal fraction Z_L Z_L. Using fast simulation with the Delphes framework, a clear improvement is found using a previously investigated 'particle-based' deep neural network on a preprocessed dataset and applying principle component analysis to the outputs.A significance of around 1.7 standard deviations can be achieved with the integrated luminosity of 3000 fb-1 that will be recorded at the High-Luminosity LHC.

hep-ph↗
arXiv2019

A square root of Hurwitz numbers

We exhibit a generating function of spin Hurwitz numbers analogous to (disconnected) double Hurwitz numbers that is a tau function of the two-component BKP (2-BKP) hierarchy and is a square root of a tau function of the two-component KP (2-KP) hierarchy defined by related Hurwitz numbers.

math.AG↗
arXiv2019

Boosting CNN beyond Label in Inverse Problems

Convolutional neural networks (CNN) have been extensively used for inverse problems. However, their prediction error for unseen test data is difficult to estimate a priori since the neural networks are trained using only selected data and their architecture are largely considered a blackbox. This poses a fundamental challenge to neural networks for unsupervised learning or improvement beyond the label. In this paper, we show that the recent unsupervised learning methods such as Noise2Noise, Stein's unbiased risk estimator (SURE)-based denoiser, and Noise2Void are closely related to each other in their formulation of an unbiased estimator of the prediction error, but each of them are associated with its own limitations. Based on these observations, we provide a novel boosting estimator for the prediction error. In particular, by employing combinatorial convolutional frame representation of encoder-decoder CNN and synergistically combining it with the batch normalization, we provide a close form formulation for the unbiased estimator of the prediction error that can be minimized for neural network training beyond the label. Experimental results show that the resulting algorithm, what we call Noise2Boosting, provides consistent improvement in various inverse problems under both supervised and unsupervised learning setting.

cs.CV↗
arXiv2019

Polarization fraction measurement in same-sign WW scattering using deep learning

Studying the longitudinally polarized fraction of $W^\pm W^\pm$ scattering at the LHC is crucial to examine the unitarization mechanism of the vector boson scattering amplitude through Higgs and possible new physics. We apply here for the first time a Deep Neural Network classification to extract the longitudinal fraction. Based on fast simulation implemented with the Delphes framework, significant improvement from a deep neural network is found to be achievable and robust over all dijet mass region. A conservative estimation shows that a high significance of four standard deviations can be reached with the High-Luminosity LHC designed luminosity of 3000 $fb^{-1}$

hep-ph↗
arXiv2018

A note on Gunningham's formula

Gunningham [G] constructed an extended topological quantum field theory (TQFT) to obtain a closed formula for all spin Hurwitz numbers. In this note, we use the gluing theorem in [LP2] to re-prove the Gunningham's formula. We also describe a TQFT formalism naturally induced from the gluing theorem.

math.SG↗
arXiv2017

Microstructure and elevated-temperature mechanical properties of refractory AlMo0.5NbTa0.5TiZr High Entropy Alloy fabricated by powder metallurgy

New approaches for the design of alloy systems with multiprincipal elements is recently researched in refractory materials field. However, most research aimed at arc melting process with weakness of coarsening of grains and inhomogeneous microstructure of segregation of elements during the cooling. This study aims to design and fabricate high-entropy alloy with powder metallurgy. In this study, a refractory high entropy alloys with composition near AlMo0.5NbTa0.5TiZr were produced by powder metallurgy. The alloy consists of two body-centered cubic (BCC) phases. One phase was disordered BCC enriched with Mo, Nb and Ta and the other phase was ordered BCC enriched with Al and Zr. The AlMo0.5NbTa0.5TiZr alloy had a density of 7.46g/cm3 and Vickers microhardness of 678HV. Its compressive yield strength was 2466MPa at 298K and 964MPa at 1273K. The properties of the alloy and the beneficial effects from powder metallurgy on the microstructure and properties were outlined.

cond-mat.mtrl-sci↗