no code implementations • ICLR 2019 • Daeyoung Choi, Wonjong Rhee, Kyungeun Lee, Changho Shin
It has been common to argue or imply that a regularizer can be used to alter a statistical property of a hidden layer's representation and thus improve generalization or performance of deep networks.
1 code implementation • 27 Mar 2024 • Wonkyun Kim, Changin Choi, Wonseok Lee, Wonjong Rhee
Recently, an alternative strategy has surfaced, employing readily available foundation models, such as VideoLMs and LLMs, across multiple stages for modality bridging.
Ranked #1 on Zero-Shot Video Question Answer on IntentQA
no code implementations • 22 Mar 2024 • Jaeill Kim, Wonseok Lee, Moonjung Eo, Wonjong Rhee
Consequently, RFR achieves dual objectives in backward and forward compatibility: minimizing feature extractor modifications and enhancing novel task performance, respectively.
no code implementations • 22 Mar 2024 • Jimyeong Kim, Jungwon Park, Wonjong Rhee
In text-to-image personalization, a timely and crucial challenge is the tendency of generated images overfitting to the biases present in the reference images.
no code implementations • 21 Mar 2024 • Yeji Song, Jimyeong Kim, Wonhark Park, Wonsik Shin, Wonjong Rhee, Nojun Kwak
In a surge of text-to-image (T2I) models and their customization methods that generate new images of a user-provided subject, current works focus on alleviating the costs incurred by a lengthy per-subject optimization.
no code implementations • 17 Jan 2024 • Euna Jung, Jaekeol Choi, Eunggu Yun, Wonjong Rhee
Specifically, we consider \textit{On-Off pattern} and \textit{PathCount} for investigating how information is stored in deep representations.
no code implementations • 11 Jan 2024 • Jaeill Kim, Duhun Hwang, Eunjung Lee, Jangwon Suh, Jimyeong Kim, Wonjong Rhee
In the past few years, contrastive learning has played a central role for the success of visual unsupervised representation learning.
no code implementations • 21 Sep 2023 • Moonjung Eo, Suhyun Kang, Wonjong Rhee
In this study, we develop a \textit{Differentiable Framework~(DF)} that can express filter selection, rank selection, and budget constraint into a single analytical formulation.
no code implementations • 30 Aug 2023 • Kyungeun Lee, Jaeill Kim, Suhyun Kang, Wonjong Rhee
Contrastive learning has emerged as a cornerstone in recent achievements of unsupervised representation learning.
1 code implementation • CVPR 2023 • Jaeill Kim, Suhyun Kang, Duhun Hwang, Jungwook Shin, Wonjong Rhee
Since the introduction of deep learning, a wide scope of representation properties, such as decorrelation, whitening, disentanglement, rank, isotropy, and mutual information, have been studied to improve the quality of representation.
Ranked #15 on Domain Generalization on TerraIncognita
1 code implementation • CVPR 2023 • Suhyun Kang, Duhun Hwang, Moonjung Eo, Taesup Kim, Wonjong Rhee
In this study, we propose Geometry-Adaptive Preconditioned gradient descent (GAP) that can overcome the limitations in MAML; GAP can efficiently meta-learn a preconditioner that is dependent on task-specific parameters, and its preconditioner can be shown to be a Riemannian metric.
1 code implementation • 20 Mar 2023 • Jungwook Shin, Jaeill Kim, Kyungeun Lee, Hyunghun Cho, Wonjong Rhee
To improve the diversity of the whole-body object construction, we develop an iterative method that stochastically combines multiple objects observed from the real world into a single object.
1 code implementation • 23 Nov 2022 • Jangwon Suh, Jimyeong Kim, Euna Jung, Wonjong Rhee
The performance of cardiac arrhythmia detection with electrocardiograms(ECGs) has been considerably improved since the introduction of deep learning models.
1 code implementation • 1 Sep 2022 • Euna Jung, Jungwon Park, Jaekeol Choi, Sungyoon Kim, Wonjong Rhee
In particular, many of the high-performing dense retrieval models evaluate representations of query and document using BERT, and subsequently apply a cosine-similarity based scoring to determine the relevance.
no code implementations • 24 Feb 2022 • Jaekeol Choi, Euna Jung, Sungjun Lim, Wonjong Rhee
The traditional approach, however, is being rapidly replaced by Neural Ranking Models (NRMs) that can exploit semantic features.
1 code implementation • 7 Feb 2022 • Hyunghun Cho, Jungwook Shin, Wonjong Rhee
The early pioneering Neural Architecture Search (NAS) works were multi-trial methods applicable to any general search space.
no code implementations • 30 Nov 2021 • Moonjung Eo, Suhyun Kang, Wonjong Rhee
The resulting BSR (Beam-search and Stable Rank) algorithm requires only a single hyperparameter to be tuned for the desired compression ratio.
1 code implementation • 28 Oct 2021 • Euna Jung, Jaekeol Choi, Wonjong Rhee
The results confirm that both lightweight fine-tuning and semi-Siamese are considerably helpful for improving BERT-based bi-encoders.
no code implementations • 29 Sep 2021 • Jaeill Kim, Wonjong Rhee
In this work, we derive a principled non-contrastive method where mutual information is estimated as a difference of entropies and thus no need for negative sampling.
no code implementations • ICML Workshop AML 2021 • Duhun Hwang, Eunjung Lee, Wonjong Rhee
AID-purifier is an auxiliary network that works as an add-on to an already trained main classifier.
1 code implementation • 11 Mar 2021 • Jaekeol Choi, Euna Jung, Jangwon Suh, Wonjong Rhee
When monoBERT is used as the cross-encoder teacher, together with either TwinBERT or ColBERT as the bi-encoder teacher, TRMD produces a student bi-encoder that performs better than the corresponding baseline bi-encoder.
no code implementations • 28 Aug 2020 • Kyungeun Lee, Moonjung Eo, Euna Jung, Yoonjin Yoon, Wonjong Rhee
2) We briefly explain a wide range of DNN techniques from the earliest networks, including Restricted Boltzmann Machines, to the most recent, including graph-based and meta-learning networks.
no code implementations • 12 May 2020 • Jaekeol Choi, Jungin Choi, Wonjong Rhee
However, explaining the ranking results has become even more difficult with NRM due to the complex structure of neural networks.
2 code implementations • 29 May 2019 • Kyungeun Lee, Wonjong Rhee
In this paper, we identify two essential spatial dependencies in traffic forecasting in addition to distance, direction and positional relationship, for designing basic graph elements as the fundamental building blocks.
1 code implementation • NeurIPS19 under review 2019 • Hyunghun Cho, Yongjin Kim, Eunjung Lee, Daeyoung Choi, YongJae lee, Wonjong Rhee
The performance of deep neural networks (DNN) is very sensitive to the particular choice of hyper-parameters.
no code implementations • 16 Nov 2018 • Changho Shin, Sunghwan Joo, Jaeryun Yim, Hyoseop Lee, Taesup Moon, Wonjong Rhee
In this work, we focus on the idea that appliances have on/off states, and develop a deep network for further performance improvements.
no code implementations • 8 Nov 2018 • Daeyoung Choi, Kyungeun Lee, Duhun Hwang, Wonjong Rhee
In this study, the effects of eight representation regularization methods are investigated, including two newly developed rank regularizers (RR).
1 code implementation • 25 Sep 2018 • Daeyoung Choi, Wonjong Rhee
Motivated by the idea, we design two class-wise regularizers that explicitly utilize class information: class-wise Covariance Regularizer (cw-CR) and class-wise Variance Regularizer (cw-VR).
1 code implementation • 4 Jul 2018 • Wonkyung Jung, Daejin Jung, and Byeongho Kim, Sunjung Lee, Wonjong Rhee, Jung Ho Ahn
Batch Normalization (BN) has become a core design block of modern Convolutional Neural Networks (CNNs).
no code implementations • ICLR 2018 • Daeyoung Choi, Changho Shin, Hyunghun Cho, Wonjong Rhee
Performance of Deep Neural Network (DNN) heavily depends on the characteristics of hidden layer representations.