Dongyoung Kim

Dong-Young Kim, Dong-Kyun Han, Hye-Bin Shin

Drowsiness reduces concentration and increases response time, which causes fatal road accidents. Monitoring drivers' drowsiness levels by electroencephalogram (EEG) and taking action may prevent road accidents. EEG signals effectively monitor the driver's mental state as they can monitor brain dynamics. However, calibration is required in advance because EEG signals vary between and within subjects. Because of the inconvenience, calibration has reduced the accessibility of the brain-computer interface (BCI). Developing a generalized classification model is similar to domain generalization, which overcomes the domain shift problem. Especially data augmentation is frequently used. This paper proposes a calibration-free framework for driver drowsiness state classification using manifold-level augmentation. This framework increases the diversity of source domains by utilizing features. We experimented with various augmentation methods to improve the generalization performance. Based on the results of the experiments, we found that deeper models with smaller kernel sizes improved generalizability. In addition, applying an augmentation at the manifold-level resulted in an outstanding improvement. The framework demonstrated the capability for calibration-free BCI.

Jaehyung Kim, Dongyoung Kim, Yiming Yang

An open challenge in recent machine learning is about how to improve the reasoning capability of large language models (LLMs) in a black-box setting, i.e., without access to detailed information such as output token probabilities. Existing approaches either rely on accessibility (which is often unrealistic) or involve significantly increased train- and inference-time costs. This paper addresses those limitations or shortcomings by proposing a novel approach, namely CoBB (Correct for improving QA reasoning of Black-Box LLMs). It uses a trained adaptation model to perform a seq2seq mapping from the often-imperfect reasonings of the original black-box LLM to the correct or improved reasonings. Specifically, the adaptation model is initialized with a relatively small open-source LLM and adapted over a collection of sub-sampled training pairs. To select the representative pairs of correct and incorrect reasonings, we formulated the dataset construction as an optimization problem that minimizes the statistical divergence between the sampled subset and the entire collection, and solved it via a genetic algorithm. We then train the adaptation model over the sampled pairs by contrasting the likelihoods of correct and incorrect reasonings. Our experimental results demonstrate that CoBB significantly improves reasoning accuracy across various QA benchmarks, compared to the best-performing adaptation baselines.

Huiwon Jang, Dongyoung Kim, Junsu Kim et al.

Self-supervised learning of image representations by predicting future frames is a promising direction but still remains a challenge. This is because of the under-determined nature of frame prediction; multiple potential futures can arise from a single current frame. To tackle this challenge, in this paper, we revisit the idea of stochastic video generation that learns to capture uncertainty in frame prediction and explore its effectiveness for representation learning. Specifically, we design a framework that trains a stochastic frame prediction model to learn temporal information between frames. Moreover, to learn dense information within each frame, we introduce an auxiliary masked image modeling objective along with a shared decoder architecture. We find this architecture allows for combining both objectives in a synergistic and compute-efficient manner. We demonstrate the effectiveness of our framework on a variety of tasks from video label propagation and vision-based robot learning domains, such as video segmentation, pose tracking, vision-based robotic locomotion, and manipulation tasks. Code is available on the project webpage: https://sites.google.com/view/2024rsp.

Dongyoung Kim, Jinwoo Shin, Pieter Abbeel et al.

A promising technique for exploration is to maximize the entropy of visited state distribution, i.e., state entropy, by encouraging uniform coverage of visited state space. While it has been effective for an unsupervised setup, it tends to struggle in a supervised setup with a task reward, where an agent prefers to visit high-value states to exploit the task reward. Such a preference can cause an imbalance between the distributions of high-value states and low-value states, which biases exploration towards low-value state regions as a result of the state entropy increasing when the distribution becomes more uniform. This issue is exacerbated when high-value states are narrowly distributed within the state space, making it difficult for the agent to complete the tasks. In this paper, we present a novel exploration technique that maximizes the value-conditional state entropy, which separately estimates the state entropies that are conditioned on the value estimates of each state, then maximizes their average. By only considering the visited states with similar value estimates for computing the intrinsic bonus, our method prevents the distribution of low-value states from affecting exploration around high-value states, and vice versa. We demonstrate that the proposed alternative to the state entropy baseline significantly accelerates various reinforcement learning algorithms across a variety of tasks within MiniGrid, DeepMind Control Suite, and Meta-World benchmarks. Source code is available at https://sites.google.com/view/rl-vcse.

Dong-Young Kim, Dong-Kyun Han, Seo-Hyeon Park et al.

Abnormal driver states, particularly have been major concerns for road safety, emphasizing the importance of accurate drowsiness detection to prevent accidents. Electroencephalogram (EEG) signals are recognized for their effectiveness in monitoring a driver's mental state by monitoring brain activities. However, the challenge lies in the requirement for prior calibration due to the variation of EEG signals among and within individuals. The necessity of calibration has made the brain-computer interface (BCI) less accessible. We propose a practical generalized framework for classifying driver drowsiness states to improve accessibility and convenience. We separate the normalization process for each driver, treating them as individual domains. The goal of developing a general model is similar to that of domain generalization. The framework considers the statistics of each domain separately since they vary among domains. We experimented with various normalization methods to enhance the ability to generalize across subjects, i.e. the model's generalization performance of unseen domains. The experiments showed that applying individual domain-specific normalization yielded an outstanding improvement in generalizability. Furthermore, our framework demonstrates the potential and accessibility by removing the need for calibration in BCI applications.

Dong-Kyun Han, Dong-Young Kim, Geun-Deok Jang

A brain-computer interface (BCI) can't be effectively used since electroencephalography (EEG) varies between and within subjects. BCI systems require calibration steps to adjust the model to subject-specific data. It is widely acknowledged that this is a major obstacle to the development of BCIs. To address this issue, previous studies have trained a generalized model by removing the subjects' information. In contrast, in this work, we introduce a style information encoder as an auxiliary task that classifies various source domains and recognizes open-set domains. Open-set recognition method was used as an auxiliary task to learn subject-related style information from the source subjects, while at the same time helping the shared feature extractor map features in an unseen target. This paper compares various OSR methods within an open-set subject recognition (OSSR) framework. As a result of our experiments, we found that the OSSR auxiliary network that encodes domain information improves generalization performance.

Jeongho Min, Dongyoung Kim, Jaehyup Lee

Cross-view image retrieval, particularly street-to-satellite matching, is a critical task for applications such as autonomous navigation, urban planning, and localization in GPS-denied environments. However, existing approaches often require supervised training on curated datasets and rely on panoramic or UAV-based images, which limits real-world deployment. In this paper, we present a simple yet effective cross-view image retrieval framework that leverages a pretrained vision encoder and a large language model (LLM), requiring no additional training. Given a monocular street-view image, our method extracts geographic cues through web-based image search and LLM-based location inference, generates a satellite query via geocoding API, and retrieves matching tiles using a pretrained vision encoder (e.g., DINOv2) with PCA-based whitening feature refinement. Despite using no ground-truth supervision or finetuning, our proposed method outperforms prior learning-based approaches on the benchmark dataset under zero-shot settings. Moreover, our pipeline enables automatic construction of semantically aligned street-to-satellite datasets, which is offering a scalable and cost-efficient alternative to manual annotation. All source codes will be made publicly available at https://jeonghomin.github.io/street2orbit.github.io/.

Jeongho Min, Yejun Lee, Dongyoung Kim et al.

Recently, reference-based image super-resolution (RefSR) has shown excellent performance in image super-resolution (SR) tasks. The main idea of RefSR is to utilize additional information from the reference (Ref) image to recover the high-frequency components in low-resolution (LR) images. By transferring relevant textures through feature matching, RefSR models outperform existing single image super-resolution (SISR) models. However, their performance significantly declines when a domain gap between Ref and LR images exists, which often occurs in real-world scenarios, such as satellite imaging. In this letter, we introduce a Domain Matching (DM) module that can be seamlessly integrated with existing RefSR models to enhance their performance in a plug-and-play manner. To the best of our knowledge, we are the first to explore Domain Matching-based RefSR in remote sensing image processing. Our analysis reveals that their domain gaps often occur in different satellites, and our model effectively addresses these challenges, whereas existing models struggle. Our experiments demonstrate that the proposed DM module improves SR performance both qualitatively and quantitatively for remote sensing super-resolution tasks.

Dongyoung Kim, Sumin Park, Huiwon Jang et al.

Large Vision-Language Models (LVLMs) have recently shown great promise in advancing robotics by combining embodied reasoning with robot control. A common approach involves training on embodied reasoning tasks related to robot control using Supervised Fine-Tuning (SFT). However, SFT datasets are often heuristically constructed and not explicitly optimized for improving robot control. Furthermore, SFT often leads to issues such as catastrophic forgetting and reduced generalization performance. To address these limitations, we introduce Robot-R1, a novel framework that leverages reinforcement learning to enhance embodied reasoning specifically for robot control. Robot-R1 learns to predict the next keypoint state required for task completion, conditioned on the current scene image and environment metadata derived from expert demonstrations. Inspired by the DeepSeek-R1 learning approach, Robot-R1 samples reasoning-based responses and reinforces those that lead to more accurate predictions. Our experiments show that models trained with Robot-R1 outperform SFT methods on embodied reasoning tasks. Despite having only 7B parameters, Robot-R1 even surpasses GPT-4o on reasoning tasks related to low-level action control, such as spatial and primitive movement reasoning.

Dongseok Lee, Jimyung Hong, Dongyoung Kim et al.

Although LLMs have achieved remarkable performance, the inherent stochasticity of their reasoning process and varying conclusions present significant challenges. Majority voting or Best-of-N with external verification models has been explored to find the most promising solution among multiple LLM outputs. However, these approaches have certain limitations, such as limited applicability or the cost of an additional training step. To address this problem, we propose a novel and effective framework that Recycles Few-shot examples to verify LLM outputs (ReFeri). Our key idea is to additionally utilize the given few-shot examples to evaluate the candidate outputs of the target query, not only using them to generate outputs as the conventional few-shot prompting setup. Specifically, ReFeri evaluates the generated outputs by combining two different scores, designed motivated from Bayes' rule, and subsequently selects the candidate that is both confidently determined and contextually coherent through a few additional LLM inferences. Experiments with three different LLMs and across seven diverse tasks demonstrate that our framework significantly improves the accuracy of LLMs-achieving an average gain of 4.8%-through effective response selection, without additional training.

Daekyu Kwon, Dongyoung Kim, Sehwan Ki et al.

In no-reference image quality assessment (NR-IQA), the challenge of limited dataset sizes hampers the development of robust and generalizable models. Conventional methods address this issue by utilizing large datasets to extract rich representations for IQA. Also, some approaches propose vision language models (VLM) based IQA, but the domain gap between generic VLM and IQA constrains their scalability. In this work, we propose a novel pretraining framework that constructs a generalizable representation for IQA by selectively extracting quality-related knowledge from VLM and leveraging the scalability of large datasets. Specifically, we select optimal text prompts for five representative image quality attributes and use VLM to generate pseudo-labels. Numerous attribute-aware pseudo-labels can be generated with large image datasets, allowing our IQA model to learn rich representations about image quality. Our approach achieves state-of-the-art performance on multiple IQA datasets and exhibits remarkable generalization capabilities. Leveraging these strengths, we propose several applications, such as evaluating image generation models and training image enhancement models, demonstrating our model's real-world applicability.

Dong-Kyun Han, Serkan Musellim, Dong-Young Kim et al.

The inter/intra-subject variability of electroencephalography (EEG) makes the practical use of the brain-computer interface (BCI) difficult. In general, the BCI system requires a calibration procedure to tune the model every time the system is used. This problem is recognized as a major obstacle to BCI, and to overcome it, approaches based on transfer learning (TL) have recently emerged. However, many BCI paradigms are limited in that they consist of a structure that shows labels first and then measures "imagery", the negative effects of source subjects containing data that do not contain control signals have been ignored in many cases of the subject-to-subject TL process. The main purpose of this paper is to propose a method of excluding subjects that are expected to have a negative impact on subject-to-subject TL training, which generally uses data from as many subjects as possible. In this paper, we proposed a BCI framework using only high-confidence subjects for TL training. In our framework, a deep neural network selects useful subjects for the TL process and excludes noisy subjects, using a co-teaching algorithm based on the small-loss trick. We experimented with leave-one-subject-out validation on two public datasets (2020 international BCI competition track 4 and OpenBMI dataset). Our experimental results showed that confidence-aware TL, which selects subjects with small loss instances, improves the generalization performance of BCI.

Dongyoung Kim, Myungsung Kwak, Eunji Won et al.

Text localization from the digital image is the first step for the optical character recognition task. Conventional image processing based text localization performs adequately for specific examples. Yet, a general text localization are only archived by recent deep-learning based modalities. Here we present document Text Localization Generative Adversarial Nets (TLGAN) which are deep neural networks to perform the text localization from digital image. TLGAN is an versatile and easy-train text localization model requiring a small amount of data. Training only ten labeled receipt images from Robust Reading Challenge on Scanned Receipts OCR and Information Extraction (SROIE), TLGAN achieved 99.83% precision and 99.64% recall for SROIE test data. Our TLGAN is a practical text localization solution requiring minimal effort for data labeling and model training and producing a state-of-art performance.

Sungjoo Ha, Martin Kersner, Beomsu Kim et al.

When there is a mismatch between the target identity and the driver identity, face reenactment suffers severe degradation in the quality of the result, especially in a few-shot setting. The identity preservation problem, where the model loses the detailed information of the target leading to a defective output, is the most common failure mode. The problem has several potential sources such as the identity of the driver leaking due to the identity mismatch, or dealing with unseen large poses. To overcome such problems, we introduce components that address the mentioned problem: image attention block, target feature alignment, and landmark transformer. Through attending and warping the relevant features, the proposed architecture, called MarioNETte, produces high-quality reenactments of unseen identities in a few-shot setting. In addition, the landmark transformer dramatically alleviates the identity preservation problem by isolating the expression geometry through landmark disentanglement. Comprehensive experiments are performed to verify that the proposed framework can generate highly realistic faces, outperforming all other baselines, even under a significant mismatch of facial characteristics between the target and the driver.

Seungwoo Choi, Seokjun Seo, Beomjun Shin et al.

Keyword spotting (KWS) plays a critical role in enabling speech-based user interactions on smart devices. Recent developments in the field of deep learning have led to wide adoption of convolutional neural networks (CNNs) in KWS systems due to their exceptional accuracy and robustness. The main challenge faced by KWS systems is the trade-off between high accuracy and low latency. Unfortunately, there has been little quantitative analysis of the actual latency of KWS models on mobile devices. This is especially concerning since conventional convolution-based KWS approaches are known to require a large number of operations to attain an adequate level of performance. In this paper, we propose a temporal convolution for real-time KWS on mobile devices. Unlike most of the 2D convolution-based KWS approaches that require a deep architecture to fully capture both low- and high-frequency domains, we exploit temporal convolutions with a compact ResNet architecture. In Google Speech Command Dataset, we achieve more than \textbf{385x} speedup on Google Pixel 1 and surpass the accuracy compared to the state-of-the-art model. In addition, we release the implementation of the proposed and the baseline models including an end-to-end pipeline for training models and evaluating them on mobile devices.