Kyung Ho Park

Kyung Ho Park

Visual localization is an essential component of intelligent transportation systems, enabling broad applications that require understanding one's self location when other sensors are not available. It is mostly tackled by image retrieval such that the location of a query image is determined by its closest match in the previously collected images. Existing approaches focus on large scale localization where landmarks are helpful in finding the location. However, visual localization becomes challenging in small scale environments where objects are hardly recognizable. In this paper, we propose a visual localization framework that robustly finds the match for a query among the images collected from indoor parking lots. It is a challenging problem when the vehicles in the images share similar appearances and are frequently replaced such as parking lots. We propose to employ a deep dense local feature matching that resembles human perception to find correspondences and eliminating matches from vehicles automatically with a vehicle detector. The proposed solution is robust to the scenes with low textures and invariant to false matches caused by vehicles. We compare our framework with alternatives to validate our superiority on a benchmark dataset containing 267 pre-collected images and 99 query images taken from 34 sections of a parking lot. Our method achieves 86.9 percent accuracy, outperforming the alternatives.

Kyung Ho Park, Hyunhee Chung, Soonwoo Kwon

Upon the significant performance of the supervised deep neural networks, conventional procedures of developing ML system are \textit{task-centric}, which aims to maximize the task accuracy. However, we scrutinized this \textit{task-centric} ML system lacks in engineering efficiency when the ML practitioners solve multiple tasks in their domain. To resolve this problem, we propose an \textit{efficiency-centric} ML system that concatenates numerous datasets, classifiers, out-of-distribution detectors, and prediction tables existing in the practitioners' domain into a single ML pipeline. Under various image recognition tasks in the real world car-sharing platform, our study illustrates how we established the proposed system and lessons learned from this journey as follows. First, the proposed ML system accomplishes supreme engineering efficiency while achieving a competitive task accuracy. Moreover, compared to the \textit{task-centric} paradigm, we discovered that the \textit{efficiency-centric} ML system yields satisfactory prediction results on multi-labelable samples, which frequently exist in the real world. We analyze these benefits derived from the representation power, which learned broader label spaces from the concatenated dataset. Last but not least, our study elaborated how we deployed this \textit{efficiency-centric} ML system is deployed in the real world live cloud environment. Based on the proposed analogies, we highly expect that ML practitioners can utilize our study to elevate engineering efficiency in their domain.

Kyung Ho Park, Hyunhee Chung

Recent progress of deep learning has empowered various intelligent transportation applications, especially in car-sharing platforms. While the traditional operations of the car-sharing service highly relied on human engagements in fleet management, modern car-sharing platforms let users upload car images before and after their use to inspect the cars without a physical visit. To automate the aforementioned inspection task, prior approaches utilized deep neural networks. They commonly employed pre-training, a de-facto technique to establish an effective model under the limited number of labeled datasets. As candidate practitioners who deal with car images would presumably get suffered from the lack of a labeled dataset, we analyzed a sophisticated analogy into the effectiveness of pre-training is important. However, prior studies primarily shed a little spotlight on the effectiveness of pre-training. Motivated by the aforementioned lack of analysis, our study proposes a series of analyses to unveil the effectiveness of various pre-training methods in image recognition tasks at the car-sharing platform. We set two real-world image recognition tasks in the car-sharing platform in a live service, established them under the many-shot and few-shot problem settings, and scrutinized which pre-training method accomplishes the most effective performance in which setting. Furthermore, we analyzed how does the pre-training and fine-tuning convey different knowledge to the neural networks for a precise understanding.

Young Ah Choi, Kyung Ho Park, Eunji Park et al.

In the era of intelligent transportation, driver behavior profiling has become a beneficial technology as it provides knowledge regarding the driver's aggressiveness. Previous approaches achieved promising driver behavior profiling performance through establishing statistical heuristics rules or supervised learning-based models. Still, there exist limits that the practitioner should prepare a labeled dataset, and prior approaches could not classify aggressive behaviors which are not known a priori. In pursuit of improving the aforementioned drawbacks, we propose a novel approach to driver behavior profiling leveraging an unsupervised learning paradigm. First, we cast the driver behavior profiling problem as anomaly detection. Second, we established recurrent neural networks that predict the next feature vector given a sequence of feature vectors. We trained the model with normal driver data only. As a result, our model yields high regression error given a sequence of aggressive driver behavior and low error given at a sequence of normal driver behavior. We figured this difference of error between normal and aggressive driver behavior can be an adequate flag for driver behavior profiling and accomplished a precise performance in experiments. Lastly, we further analyzed the optimal level of sequence length for identifying each aggressive driver behavior. We expect the proposed approach to be a useful baseline for unsupervised driver behavior profiling and contribute to the efficient, intelligent transportation ecosystem.

Eunji Park, Kyung Ho Park, Huy Kang Kim

The game industry has long been troubled by malicious activities utilizing game bots. The game bots disturb other game players and destroy the environmental system of the games. For these reasons, the game industry put their best efforts to detect the game bots among players' characters using the learning-based detections. However, one problem with the detection methodologies is that they do not provide rational explanations about their decisions. To resolve this problem, in this work, we investigate the explainabilities of the game bot detection. We develop the XAI model using a dataset from the Korean MMORPG, AION, which includes game logs of human players and game bots. More than one classification model has been applied to the dataset to be analyzed by applying interpretable models. This provides us explanations about the game bots' behavior, and the truthfulness of the explanations has been evaluated. Besides, interpretability contributes to minimizing false detection, which imposes unfair restrictions on human players.

Kyung Ho Park, Eunji Park, Huy Kang Kim

Along with the importance of safety, an IDS has become a significant task in the real world. Prior studies proposed various intrusion detection models for the UAV. Past rule-based approaches provided a concrete baseline IDS model, and the machine learning-based method achieved a precise intrusion detection performance on the UAV with supervised learning models. However, previous methods have room for improvement to be implemented in the real world. Prior methods required a large labeling effort on the dataset, and the model could not identify attacks that were not trained before. To jump over these hurdles, we propose an IDS with unsupervised learning. As unsupervised learning does not require labeling, our model let the practitioner not to label every type of attack from the flight data. Moreover, the model can identify an abnormal status of the UAV regardless of the type of attack. We trained an autoencoder with the benign flight data only and checked the model provides a different reconstruction loss at the benign flight and the flight under attack. We discovered that the model produces much higher reconstruction loss with the flight under attack than the benign flight; thus, this reconstruction loss can be utilized to recognize an intrusion to the UAV. With consideration of the computation overhead and the detection performance in the wild, we expect our model can be a concrete and practical baseline IDS on the UAV.

Kyung Ho Park, Huy Kang Kim

As a car becomes more connected, a countermeasure against automobile theft has become a significant task in the real world. To respond to automobile theft, data mining, biometrics, and additional authentication methods are proposed. Among current countermeasures, data mining method is one of the efficient ways to capture the owner driver's unique characteristics. To identify the owner driver from thieves, previous works applied various algorithms toward driving data. Such data mining methods utilized supervised learning, thus required labeled data set. However, it is unrealistic to gather and apply the thief's driving pattern. To overcome this problem, we propose driver identification method with GAN. GAN has merit to build identification model by learning the owner driver's data only. We trained GAN only with owner driver's data and used trained discriminator to identify the owner driver. From actual driving data, we evaluated our identification model recognizes the owner driver well. By ensembling various driver authentication methods with the proposed model, we expect industry can develop automobile theft countermeasures available in the real world.

Yong Goo Kang, Kyung Ho Park, Huy Kang Kim

As automobiles become intelligent, automobile theft methods are evolving intelligently. Therefore automobile theft detection has become a major research challenge. Data-mining, biometrics, and additional authentication methods have been proposed to address automobile theft, in previous studies. Among these methods, data-mining can be used to analyze driving characteristics and identify a driver comprehensively. However, it requires a labeled driving dataset to achieve high accuracy. It is impractical to use the actual automobile theft detection system because real theft driving data cannot be collected in advance. Hence, we propose a method to detect an automobile theft attempt using only owner driving data. We cluster the key features of the owner driving data using the k-means algorithm. After reconstructing the driving data into one of these clusters, theft is detected using an error from the original driving data. To validate the proposed models, we tested our actual driving data and obtained 99% accuracy from the best model. This result demonstrates that our proposed method can detect vehicle theft by using only the car owner's driving data.

Kyung Ho Park, Eunjo Lee, Huy Kang Kim

With the rapid growth of MMORPG market, game bot detection has become an essential task for maintaining stable in-game ecosystem. To classify bots from normal users, detection methods are proposed in both game client and server-side. Among various classification methods, data mining method in server-side captured unique characteristics of bots efficiently. For features used in data mining, behavioral and social actions of character are analyzed with numerous algorithms. However, bot developers can evade the previous detection methods by changing bot's activities continuously. Eventually, overall maintenance cost increases because the selected features need to be updated along with the change of bot's behavior. To overcome this limitation, we propose improved bot detection method with financial analysis. As bot's activity absolutely necessitates the change of financial status, analyzing financial fluctuation effectively captures bots as a key feature. We trained and tested model with actual data of Aion, a leading MMORPG in Asia. Leveraging that LSTM efficiently recognizes time-series movement of data, we achieved meaningful detection performance. Further on this model, we expect sustainable bot detection system in the near future.