arXiv subjects

Hojun Jin

Publications and source records attributed to Hojun Jin.

10 recordsLinked to original sources

arXiv2025

Beyond Hard Sharing: Efficient Multi-Task Speech-to-Text Modeling with Supervised Mixture of Experts

Hard-parameter sharing is a common strategy to train a single model jointly across diverse tasks. However, this often leads to task interference, impeding overall model performance. To address the issue, we propose a simple yet effective Supervised Mixture of Experts (S-MoE). Unlike traditional Mixture of Experts models, S-MoE eliminates the need for training gating functions by utilizing special guiding tokens to route each task to its designated expert. By assigning each task to a separate feedforward network, S-MoE overcomes the limitations of hard-parameter sharing. We further apply S-MoE to a speech-to-text model, enabling the model to process mixed-bandwidth input while jointly performing automatic speech recognition (ASR) and speech translation (ST). Experimental results demonstrate the effectiveness of the proposed S-MoE, achieving a 6.35% relative improvement in Word Error Rate (WER) when applied to both the encoder and decoder.

cs.CL↗

arXiv2022

Reinforcement Learning-Based Cooperative P2P Power Trading between DC Nanogrid Clusters with Wind and PV Energy Resources

In replacing fossil fuels with renewable energy resources for carbon neutrality, the unbalanced resource production of intermittent wind and photovoltaic (PV) power is a critical issue for peer-to-peer (P2P) power trading. To address this issue, a reinforcement learning (RL) technique is introduced in this paper. For RL, a graph convolutional network (GCN) and a bi-directional long short-term memory (Bi-LSTM) network are jointly applied to P2P power trading between nanogrid clusters, based on cooperative game theory. The flexible and reliable DC nanogrid is suitable for integrating renewable energy for a distribution system. Each local nanogrid cluster takes the position of prosumer, focusing on power production and consumption simultaneously. For the power management of nanogrid cluster, multi-objective optimization is applied to each local nanogrid cluster with the Internet of Things (IoT) technology. Charging/discharging of an electric vehicle (EV) is executed considering the intermittent characteristics of wind and PV power production. RL algorithms, such as GCN- convolutional neural network (CNN) layers for deep Q-learning network (DQN), GCN-LSTM layers for deep recurrent Q-learning network (DRQN), GCN-Bi-LSTM layers for DRQN, and GCN-Bi-LSTM layers for proximal policy optimization (PPO), are used for simulations. Consequently, the cooperative P2P power trading system maximizes the profit by considering the time of use (ToU) tariff-based electricity cost and the system marginal price (SMP), and minimizes the amount of grid power consumption. Power management of nanogrid clusters with P2P power trading is simulated on a distribution test feeder in real time, and the proposed GCN-Bi-LSTM-PPO technique achieving the lowest electricity cost among the RL algorithms used for comparison reduces the electricity cost by 36.7%, averaging over nanogrid clusters.

cs.LG↗

arXiv2022

Learning Dense Features for Point Cloud Registration Using a Graph Attention Network

Point cloud registration is a fundamental task in many applications such as localization, mapping, tracking, and reconstruction. Successful registration relies on extracting robust and discriminative geometric features. Though existing learning based methods require high computing capacity for processing a large number of raw points at the same time, computational capacity limitation is not an issue thanks to the powerful parallel computing process using GPU. In this paper, we introduce a framework that efficiently and economically extracts dense features using a graph attention network for point cloud matching and registration (DFGAT). The detector of the DFGAT is responsible for finding highly reliable key points in large raw data sets. The descriptor of the DFGAT takes these keypoints combined with their neighbors to extract invariant density features in preparation for the matching. The graph attention network (GAT) uses the attention mechanism that enriches the relationships between point clouds. Finally, we consider this as an optimal transport problem and use the Sinkhorn algorithm to find positive and negative matches. We perform thorough tests on the KITTI dataset and evaluate the effectiveness of this approach. The results show that this method with the efficiently compact keypoint selection and description can achieve the best performance matching metrics and reach the highest success ratio of 99.88% registration in comparison with other state of the art approaches.

cs.CV↗

arXiv2022

Power Management in Smart Residential Building with Deep Learning Model for Occupancy Detection by Usage Pattern of Electric Appliances

With the growth of smart building applications, occupancy information in residential buildings is becoming more and more significant. In the context of the smart buildings' paradigm, this kind of information is required for a wide range of purposes, including enhancing energy efficiency and occupant comfort. In this study, occupancy detection in residential building is implemented using deep learning based on technical information of electric appliances. To this end, a novel approach of occupancy detection for smart residential building system is proposed. The dataset of electric appliances, sensors, light, and HVAC, which is measured by smart metering system and is collected from 50 households, is used for simulations. To classify the occupancy among datasets, the support vector machine and autoencoder algorithm are used. Confusion matrix is utilized for accuracy, precision, recall, and F1 to demonstrate the comparative performance of the proposed method in occupancy detection. The proposed algorithm achieves occupancy detection using technical information of electric appliances by 95.7~98.4%. To validate occupancy detection data, principal component analysis and the t-distributed stochastic neighbor embedding (t-SNE) algorithm are employed. Power consumption with renewable energy system is reduced to 11.1~13.1% in smart buildings by using occupancy detection.

eess.SP↗

arXiv2022

Anomaly Detection of Smart Metering System for Power Management with Battery Storage System/Electric Vehicle

A novel smart metering technique capable of anomaly detection was proposed for real-time home power management system. Smart meter data generated in real-time was obtained from 900 households of single apartments. To detect outliers and missing values in smart meter data, a deep learning model, the autoencoder, consisting of a graph convolutional network and bidirectional long short-term memory network, was applied to the smart metering technique. Power management based on the smart metering technique was performed by multi-objective optimization in the presence of a battery storage system and an electric vehicle. The results of the power management employing the proposed smart metering technique indicate a reduction in electricity cost and amount of power supplied by the grid compared to the results of power management without anomaly detection.

eess.SY↗

arXiv2022

Development of Charging, Discharging Scheduling Algorithm for Economical and Energy Efficient Operation of Multi EV Charging Station

As the number of electric vehicles (EVs) significantly increases, the excessive charging demand of parked EVs in the charging station may incur an instability problem to the electricity network during peak hours. For the charging station to take a microgrid (MG) structure, an economical and energy-efficient power management scheme is required for the power provision of EVs while considering the local load demand of the MG. For these purposes, this study presents the power management scheme of interdependent MG and EV fleets aided by a novel EV charg-ing/discharging scheduling algorithm. In this algorithm, the maximum amount of discharging power from parked EVs is determined based on the difference between local load demand and photovoltaic (PV) power production to alleviate imbalances occurred between them. For the power management of the MG with charging/discharging scheduling of parked EVs in the PV-based charging station, multi-objective optimization is performed to minimize the operating cost and grid dependency. In addition, the proposed scheme maximizes the utilization of EV charging/discharging while satisfying the charging requirements of parked EVs. Moreover, a more economical and energy-efficient PV-based charging station is established using the future trends of local load demand and PV power production predicted by a gated recurrent unit (GRU) network. With the proposed EV charging/discharging scheduling algorithm, the operating cost of PV-based charging station is decreased by 167.71% and 28.85% compared with the EV charg-ing scheduling algorithm and the conventional EV charging/discharging scheduling algorithm, respectively. It is obvious that the economical and energy-efficient operation of PV-based charging station can be accomplished by applying the power management scheme with the proposed EV charging/discharging scheduling strategy.

eess.SY↗

arXiv2021

Optimal Scheduling of Energy Storage for Power System with Capability of Sensing Short-term Future PV Power Production

Constant rise in energy consumption that comes with the population growth and introduction of new technologies has posed critical issues such as efficient energy management on the consumer side. That has elevated the importance of the use of renewable energy sources, particularly photovoltaic (PV) system and wind turbine. This work models and discusses design options based on the hybrid power system of grid and battery storage. The effects of installed capacity on renewable penetration (RP) and cost of electricity (COE) are investigated for each modality. For successful operation of hybrid power system and electricity trading in power market, accurate predictions of PV power production and load demand are taken into account. A machine learning (ML) model is introduced for scheduling, and predicting variations of the PV power production and load demand. Fitness of the ML model shows, when employing a linear regression model, the mean squared error (MSE) of 0.000012, root mean square error (RMSE) of 0.003560 and R2 of 0.999379. Using predicted PV power production and load demand, reduction of electricity cost is 37.5 % when PV and utility grid are utilized, and is 43.06% when PV, utility grid, and storage system are utilized.

eess.SY↗

arXiv2021

Power Management of Microgrid Integrated with Electric Vehicles in Residential Parking Station

Lately, increasing number of electric vehicles (EVs) in residential parking station has become an important issue, because excessive number of EVs can destabilize the power system during peak hours with high charging power requested. When the power system of the residential parking station takes the structure of microgrid (MG), power provision for the EVs requires efficient power management scheme. To minimize the maintenance cost of the MG and maintain the grid stability, the MG needs to balance the charging/discharging power of EVs in the parking station. To achieve these goals, this paper proposes a charging/discharging algorithm suitable for the power management of the MG configured with EVs. Multi-objective optimization is taken to MG to minimize the maintenance cost and the grid dependency while maximizing the use of photovoltaic (PV) power and the utilization of EVs as energy storage systems (ESSs). In our approach, to increase the usefulness of discharging power of EVs, the base load and the PV power production are considered together for power management to mitigate imbalances incurred between them. As a result, proposed approach demonstrates superior power management performance as compared to other comparable ones.

eess.SY↗

arXiv2021

Smart Metering System Capable of Anomaly Detection by Bi-directional LSTM Autoencoder

Anomaly detection is concerned with a wide range of applications such as fault detection, system monitoring, and event detection. Identifying anomalies from metering data obtained from smart metering system is a critical task to enhance reliability, stability, and efficiency of the power system. This paper presents an anomaly detection process to find outliers observed in the smart metering system. In the proposed approach, bi-directional long short-term memory (BiLSTM) based autoencoder is used and finds the anomalous data point. It calculates the reconstruction error through autoencoder with the non-anomalous data, and the outliers to be classified as anomalies are separated from the non-anomalous data by predefined threshold. Anomaly detection method based on the BiLSTM autoencoder is tested with the metering data corresponding to 4 types of energy sources electricity/water/heating/hot water collected from 985 households.

cs.LG↗

arXiv2020

Power Management of Nanogrid Cluster with P2P Electricity Trading Based on Future Trends of Load Demand and PV Power Production

This paper presents the power management of the nanogrid clusters assisted by a novel peer-to-peer(P2P) electricity trading. In our work, unbalance of power consumption among clusters is mitigated by the proposed P2P trading method. For power management of individual clusters, multi-objective optimization simultaneously minimizing total power consumption, portion of grid power consumption, and total delay incurred by scheduling is attempted. A renewable power source photovoltaic(PV) system is adopted for each cluster as a secondary source. The temporal surplus of self-supply PV power of a cluster can be sold through P2P trading to another cluster (s) experiencing temporal power shortage. The cluster in temporal shortage of electric power buys the PV power to reduce peak load and total delay. In P2P trading, a cooperative game model is used for buyers and sellers to maximize their welfare. To increase P2P trading efficiency, future trends of load demand and PV power production are considered for power management of each cluster to resolve instantaneous unbalance between load demand and PV power production. To this end, a gated recurrent unit network is used to forecast future load demand and future PV power production. Simulations verify the effectiveness of the proposed P2P trading for nanogrid clusters.

eess.SY↗