Abstract A lot of malicious applications appears every day, threatening numerous users. Therefore, a surge studies have been conducted to protect users from newly emerging malware by using machine learning algorithms. Albeit existing or deep learning-based Android detection approaches achieve high accuracy combination multiple features, it is not possible employ them on our mobile devices due the cost for them. In this paper, we propose MAPAS , system, that achieves and adaptable usages...

In this paper, we investigate a behavior of the cell average ergodic capacity for distributed antenna systems (DAS) in composite fading channel model which contains small-scale and large-scale fadings. For Rayleigh fadings, based on proof asymptotic normality, mean variance instantaneous were recently presented as closed form solution. However, solution is too complicated to be applied directly obtaining over cell. work, derive simple accurate expression by utilizing high signal noise ratio...

For multiple-input multiple-output (MIMO) systems, the optimum maximum likelihood (ML) detection requires tremendous complexity as number of antennas or modulation level increases. This paper proposes a new algorithm which attains ML performance with significantly reduced complexity. Based on minimum mean square error (MMSE) criterion, proposed scheme reduces search space by excluding unreliable candidate symbols in data streams. Utilizing probability metric evaluates reliability normalized...

In K-user multiple-input multiple-output (MIMO) interference channels, it was shown that alignment (IA) achieves a full spatial multiplexing gain when perfect channel state information (CSI) is available at each transmitter in the network. When CSI fed back from receivers using limited number of feedback bits, significant performance loss inevitable IA due to quantized knowledge. this paper, we propose new quantization strategy optimize with feedback. our proposed scheme, introduce an...

In this paper, we study coordinated multi-cell downlink systems where multiple base stations jointly design a transmission strategy by sharing channel state information. Particularly, tackle the signal-to-interference-plus-noise ratio (SINR) balancing problem to maximize worst-user rate. We consider single-input single-output (SISO) interference channels (IFC) all nodes are equipped with single antenna, and there is one active user in each cell. First, achievable rate regions symmetric...

In this paper, we analyze the average sum rate of downlink multi-antenna systems with zero-forcing beamforming (ZFBF). practical implementations, each antenna is equipped its own power amplifier and limited individually by linearity amplifier. Thus, paper adopts a more realistic per-antenna constraint instead conventional sum-power on transmit antennas. To end, first show that distribution received signal-to-noise ratio (SNR) ZFBF scheme equal-rate metric can be approximated as minimum...

In this letter, we analyze asymptotic ergodic capacity of MIMO amplify-and-forward (AF) relaying systems which employ linear processing at the relay. By exploiting results for eigenvalue distributions, derive in various antenna regimes as a closed-form expression with arbitrary system parameters. The analyzed demonstrate that increasing number source antennas causes shrink phenomenon is analogous to channel hardening effect multi-user systems. Although assume asymptotically large obtain...

In the forthcoming era of IoT, where everything will be connected, mobile devices play a key role in providing data sharing and user-centric services between devices. such service environment, if application is vulnerable to security threats exposed malicious behavior, malware can spread hundreds millions connected particular, it important isolate respond quickly code. This requires prediction behavior. Currently, risk assessment schemes based on permission use description or user review,...

Pitch control system plays an important role in large wind turbine generator (LWTG) because of the direction and intensity is changing every moment. Therefore, pitch difficult to be set up tuning parameters controller. Conventional PID controller not suit for all operating point. Because, conventional are consistently changed. In this paper, a self method using reinforcement learning proposed (LWTG). Actor-Critic used tune adaptive way by taking advantage model-free on-line properties...

In this paper, we introduce multi-link two-way relay channels where multiple systems are interfering with each other. We study the capacity of system by investigating degree freedom (DOF) various message settings. Specifically, consider two cases interference and X channels. show that channel all nodes have M antennas obtains DOF 2M compare multipair Next, general settings for For case user is equipped 3 relays 4 antennas, prove 8 achieved employing network coding.

Research on human-robot interaction has recently been getting increasing attention. In the research field of interaction, speech signal processing in particular is source much interest. this paper, we report a speaker localization system with six microphones for humanoid robot called MAHRU KIST and propose time delay arrival (TDOA)-based feature matrix its algorithm based minimum sum absolute errors (MSAE) sound localization. The TDOA-based defined as simple database calculated from pairs...

Voice-based interaction generates a lot of interest for natural human–robot interaction. Auditory information from microphones is an essential clue robot's attention to person. Fault detection required in order improve the reliability voice-based This paper proposes new approach real-time fault microphone array conversation without calibration signal and known sound source position by intercorrelation features voice activity detection. The successfully applied six-microphone system,...

It has been shown that the performance of multiple-input multiple-output (MIMO) spatial multiplexing systems is significantly degraded when correlation exists between transmit and receive antenna pairs. In this paper, we investigate designs a new statistical precoder for with maximum likelihood (ML) receiver which requires only statistics at transmitter. Two kinds closed-form solution precoders based on rotation power allocation are proposed by means maximizing minimum Euclidean distance...

Tokenization is a significant primary step for the training of Pre-trained Language Model (PLM), which alleviates challenging Out-of-Vocabulary problem in area Natural Processing. As tokenization strategies can change linguistic understanding, it essential to consider composition input features based on characteristics language model performance. This study answers question "Which strategy enhances Korean Named Entity Recognition (NER) task model?" focusing tokenization, significantly...

The dialogue-based relation extraction (DialogRE) task aims to predict the relations between argument pairs that appear in dialogue. Most previous studies utilize fine-tuning pre-trained language models (PLMs) only with extensive features supplement low information density of dialogue by multiple speakers. To effectively exploit inherent knowledge PLMs without extra layers and consider scattered semantic cues on arguments, we propose a Guiding model RelAtional Semantics using Prompt (GRASP)....

To build a conversational agent that interacts fluently with humans, previous studies blend knowledge or personal profile into the pre-trained language model. However, model considers and persona at same time is still limited, leading to hallucination passive way of using personas. We propose an effective dialogue grounds external simultaneously. The selects proper use for generating answers our candidate scoring implemented poly-encoder. Then, generates utterance lesser more engagingness...

Jungwoo Lim, Myunghoon Kang, Yuna Hur, Seung Won Jeong, Jinsung Kim, Yoonna Jang, Dongyub Lee, Hyesung Ji, DongHoon Shin, Seungryong Heuiseok Lim. Findings of the Association for Computational Linguistics: EMNLP 2022.

Address driver circuits are responsible for a significant portion of power consumption by plasma display panel television. In the latest effort reduction address circuits, power-efficient operation mode is selected controlling voltage capacitor that affects <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">LC</i> resonance. This paper presents an analysis and observes change in previous work too slow to achieve efficient when number data switchings...

In this paper, we study K-user multiple-input multiple-output (MIMO) interference channels (IC) in the presence of antenna correlations. Although alignment (IA) achieves maximum multiplexing gain, i.e., degrees freedom (DOF), actual performance is degraded due to ill-conditioned caused by To enhance overall system performance, thus consider a transmission mode selection scheme which adaptively determines number data streams at each transmitter-receiver pair distributed fashion using local...

Block diagonalization (BD) is an attractive method for approaching sum capacity in multiuser multiple-input multiple-output (MIMO) broadcast channels. However, when channel state information (CSI) limited, the performance of BD degraded because inter-user interference cannot be completely eliminated. In this paper, we study limited feedback systems where a large number users are supported per cell. Based on estimate received signal-to-interference-plus-noise ratio (SINR), present efficient...

This study aims to develop English listening ability of middle school students in below-average level through pop songs and investigate the effectiveness by before after surveys, national tests achievement 10 months. Learning includes motivation, listening, speaking, reading writing. Most all, motivation skills have been proved most fundamental important two factors. For teaching learning skills, can be used as an effective way useful pattern practices related very helpful. help feel more...

Dialogue relation extraction identifies semantic relations between entity pairs in dialogues. This research explores a methodology harnessing the potential of prompt-based fine-tuning paired with trigger-generation approach. Capitalizing on intrinsic knowledge pre-trained language models, this strategy employs triggers that underline entities decisively. In particular, diverging from conventional extractive methods seen earlier research, our study leans towards generative manner for trigger...