A5101694303- Speech and Audio Processing
- Speech Recognition and Synthesis
- Mobile Ad Hoc Networks
- Music and Audio Processing
- Education and Work Dynamics
- Higher Education and Teaching Methods
- Ideological and Political Education
- Energy Efficient Wireless Sensor Networks
- Digital Media Forensic Detection
- Voice and Speech Disorders
- Cryptographic Implementations and Security
- Opportunistic and Delay-Tolerant Networks
- Security in Wireless Sensor Networks
- Advanced Authentication Protocols Security
- Educational Technology and Pedagogy
- Composite Material Mechanics
- Advanced Numerical Methods in Computational Mathematics
- Scheduling and Timetabling Solutions
- Advanced Computational Techniques and Applications
- Coal and Its By-products
- Advanced Wireless Communication Techniques
- Service-Oriented Architecture and Web Services
- Advanced Mathematical Modeling in Engineering
- Advanced Wireless Network Optimization
- Chaos-based Image/Signal Encryption
Duke Kunshan University
2022-2023
Cytoskeleton (United States)
2023
Institute for Forecasting of the Slovak Academy of Sciences
2023
Wuhan University
2007-2022
Anhui University of Science and Technology
2020
Jeonbuk National University
2010
Daqing Normal University
2009
University of Chinese Academy of Sciences
2009
Beihang University
2008
Anhui Agricultural University
2007
Wireless local area networks are presently widely deployed so that different stations can potentially associate with an access point (AP). The AP selection strategy is one of the significant research areas for wireless networks. solves relevant problem which be selected and associated a station efficient resource utilization obtained. Rather than merely choosing strongest received signal strength, however, we calculate effective throughput each neighboring use result as basis selection....
This paper describes our DKU-OPPO system for the 2022 Spoofing-Aware Speaker Verification (SASV) Challenge.First, we split joint task into speaker verification (SV) and spoofing countermeasure (CM), these two tasks which are optimized separately.For ASV systems, four state-of-the-art methods employed.For CM propose on top of challenge baseline to further improve performance, namely Embedding Random Sampling Augmentation (ERSA) One-Class Confusion Loss(OCCL).Second, also explore whether SV...
Despite the great performance of language identification (LID), there is a lack large-scale singing LID databases to support research (SLID). This paper proposed over 3200 hours dataset used for identification, called Slingua. As baseline, we explore two self-supervised learning (SSL) models, WavLM and Wav2vec2, as feature extractors both SLID universal speech (ULID), compared with traditional handcraft feature. Moreover, by training corpus, compare difference identification. The final...
In popular symmetric ciphers, S-box substitution is the core operation that dominates executions of cryptographic algorithms. this paper, a method application-specific instruction-set extension used for accelerating key in cryptography. Two instructions access are designed by constructing novel flexible on-chip parallel box unit consists multiple lookup tables and post-processing module. The integrated into 32-bit configurable Leon2 processor. Configuration presented. Implementing extended...
Recent anti-spoofing systems focus on spoofing detection, where the task is only to determine whether test audio fake. However, there are few studies putting attention identifying methods of generating fake speech. Common attack algorithms in logical access (LA) scenario, such as voice conversion and speech synthesis, can be divided into several stages: input processing, conversion, waveform generation, etc. In this work, we propose a system for classifying different attributes, representing...
A reliable voice anti-spoofing countermeasure system needs to robustly protect automatic speaker verification (ASV) systems in various kinds of spoofing scenarios. However, the performance could be degraded by channel effects and codecs. In this paper, we show that using low-frequency subbands signals as input can mitigate negative impact introduced codecs on systems. To validate this, two types low-pass filters with different cut-off frequencies are applied systems, equal error rate (EER)...
Current research in synthesized speech detection primarily focuses on the generalization of systems to unknown spoofing methods noise-free speech. However, performance anti-spoofing countermeasures (CM) system is often don't work as well more challenging scenarios, such those involving noise and reverberation. To address problem enhancing robustness CM systems, we propose a transfer learning-based enhancement front-end joint optimization (TL-SEJ) method, investigating its effectiveness...
Recently, much attention is drawn for integrating MANET with Internet in order to extend network coverage and applications domain of ad hoc networks. In this paper, we propose a modified Dynamic Sequence Routing method providing route between through gateway. the proposed scheme introduce flag field packet differentiate whether destined or Internet. The gateway provides address translation functionality by maintaining table mapping local global addresses. verify performance our system have...
This paper describes a new tree-based routing method which can be used to connect heterogeneous networks. In the proposed method, source node located in one network sends packets group of nodes another network. our scenario, normally belongs mobile ad hoc network, while consists sensor specified geographical region. The packet destination order aggregate data. Our protocol attains energy efficiency and accuracy by incorporating metric for constructing tree according routes. is calculated...
Recent anti-spoofing systems focus on spoofing detection, where the task is only to determine whether test audio fake. However, there are few studies putting attention identifying methods of generating fake speech. Common attack algorithms in logical access (LA) scenario, such as voice conversion and speech synthesis, can be divided into several stages: input processing, conversion, waveform generation, etc. In this work, we propose a system for classifying different attributes, representing...
poster Share on Why and How to Teach Machine Learning in High School –A Practice Bejing National Day China Authors: Zijie Zheng Department of Math, Beijing School, 0009-0003-8979-028XSearch about this author , Siyao Han Computer Science, 0009-0007-2235-3930Search Weiqi Zhang 0009-0002-5665-8189Search Xingming Wang 0009-0004-1301-7597Search Chao Gong Tsinghua University, 0000-0003-1160-0565Search Authors Info & Claims ACM TURC '23: Proceedings the Turing Award Celebration Conference -...
AdaBoost regression predicts the ranking of college students using super star learning APP
Due to the resource limitations of sensor nodes, providing security protocols is a particular challenge in networks. A popular proposed method neighborhood based key agreement protocol (NEKAP), which an improvement over well-known Localized Encryption and Authentication Protocol (LEAP). NEKAP efficient light-weight protocol, but includes loopholes through adversaries may launch replay attack by successfully masquerading as legitimate nodes thereby compromise communications network. In this...
In recent years, the combination of case-based reasoning (CBR) and domain knowledge has become a hotspot in CBR field. Different representation ways bring about different influences system performance. This paper makes summarization an analysis deficiencies integration between traditional CBR, based on which it proposes scalable case model using description logic (DL). Moreover, corresponding retrieval algorithm is also presented. The mainly serves for revision. Furthermore, usage...
Wireless mesh networks (WMNs) are drawing attention for commercial applications. Service should be uninterrupted when roaming from one domain to another, security maintained, and billing accurate. However, conventional authentication methods have drawbacks such as coordination between the home foreign domains connectivity them. In this paper, we first present a re-authentication method which enables mobile node authenticated by may not directly connected domain. Even though current does...