A5100677770- Natural Language Processing Techniques
- Cryptography and Data Security
- Topic Modeling
- Cryptography and Residue Arithmetic
- Remote Sensing and LiDAR Applications
- Advanced Vision and Imaging
- Coding theory and cryptography
- Biomedical Text Mining and Ontologies
- Cryptographic Implementations and Security
- Robotics and Sensor-Based Localization
- Automated Road and Building Extraction
- Image and Object Detection Techniques
- Data Management and Algorithms
- Chaos-based Image/Signal Encryption
- Advanced Text Analysis Techniques
- Optical measurement and interference techniques
Villanova University
2022-2024
Southeast University
2022-2023
Horizon Robotics (China)
2023
Boise State University
2022
United States Air Force Research Laboratory
2022
U.S. Air Force Research Laboratory Information Directorate
2022
Recent work for extracting relations from texts has achieved excellent performance. However, most existing methods pay less attention to the efficiency, making it still challenging quickly extract massive or streaming text data in realistic scenarios. The main efficiency bottleneck is that these use a Transformer-based pre-trained language model encoding, which heavily affects training speed and inference speed. To address this issue, we propose fast relation extraction (FastRE) based on...
High-definition (HD) map serves as the essential infrastructure of autonomous driving. In this work, we build up a systematic vectorized annotation framework (termed VMA) for efficiently generating HD large-scale driving scene. We design divide-and-conquer scheme to solve spatial extensibility problem generation, and abstract elements with variety geometric patterns unified point sequence representation, which can be extended most in VMA is highly efficient extensible, requiring negligible...
Recent work for extracting relations from texts has achieved excellent performance. However, most existing methods pay less attention to the efficiency, making it still challenging quickly extract massive or streaming text data in realistic scenarios. The main efficiency bottleneck is that these use a Transformer-based pre-trained language model encoding, which heavily affects training speed and inference speed. To address this issue, we propose fast relation extraction (FastRE) based on...
Approximate arithmetic-based homomorphic encryption (HE) scheme CKKS [CKKS17] is arguably the most suitable one for real-world data-privacy applications due to its wider computation range than other HE schemes such as BGV [BGV14], FV and BFV [Bra12, FV12]. However, crucial operation of called key-switching induces a great amount computational burden in actual deployment situations, creates scalability challenges hardware acceleration. In this paper, we present novel Compact And Scalable...
Estimating the 3D structure of drivable surface and surrounding environment is a crucial task for assisted autonomous driving. It commonly solved either by using sensors such as LiDAR or directly predicting depth points via deep learning. However, former expensive, latter lacks use geometry information scene. In this paper, instead following existing methodologies, we propose Road Planar Parallax Attention Network (RPANet), new neural network sensing from monocular image sequences based on...
Recent work for extracting relations from texts has achieved excellent performance. However, existing studies mainly focus on simple relation extraction, these methods perform not well overlapping triple problem because the tags of shared entities would conflict with each other. Especially, are common and indispensable in Chinese. To address this issue, paper proposes PasCore, which utilizes a global pointer annotation strategy extraction PasCore first obtains sentence vector via general...
<p>Along with the National Institute of Standards and Technology (NIST) post-quantum cryptography (PQC) standardization process, lightweight PQC-related research development have also gained substantial attention from community recently. Ring-Binary-Learning-with-Errors (RBLWE), a variant Ring-LWE, which uses binary errors to replace regular Gaussian distributed achieve smaller complexity, has great potential built such PQC scheme for emerging Internet-of-Things (IoT) edge computing...
<p>Along with the National Institute of Standards and Technology (NIST) post-quantum cryptography (PQC) standardization process, lightweight PQC-related research development have also gained substantial attention from community recently. Ring-Binary-Learning-with-Errors (RBLWE), a variant Ring-LWE, which uses binary errors to replace regular Gaussian distributed achieve smaller complexity, has great potential built such PQC scheme for emerging Internet-of-Things (IoT) edge computing...
Estimating the 3D structure of drivable surface and surrounding environment is a crucial task for assisted autonomous driving. It commonly solved either by using sensors such as LiDAR or directly predicting depth points via deep learning. However, former expensive, latter lacks use geometry information scene. In this paper, instead following existing methodologies, we propose Road Planar Parallax Attention Network (RPANet), new neural network sensing from monocular image sequences based on...