A5060660400- Advanced Vision and Imaging
- Computer Graphics and Visualization Techniques
- 3D Shape Modeling and Analysis
- Advanced Harmonic Analysis Research
- Advanced Wireless Communication Techniques
- Advanced Memory and Neural Computing
- Holomorphic and Operator Theory
- Advanced Neural Network Applications
- Electric and Hybrid Vehicle Technologies
- Algebraic and Geometric Analysis
- Neural Networks and Applications
- Tea Polyphenols and Effects
- Electric Vehicles and Infrastructure
- Traffic control and management
- Wireless Communication Networks Research
- Natural Language Processing Techniques
- Ferroelectric and Negative Capacitance Devices
- Advanced Battery Technologies Research
- 3D Surveying and Cultural Heritage
- Web Data Mining and Analysis
- PAPR reduction in OFDM
- Robotics and Sensor-Based Localization
- Advanced Image and Video Retrieval Techniques
- Terahertz technology and applications
- Software Engineering Research
ZheJiang Academy of Agricultural Sciences
2024-2025
Zhejiang A & F University
2024-2025
Kunming University of Science and Technology
2025
Georgia Institute of Technology
2023-2024
Texas A&M University
2023-2024
University of Electronic Science and Technology of China
2006-2023
West China Medical Center of Sichuan University
2023
Jilin Agricultural University
2023
Sichuan University
2023
Jinan University
2021-2022
In the recent years, spiking neural network (SNN) has attracted increasing attention due to its low energy consumption and online learning potential. However, design of SNN processor not been thoroughly investigated in past, resulting limited performance consumption. this work, a fast energy-efficient with adaptive clock/event-driven computation scheme capability proposed. Several techniques have proposed reduce time consumption, including Adaptive Clock- Event-Driven Computing Scheme,...
Neural Radiance Field (NeRF) based 3D reconstruction is highly desirable for immersive Augmented and Virtual Reality (AR/VR) applications, but achieving instant (i.e., < 5 seconds) on-device NeRF training remains a challenge. In this work, we first identify the inefficiency bottleneck: need to interpolate embeddings up 200,000 times from embedding grid during each iteration. To alleviate this, propose Instant-3D, an algorithm-hardware co-design acceleration framework that achieves training....
The remarkable capabilities and intricate nature of Artificial Intelligence (AI) have dramatically escalated the imperative for specialized AI accelerators. Nonetheless, designing these accelerators various workloads remains both labor- time-intensive. While existing design exploration automation tools can partially alleviate need extensive human involvement, they still demand substantial hardware expertise, posing a barrier to non-experts stifling accelerator development. Motivated by...
This paper develops a sequencing-enabled hierarchical connected automated vehicle (CAV) cooperative on-ramp merging control framework. The proposed framework consists of two-layer design: the upper-level sequences vehicles to harmonize traffic density across mainline and segments, simultaneously enhancing lower-level efficiency through mixed-integer linear programming formulation. Subsequently, control, in turn, employs longitudinal distributed model predictive (MPC) supplemented by virtual...
Roasting significantly influences the formation of heterocyclic compounds in tea. This study investigated rarely reported alkyloxazole compounds. Totally 9 alkyloxazoles were identified unequivocally by SPME and SPE tea infusion. Quantification through standard addition method revealed content was affected roasting conditions (temperature duration) showed differences between stalks leaves, with these being widely distributed roasted teas. Alkyl side chains influenced aroma characteristics,...
Scarce feature points are a critical limitation affecting the accuracy and stability of incremental structure from motion (SfM) in small-scale scenes. In this paper, we propose an SfM method for scenes, combined with auxiliary calibration plate. This approach increases number sparse regions, randomly generate within those areas. At same time, obtain coarse matching set using pairwise polar geometric constraints. The positional results constraints plate then used to filter out high-precision...
Neural Radiance Field (NeRF) based rendering has attracted growing attention thanks to its state-of-the-art (SOTA) quality and wide applications in Augmented Virtual Reality (AR/VR). However, immersive real-time (> 30 FPS) NeRF enabled interactions are still limited due the low achievable throughput on AR/VR devices. To this end, we first profile SOTA efficient algorithms commercial devices identify two primary causes of aforementioned inefficiency: (1) uniform point sampling (2) dense...
The ECG classification processor is a key component in wearable intelligent monitoring devices which monitor the signals real time and detect abnormality automatically. state-of-the-art processors for are faced with two challenges, including ultra-low energy consumption demand high accuracy against patient-to-patient variability. To address above this work, an ultra-energy-efficient proposed. Several design techniques have been proposed, reconfigurable SNN/ANN inference architecture reducing...
The National Institute of Diabetes and Digestive Kidney Diseases (NIDDK) Central Repository makes data biospecimens from NIDDK-funded research available to the broader scientific community. It thereby facilitates: testing new hypotheses without or biospecimen collection; pooling across several studies increase statistical power; informative genetic analyses using Repository's well-curated phenotypic data. This article describes initial database plan for its revision a simpler model. Among...
In vehicle operation, in order to maximize the fuel economy, propulsion system control can easily adapt pressure and temperature variations as these be measured by sensors. However, it is challenging detect driving cycles. With growing progress made artificial intelligence field, pattern recognition gains momentum various applications. This study presents a on cycle based supervised learning. Training data 2-D visualization achieved t-distributed stochastic neighbor embedding (t-SNE)...
Novel view synthesis is an essential functionality for enabling immersive experiences in various Augmented- and Virtual-Reality (AR/VR) applications, which Neural Radiance Field (NeRF) has emerged as the state-of-the-art (SOTA) technique. In particular, generalizable NeRFs have gained increasing popularity thanks to their cross-scene generalization capability, enables be instantly serviceable new scenes without per-scene training. Despite promise, aggravate prohibitive complexity of due...
In this paper, we proposed a novel frequency synchronization method which has larger estimation range than conventional by Tufvesson. A two steps offset is performed in the new method. first step received signal correlated with local training sequence to eliminate influence of makes independent period repeated data sequence. The its delay second obtain fine offset. Precise can be achieved Tufvesson's determines method, while they are irrelevant. Tens subcarriers spacing estimated used other...
Score-based Generative Models (SGMs) is one leading method in generative modeling, renowned for their ability to generate high-quality samples from complex, high-dimensional data distributions. The enjoys empirical success and supported by rigorous theoretical convergence properties. In particular, it has been shown that SGMs can a distribution close the ground-truth if underlying score function learned well, suggesting of SGM as model. We provide counter-example this paper. Through sample...
Federated learning is an important framework in modern machine that seeks to integrate the training of models from multiple users, each user having their own local data set, a way sensitive privacy and communication loss constraints. In clustered federated learning, one assumes additional unknown group structure among goal train are useful for group, rather than simply single global model all users. this paper, we propose novel solution problem inspired by ideas consensus-based optimization...