A5033649899- Advanced Optical Network Technologies
- Cloud Computing and Resource Management
- Software-Defined Networks and 5G
- Advanced Image and Video Retrieval Techniques
- Video Surveillance and Tracking Methods
- Optical Network Technologies
- Domain Adaptation and Few-Shot Learning
- Fuel Cells and Related Materials
- Interconnection Networks and Systems
- Astrophysics and Cosmic Phenomena
- Radio Astronomy Observations and Technology
- Currency Recognition and Detection
- Advanced Computational Techniques and Applications
- Particle Dynamics in Fluid Flows
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Internet Traffic Analysis and Secure E-voting
- Optical Wireless Communication Technologies
- Astronomical Observations and Instrumentation
- Educational Technology and Assessment
- Error Correcting Code Techniques
- Organizational and Employee Performance
- Graph theory and applications
- Neural Networks and Reservoir Computing
- Privacy-Preserving Technologies in Data
- Advanced SAR Imaging Techniques
Inspur (China)
2023-2024
Peking University First Hospital
2024
National Medical Products Administration
2024
Peking University
2024
China Electronic Information Industry Development
2023
Zhejiang Normal University
2023
Heilongjiang University
2023
Huawei Technologies (United Kingdom)
2022
Huawei Technologies (China)
2021
Beijing University of Posts and Telecommunications
2017-2020
Deep metric learning, which learns discriminative features to process image clustering and retrieval tasks, has attracted extensive attention in recent years. A number of deep learning methods, ensure that similar examples are mapped close each other dissimilar farther apart, have been proposed construct effective structures for loss functions shown promising results. In this paper, different from the approaches on structures, we propose a robust SNR distance based Signal-to-Noise Ratio...
This study aims to characterise the polarized foreground emission in ELAIS-N1 field and address its possible implications for extraction of cosmological 21-cm signal from Low-Frequency Array - Epoch Reionization (LOFAR-EoR) data. We use high band antennas LOFAR image this region RM-synthesis unravel structures at Galactic latitudes. The brightness temperature detected is on average 4 K intensity covers range -10 +13rad m^-2 Faraday depth. total polarization angle show a wide morphological...
KOH-activated carbon materials prepared from biomass-derived source, cellulose and lignin, were compared. Mixtures of different ratios lignin used to partially mimic biomass as source. This allows tailoring optimizing the KOH activated by getting rid restriction intrinsic proportion in specific biomass. The results indicate that use a more porous structure, whereas leads graphite structure. material (CL1) blend with mass ratio 1 : exhibits high surface area 2000.39 m2 g-1 , TEABF4 /ACN...
Distributed deep learning has become an essential technique for accelerating learning, but its performance is often influenced by the heterogeneous computing nodes and communication networks within distributed platform. Due to high costs of practical deployment running it almost impossible researchers optimize training strategy tasks in real-world environments. In this paper, we propose HeterSim, a simulator specifically designed platforms. HeterSim enables flexible configuration node...
Experience-driven networking has emerged as a new and highly effective approach for resource allocation in complex communication networks. Deep Reinforcement Learning (DRL) been shown to be useful technique enabling experience-driven networking. In this paper, we focus on practical fundamental problem networking: when network configurations are changed, how train DRL agent effectively quickly adapt the environment. We present an Actor-Critic-based Transfer learning framework Traffic...
Hybrid optical-electrical switching based data center network (HOE-DCN) has been regarded as a promising architecture for the next generation (DCN). To achieve traffic optimization, main superiority of HOE-DCN is its capability to offload long-lived `elephant' flows by optical interconnections, and transmit latency-sensitive `mice' electrical switching. However, most previous works identify schedule according fixed flow size threshold, which can hardly handle highly dynamic conditions in...
The performance of routing strategies on complex networks can be characterized by two measurements, i.e., the traffic capacity and average packets travel time. By efficiently synthesizing degree dynamic queue length nodes, we propose global hybrid (GH) strategy. It achieve higher capacity, as well shorter packets, time compared with state-of-the-art (GD) strategy efficient (ER) Moreover, such superiority maintained through information corresponding paths are updated periodically. simulation...
Big data applications in Hadoop usually cause heavy bandwidth demand and network bottleneck the current center (DCN). On one hand, design of DCN does not take traffic patterns into account. other suffers from inherent performance limitations due to its solution for transmitting massive sets based on application-layer overlays, which ignores architecture DCN. To improve applications, this paper we propose OEHadoop, a modified is built by co-designing with hybrid optical electrical network....
All optical or hybrid electrical/optical switching data center network (DCN) architectures are capable of adapting the optical-layer topology to various traffic demands based on flexibility technology. The reconfiguration strategies in previous research simply attempt reconfigure circuit (OCS) links offload heavier rack-to-rack traffic. However, due complex multi-layer interconnections and patterns DCNs, OCS will not only between its source destination racks, but also forward that belongs...
The traffic flows in parallel computing systems show clustered, correlative nature and the are always latency-sensitive. These have been abstracted as "Coflow" to pursue overall optimization. Concurrent Coflows on network very novel patterns. On other hand, multiple optical interconnection architectures proposed enable adaption topology reconstructions. Nevertheless, reconstruction strategies application-agnostic, their optimization objective of performance cannot meet Coflow demand. In...
Regular and adequate use of broad-spectrum sunscreen has been proven to offer significant protection against acute Ultraviolet-induced photodamage, photoaging, immunosuppression, the development skin tumors. However, concerns regarding safety standardized sunscreens persist, including potential allergenicity irritability certain organic sunscreens, impact systemic absorption on endocrine system, effect vitamin D synthesis absorption, environmental implications. Special caution is advised...
We proposed an algorithm named OCBridge to effectively reconfigure the topology of optical data center network. Simulation results show it can reduce hotspots and achieve better throughput in networks compared with relative proposals.
The non-line-of-sight single scattering channel model is significantly simplified by means of a closed-form expression. Based on such an approximation, the relationship between visibility and path loss ultraviolet communication system analyzed, result accordance with that obtained from original model. Moreover, simulations bit-error-rate indicate has lowest value data rates have highest under fixed BER requirements when critical point. which meet different are in same order magnitude at...