A5049544852- Plasma Applications and Diagnostics
- Water Treatment and Disinfection
- Advanced oxidation water treatment
- Electrohydrodynamics and Fluid Dynamics
- Effects and risks of endocrine disrupting chemicals
- Advanced Memory and Neural Computing
- Advanced Neural Network Applications
- Environmental remediation with nanomaterials
- Catalytic Processes in Materials Science
- Machine Learning and ELM
- Toxic Organic Pollutants Impact
- Nanomaterials for catalytic reactions
- Electrochemical Analysis and Applications
- Topic Modeling
- Pharmaceutical and Antibiotic Environmental Impacts
- IoT and Edge/Fog Computing
- Semiconductor materials and devices
- Constructed Wetlands for Wastewater Treatment
- Advanced Malware Detection Techniques
- Privacy-Preserving Technologies in Data
- Water Quality Monitoring and Analysis
- Personal Information Management and User Behavior
- Magnetic Properties and Synthesis of Ferrites
- Aerosol Filtration and Electrostatic Precipitation
- Analytical chemistry methods development
Jiangsu University
2016-2025
Suzhou University of Science and Technology
2024-2025
Beijing University of Posts and Telecommunications
2025
Nanjing Forestry University
2024
State Key Laboratory of Networking and Switching Technology
2023
The Synergetic Innovation Center for Advanced Materials
2016
Large foundation models, including large language models (LLMs), vision transformers (ViTs), diffusion, and LLM-based multimodal are revolutionizing the entire machine learning lifecycle, from training to deployment. However, substantial advancements in versatility performance these offer come at a significant cost terms of hardware resources. To support growth scalable environmentally sustainable way, there has been considerable focus on developing resource-efficient strategies. This survey...
Large Language Models (LLMs) such as GPTs and LLaMa have ushered in a revolution machine intelligence, owing to their exceptional capabilities wide range of learning tasks. However, the transition LLMs from data centers edge devices presents set challenges opportunities. While this shift can enhance privacy availability, it is hampered by enormous parameter sizes these models, leading impractical runtime costs. In light considerations, we introduce EdgeMoE, first on-device inference engine...
Quinoline is widely used in the production of drugs as a highly effective insecticide, and its derivatives can also be to produce dyes. It has teratogenic carcinogen wildlife humans once entering into aquatic environment. In this study, degradation mechanism quinoline drinking water by strong ionization dielectric barrier discharge (DBD) low-temperature plasma with large volume was explored. High concentration hydroxyl radical (·OH) (0.74 mmol l−1) ozone (O3) (58.2 mg produced strongly...
DNNs are ubiquitous on edge devices nowadays. With its increasing importance and use cases, it's not likely to pack all into device memory expect that each inference has been warmed up. Therefore, cold inference, the process read, initialize, execute a DNN model, is becoming commonplace performance urgently demanded be optimized. To this end, we present NNV12, first on-device engine optimizing inference. NNV12 built atop three novel optimization knobs: selecting proper kernel (i.e., operator...
The present study focused on the degradation of sulfamethoxazole (SMX) aqueous solution and toxicity processing by dielectric barrier discharge (DBD) activated persulfate (PS). effects input voltage, frequency, duty cycle, PS dosage ratio SMX efficiency were measured. Based results Response Surface Methodology (RSM), reached 83.21% which is 10.54% higher than that without PS, kinetic constant was 0.067 min
The bactericidal effect on the representative type of Gram-negative Escherichia coli (E. coli) and Gram-positive Bacillus subtilis in drinking water was investigated this paper by using dielectric barrier discharge (DBD) advanced oxidation technology. sterilizing rates under different conditions reaction time t, input voltage V, pH value, initial concentration bacteria C0 were to figure out optimum sterilization conditions. Our observations comparisons cell morphology alteration scanning...
The degradation mechanism of dimethyl phthalate (DMP) in the drinking water was investigated using strong ionization discharge technology this study. Under optimized condition, efficiency DMP up to 93% 60 min. A series analytical techniques including high-performance liquid chromatography, chromatography mass spectrometry, total organic carbon analyzer and ultraviolet–visible spectroscopy were used It found that a high concentration ozone (O3) produced by dielectric barrier reactor 74.4 mg...
A vortex electrostatic precipitator (VEP) forms a flow field within by means of the vertical staggered layout double-vortex collecting plate facing direction gas flow. The ion concentrations can be significantly increased. Correspondingly, charging and coagulation rates fine particles particle migration velocity are improved VEP. Since it effectively collect reduce size, VEPs represent new type with great application potential. In this work change curve external voltage, velocity, row...
Nitrobenzene (NB) is toxic and carcinogenic aromatic compound widely used in several industries which ultimately found their effluents. In this work, dielectric barrier discharge (DBD) reactor was employed for the degradation of nitrobenzene aqueous solution. Active species like O3 •OH produced by DBD were mixed with water degraded NB. The results indicated that lower NB concentrations slightly acidic conditions high voltage ranges showed optimum efficiencies. Moreover, impacts active...
Nonylphenol (NP) is a typical environmental endogenous disrupter with low concentration and high toxicity. This paper describes the mechanism of NP degradation in solution by strong ionization dielectric barrier discharge (SIDBD). Furthermore, performance SIDBD was tested changing equipment voltage, initial aqueous solution, pH, inorganic ions. Degradation pathways were detected using high-performance liquid chromatography-mass spectrometer. The biological effects assessed detecting...
The growing capabilities of large language models in natural understanding significantly strengthen existing agentic systems. To power performant on-device mobile agents for better data privacy, we introduce DroidCall, the first training and testing dataset accurate Android intent invocation. With a highly flexible reusable generation pipeline, constructed 10k samples DroidCall. Given task instruction language, small such as Qwen2.5-3B Gemma2-2B fine-tuned with DroidCall can approach or even...
This paper is focused on the situation that by-product ammonium sulfite produced in flue gas desulfurization of ammonia process,which carried ozone oxidation experiments.In reaction container, studied effect factors as temperature, pH value, flow and initial concentration rate oxidation.Then, conducted kinetics experiment under condition only use ozone.The results showed are both simple first-order response reactions.
DNNs are ubiquitous on edge devices nowadays. With its increasing importance and use cases, it's not likely to pack all into device memory expect that each inference has been warmed up. Therefore, cold inference, the process read, initialize, execute a DNN model, is becoming commonplace performance urgently demanded be optimized. To this end, we present NNV12, first on-device engine optimizes for NNV12 built atop 3 novel optimization knobs: selecting proper kernel (implementation) operator,...
Based on the egocentric network theory and visualization analysis method, a communication framework is proposed, combined method of dielectric barrier discharge (DBD) non-equilibrium plasma activated carbon fiber (ACF) used to remove 3,4-dichloroaniline (3,4-dichloroaniline) in water. 4-DCA), influence factors such as power, air flow, initial concentration solution, conductivity, pH value, dosage treatment efficiency was investigated. A data platform constructed. The can interactively...
A large amount of volatile organic compounds (VOCs) produced by industry have caused serious environmental pollution. In this paper, the removal effect simulated xylene strong ionization dielectric barrier discharge (DBD) plasma at atmospheric pressure and its degradation mechanism pathway were studied. The gas residence time, initial concentration was results showed that higher voltage an increase in power, with voltage, ozone nitrogen oxide reactor increased. efficiency decreased from...