A5029228290- Atmospheric chemistry and aerosols
- Atmospheric Ozone and Climate
- Speech Recognition and Synthesis
- Air Quality and Health Impacts
- Atmospheric and Environmental Gas Dynamics
- Air Quality Monitoring and Forecasting
- Music and Audio Processing
- Speech and Audio Processing
- Natural Language Processing Techniques
- Topic Modeling
- Vehicle emissions and performance
- Cellular and Composite Structures
- Water Quality Monitoring and Analysis
- Music Technology and Sound Studies
- Additive Manufacturing and 3D Printing Technologies
- COVID-19 impact on air quality
- Advanced Theoretical and Applied Studies in Material Sciences and Geometry
- Child Development and Digital Technology
- COVID-19 and Mental Health
- Scientific Computing and Data Management
- Sustainable Development and Environmental Policy
- Guidance and Control Systems
- Atmospheric aerosols and clouds
- Environmental Quality and Pollution
- Per- and polyfluoroalkyl substances research
Georgia Institute of Technology
2017-2024
Didi Chuxing (China)
2021
Microsoft (United States)
2021
Peking University
2013-2019
Beijing Foreign Studies University
2016-2017
Using new satellite observations and atmospheric inverse modeling, we report methane emissions from the Permian Basin, which is among world's most prolific oil-producing regions accounts for >30% of total U.S. oil production. Based on measurements May 2018 to March 2019, natural gas production are estimated be 2.7 ± 0.5 Tg a-1, representing largest flux ever reported a oil/gas-producing region more than two times higher bottom-up inventory-based estimates. This magnitude 3.7% gross extracted...
Since its first confirmed case at the end of 2019, COVID-19 has become a global pandemic in three months with more than 1.4 million cases worldwide, as early April 2020. Quantifying changes pollutant emissions due to and associated governmental control measures is crucial understand impacts on economy, air pollution, society. We used WRF-GC model tropospheric NO2 column observations retrieved by TROPOMI instrument derive top-down NOx emission change estimation between periods: P1 (January...
By 15 April 2020, more than 1.5 billion students worldwide experienced school closures in an effort to slow the spread of a novel coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), during disease 2019 (COVID-19) pandemic. These interruptions formal in-person educational experiences caused adverse consequences on school-age children's academic outcomes. Using pre-existing database, we calculated changes reading ability without education (i.e., summer months). The...
Photolysis of oxygenated volatile organic compounds (OVOCs) produces a primary source free radicals, including OH and inorganic peroxy radicals (HO2 RO2), consequently increasing photochemical ozone production. The amplification radical cycling through OVOC photolysis provides an important positive feedback mechanism to accelerate large production OVOCs near the surface helps promote photochemistry in whole boundary layer. This amplifier effect is most significant regions with high nitrogen...
Abstract. Recent studies have revealed a significant influx of anthropogenic aerosol from South Asia to the Himalayas and Tibetan Plateau (TP) during pre-monsoon period. In order characterize chemical composition, sources, transport processes in this area, we carried out field study June 2015 by deploying suite online instruments including an Aerodyne high-resolution time-of-flight mass spectrometer (HR-AMS) multi-angle absorption photometer (MAAP) at Nam Co station (90∘57′ E, 30∘46′ N; 4730...
Abstract In this study, we evaluated the variations of air quality in Lanzhou, a typical city Northwestern China impacted by COVID‐19 lockdown. The mass concentration and chemical composition non‐refractory submicron particulate matter (NR‐PM 1 ) were determined high‐resolution aerosol spectrometer during January‐March 2020. NR‐PM dropped 50% from before to control period. five components (sulfate, nitrate, ammonium, chloride, organic [OA]) all decreased period with biggest decrease observed...
Abstract Gas flaring is a commonly used practice in the oil and gas sector that leads to key air pollutant greenhouse emissions. Here we use multipollutant (NO 2 , SO ) satellite observations from 2005 2017 quantify activity Mexico's offshore production cluster, which produces ~50–70% of country's among world's largest fields. We estimate annual flared volume ranging 5.5 20 × 10 9 m 3 over Mexican corresponding >40% associated production, significantly larger than for instance United...
Abstract. With the improved spatial resolution of Ozone Monitoring Instrument (OMI) over earlier instruments and more than 10 years service, tropospheric NO2 retrievals from OMI have led to many influential studies on relationships between socioeconomic activities NOx emissions. Previous shown that data show different relative trends compared in situ measurements. However, sources discrepancies need further investigations. This study focuses how appropriately compare derived We retrieve...
Abstract. Long-range transport followed by deposition of black carbon on glaciers Tibet is one the key issues climate research as it induces changes radiative forcing and subsequently impacting melting glaciers. The mechanism, however, not well understood. In this study, we use short-lived reactive aromatics proxies to diagnose pollutants Tibet. situ observations across Tibetan Plateau are analyzed using a regional chemistry model. model performance current emission inventories over region...
By 15 April 2020, more than 1.5 billion students worldwide experienced school closures in an effort to slow the spread of a novel coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), during disease 2019 (COVID-19) pandemic. These interruptions formal in-person educational experiences caused adverse consequences on school-age children’s academic outcomes. Using pre-existing database, we calculated changes reading ability without education (i.e., summer months). The...
Abstract. Emission datasets of nitrogen oxides (NOx) at high horizontal resolutions (e.g., 0.05∘×0.05∘) are crucial for understanding human influences fine scales, air quality studies, and pollution control. Yet high-resolution emission data often missing or contain large uncertainties especially the developing regions. Taking advantage long-term satellite measurements dioxide (NO2), here we develop a computationally efficient method estimating NOx emissions in major urban areas 0.05∘×0.05∘...
Abstract Measurement‐model comparisons for various atmospheric species have been widely used evaluating the reliability of model simulations, yet such are quite limited isoprene due to lack systematic observations over a large scale. Here, we collected ambient air samples concurrently at 20 sites in China during 2012–2014 and compared observed mixing ratios with those simulated by coupling MEGAN v2.1 3‐D Regional chEmical trAnsport Model. The average concentration all 1,731 is 0.35 ± 0.03...
Abstract An extensive set of primary and secondary pollutants was measured at a ground site in remote location the Yellow River Delta, China during Ozone Photochemistry Export from Experiment (OPECE) March to April 2018. The measurements include volatile organic compounds (VOCs), peroxyacyl nitrates (PANs), ozone (O 3 ), particulate species, nitrogen oxides (NO x SO 2 . Observed VOC mixing ratios were comparable those heavily polluted cities U.S. China. source signatures suggest strong...
The construction of an effective good speech recognition system typically requires large amounts transcribed data, which is expensive to collect. To overcome this problem, many unsupervised pretraining methods have been proposed. Among these methods, Masked Predictive Coding achieved significant improvements on various datasets with BERT-like Reconstruction loss and transformer backbone. However, aspects MPC yet be fully investigated. In paper, we conduct a further study focus three...
Nitrogen oxides (NO x =NO+NO2) play a crucial role in the formation of ozone and secondary inorganic organic aerosols, thus affecting human health, global radiation budget, climate. The diurnal spatial variations NO2 are functions emissions, advection, deposition, vertical mixing, chemistry. Their observations, therefore, provide useful constraints our understanding these factors. We employ Regional chEmical trAnsport model (REAM) to analyze observed temporal (diurnal cycles) distributions...
Recently, a variety of acoustic tasks and related applications arised. For many tasks, the labeled data size may be limited. To handle this problem, we propose an unsupervised pre-training method using Transformer based encoder to learn general robust high-level representation for all tasks. Experiments have been conducted on three kinds tasks: speech emotion recognition, sound event detection translation. All experiments shown that its own training can significantly improve performance....
This paper introduces a new open-sourced Mandarin speech corpus, called DiDiSpeech. It consists of about 800 hours data at 48kHz sampling rate from 6000 speakers and the corresponding texts. All in corpus is recorded quiet environment suitable for various processing tasks, such as voice conversion, multi-speaker text-to-speech automatic recognition. We conduct experiments with multiple tasks evaluate performance, showing that it promising to use both academic research practical application....
Abstract High ozone concentrations have become the major summertime air quality problem in China. Extensive situ observations are deployed for developing strategies to effectively control emissions of precursors, that is, nitrogen oxides (NO X = NO + 2 ) and volatile organic compounds (VOCs). The modeling analysis often makes uses dependence peak concentration on VOC emissions, because among most widely available measurements. To extract more information from regulatory observations, we...
Self-supervised visual pretraining has shown significant progress recently. Among those methods, SimCLR greatly advanced the state of art in self-supervised and semi-supervised learning on ImageNet. The input feature representations for speech tasks are both continuous, so it is natural to consider applying similar objective representation learning. In this paper, we propose Speech SimCLR, a new During training, applies augmentation raw its spectrogram. Its combination contrastive loss that...
In this study, differential game theory was applied to propose two guidance laws of a pursuer in an orbit pursuit–evasion–defense game. One conservative law maneuvering absolute safety before pursuing the evader, and other radical playing with defender only when necessary. Both enable avoid while but reduces pursuit time by moving closer defender. The pursuer, defender, evader are all spacecraft carrying three-axis thrusters that can provide independent thrust three directions. proximity...