Source-free object detection (SFOD) aims to adapt the source detector to unlabeled target domain data in the absence of source domain data. Most SFOD methods follow the same self-training paradigm using mean-teacher (MT) framework where the student model is guided by only one single teacher model. However, such paradigm can easily fall into a training instability problem that when the teacher model collapses uncontrollably due to the domain shift, the student model also suffers drastic performance degradation. To address this issue, we propose the Periodically Exchange Teacher-Student (PETS) method, a simple yet novel approach that introduces a multiple-teacher framework consisting of a static teacher, a dynamic teacher, and a student model. During the training phase, we periodically exchange the weights between the static teacher and the student model. Then, we update the dynamic teacher using the moving average of the student model that has already been exchanged by the static teacher. In this way, the dynamic teacher can integrate knowledge from past periods, effectively reducing error accumulation and enabling a more stable training process within the MT-based framework. Further, we develop a consensus mechanism to merge the predictions of two teacher models to provide higher-quality pseudo labels for student model. Extensive experiments on multiple SFOD benchmarks show that the proposed method achieves state-of-the-art performance compared with other related methods, demonstrating the effectiveness and superiority of our method on SFOD task.

Recently, there has been increasing interest in the Source-Free Domain Adaptive Object Detection task, which involves training an object detector on the unlabeled target data using a pre-trained source model without accessing the source data. Most related methods are developed from the mean-teacher framework, which aims to train the student model closer to the teacher model via a pseudo labeling manner, where the teacher model is the exponential-moving-average of the student models at different time-steps. Following this line of works, we propose a Run-and-Chase Mutual-Learning method to strengthen the interactions between the student model and the teacher model in both feature and prediction levels. In our method, the student model is optimized to run away from the teacher model at the feature level, while chasing the teacher model at the prediction level. In this way, the student model is forced to be distinguishable at different time-steps, so that the teacher model can acquire more diverse task-related information and produce higher-accuracy pseudo labels. As the training goes, the student and teacher models are updated iteratively and promoted mutually, which can prevent the model collapse problem. Extensive experiments are conducted to validate the effectiveness of our method.

Time evolving meteorological variables of daily observations from multiple sources can be represented as simple 3-node tensors. Powerful analytic tools such as canonical polyadic (cp) tensor decomposition are available to factor streaming tensors into tensor approximations associated with dominant components. We show the potential of tensor decomposition for meteorological analysis by presenting multi-year experiments for inferring Planetary Boundary Layer Heights (PBLH) by employing cp tensor decomposition analytics applied to streaming ceilometer base Lidar data sets. The experiment compares the PBLH obtained from the 3 nodes tensor consisting of 5 minutes streams of hourly Lidar based aerosol backscatter profiles each day from 3 distinctly located ceilometers distributed along the East coast for a 1.5 years period from May 7, 2021-June 30, 2022. We compare the PBLH derived from the ceilometers with those inferred from a WRF model simulation for the same period with two nested intervals over the CONUS at 2 deg. and 0.9 deg., interpolated to the same locations as the ceilometers. The WRF model integration is forced by the interpolated hourly NOAA GFS reanalysis data set and annual averaged aerosols distributions specified geographically over the CONUS grid. We present graphs of the time dependence of the dominant tensor components of PBLH as well as the error of the dominant tensors for various assumptions of tensor rank. These dominant components show the correlations of PBLH over 24 hours over multiple regions.

Oceanic lidar has emerged as a strong technology for oceanic three-dimensional remote sensing. However, most existing oceanic lidars are bulky and high-power consumption, thus difficult to enable underwater operation. Here we present a compact single-photon lidar system for long-range underwater measurement. A single-photon detector was adopted to achieve a high signal-to-noise ratio. Benefiting from the single-photon sensitivity in detection, long-range active detection was realized with a low pulse energy laser at 1 μJ and a small-aperture coupler at 12 mm. Moreover, a narrow linewidth picosecond fiber laser with high repetition rate was employed to guarantee a high spatial resolution and high update rate. A fiber-connected configuration was specially designed for the miniaturized and robust structure in an optical receiver. In an experimental demonstration, the profile of backscattered signal from clean water was obtained over 70 m with high spatial-temporal resolution to demonstrate the capability of this lidar system. The maximum detection distance of the single-photon lidar reaches ~3.6/Kd(Kd, diffuse attenuation coefficient) for waterbody and up to 5.5/Kdfor a hard target. Furthermore, it exhibits a high update rate capability and realizes the localization and quantification of underwater bubbles up to 26 m away at a high update rate of 100 Hz. These results indicate its potential in a variety of applications including remote sensing of marine biogeochemical parameters, the quantification of seabed gas emissions, and long-range underwater imaging.

This paper proposes a federated, fair, and fast k-means algorithm (F3KM) to solve the fair clustering problem efficiently in scenarios where data cannot be shared among different parties. The proposed algorithm decomposes the fair k-means problem into multiple subproblems and assigns each subproblem to a client for local computation. Our algorithm allows each client to possess multiple sensitive attributes (or have no sensitive attributes). We propose an in-processing method that employs the alternating direction method of multipliers (ADMM) to solve each subproblem. During the procedure of solving subproblems, only the computation results are exchanged between the server and the clients, without exchanging the raw data. Our theoretical analysis shows that F3KM is efficient in terms of both communication and computation complexities. Specifically, it achieves a better trade-off between utility and communication complexity, and reduces the computation complexity to linear with respect to the dataset size. Our experiments show that F3KM achieves a better trade-off between utility and fairness than other methods. Moreover, F3KM is able to cluster five million points in one hour, highlighting its impressive efficiency.

In the ongoing evolution of the OceanBase database system, it is essential to enhance its adaptability to small-scale enterprises. The OceanBase database system has demonstrated its stability and effectiveness within the Ant Group and other commercial organizations, besides through the TPC-C and TPC-H tests. In this paper, we have designed a stand-alone and distributed integrated architecture named Paetica to address the overhead caused by the distributed components in the stand-alone mode, with respect to the OceanBase system. Paetica enables adaptive configuration of the database that allows OceanBase to support both serial and parallel executions in stand-alone and distributed scenarios, thus providing efficiency and economy. This design has been implemented in version 4.0 of the OceanBase system, and the experiments show that Paetica exhibits notable scalability and outperforms alternative stand-alone or distributed databases. Furthermore, it enables the transition of OceanBase from primarily serving large enterprises to truly catering to small and medium enterprises, by employing a single OceanBase database for the successive stages of enterprise or business development, without the requirement for migration. Our experiments confirm that Paetica has achieved linear scalability with the increasing CPU core number within the stand-alone mode. It also outperforms MySQL and Greenplum in the Sysbench and TPC-H evaluations.

Domain generalization (DG) is a fundamental yet very challenging research topic in machine learning. The existing arts mainly focus on learning domain-invariant features with limited source domains in a static model. Unfortunately, there is a lack of training-free mechanism to adjust the model when generalized to the agnostic target domains. To tackle this problem, we develop a brand-new DG variant, namely Dynamic Domain Generalization (DDG), in which the model learns to twist the network parameters to adapt to the data from different domains. Specifically, we leverage a meta-adjuster to twist the network parameters based on the static model with respect to different data from different domains. In this way, the static model is optimized to learn domain-shared features, while the meta-adjuster is designed to learn domain-specific features. To enable this process, DomainMix is exploited to simulate data from diverse domains during teaching the meta-adjuster to adapt to the agnostic target domains. This learning mechanism urges the model to generalize to different agnostic target domains via adjusting the model without training. Extensive experiments demonstrate the effectiveness of our proposed method. Code is available: https://github.com/MetaVisionLab/DDG

We have designed and developed OceanBase, a distributed relational database system from the very basics for a decade. Being a scale-out multi-tenant system, OceanBase is cross-region fault tolerant, which is based on the shared-nothing architecture. Besides sharing many similar goals with alternative distributed DBMS, such as horizontal scalability, fault-tolerance, etc., our design has been driven by the demands of typical RDBMS compatibility as well as both on-premise and off-premise deployments. OceanBase has fulfilled its design goal. It implements the salient features of certain mainstream classical RDBMS, and most applications on them can run on OceanBase, with or without a few minor modifications. Tens of thousands of OceanBase servers have been deployed in Alipay.com as well as many other commercial organizations. It has also successfully passed the TPC-C benchmark test and seized the first place with more than 707 million tpmC. This paper presents the goals, design criteria, infrastructure, and key components of OceanBase including its engines for storage and transaction processing. Further, it details how OceanBase achieves the above leading TPC-C benchmark in a distributed cluster with more than 1,500 servers from 3 zones. It also describes lessons what we have learnt in building OceanBase for more than a decade.

Multiple satellites and ground based Lidar instruments are becoming increasingly available in providing near daily coverage of aerosol distributions over the US. Aerosols are very small particles, both liquid and solid, arising from both natural and human activities. Atmospheric aerosols directly influence incoming solar radiation by scattering and reflection and indirectly by serving as nuclei for cloud formation of rain, ice and graupel. In addition, aerosols are emitted in the surface layer and pollute the Planetary Boundary Layer (PBL) affecting human health. PBL height (PBLH) determines the vertical extent to which most pollutants effectively disperse and is a key variable in air quality studies. Incorporating aerosol emissions with microphysics and chemistry parameterizations can significantly impact sub-seasonal weather forecasts directly through model radiative transfer parametrization and indirectly can influences the cloud formation, precipitation and the onset of deep convective storms through microphysical processes. Cloud resolving regional forecast models at 0.01 0 (~ 1km) spatial resolution employing such parameterizations will require a fully dedicated petascale computer to effect sub-seasonal forecasts on a weekly basis. In this paper, we initially study the potential of Neural Architecture Searches (NAS) to accurately emulate aerosol and microphysical parameterizations at coarse spatial resolutions to assess potential speed ups over conventional computational approaches. We conduct two experiments, a coarse 1-year WRF model simulation conducted at 2.5° over N. America and 0.8° over Eastern US. with 30 levels including CHEM, Bulk 2-moment microphysics to explore the feasibility of using NAS nets for sub seasonal forecasts and a second short term 6-day high resolution simulation at 0.09°with a single domain over the US and a 0.03° inner domain over eastern US. The problem running the WRF model at cloud resolving resolutions (for example 1km - 4km) is that it will require large computational times and resources. By first studying a coarse WRF experiment, we only need 10 min./sim day on 18 nodes or ~ 14 hrs/sub-seasonal forecast compared with 4hrs /day or approximately 5 dedicated days /sub-seasonal prediction. Our 6 day forecast study is used to provide computing time estimate for testing scalability and takes 20 min./day to run the 0.09° sub-seasonal simulation when using chemical and microphysics components at 3km. We employ the Auto-Keras NAS model, which accurately emulates the 1 year coarse experiment as well as gains an expected 10x speedup. This NAS deep neural net model will be used to infer the aerosol types, sizes and distributions to enable microphysics inferences and realistic mixed boundary layer heights with expected speedups of an order of magnitude or more.

Inspired by the importance of Planetary Boundary Layer Heights (PBLH) for inferring Air Quality assessments and the disappointing performance of current weather forecasts of PBLH, this paper presents the proposed impact of using a Machine Learning derived PBLH (ML-PBLH) using ground-based Ceilometer observing systems. The PBLH is vital in air pollution studies to determine the extent of vertical mixing of pollutants (e.g., particles, trace gases, etc.). On the other hand, model forecasts of the nighttime collapse of the PBLH close to the surface, grossly underestimate the actual observations. We propose using machine learning methods, employing deep neural networks for denoising and image segmentation to detect PBLH from the Ceilometer backscatter signal from the LIDAR observations. The ML-PBLH detection algorithm shows promising early results when compared with conventional ground-based Ceilometer retrieval methods such as Wavelet methods used for LIDAR identification of PBLH under normal conditions. The use of conventional methods as well as the wavelet method for LIDAR backscatter retrievals is limited in the presence of dense clouds. However, the proposed machine learning approach is able to infer PBLH even under such conditions.

Water is a fundamental resource in our daily lives and for social development. Currently, water resources are under stressed conditions in north China due to the ongoing growth of demand for water and the frequent occurrences of natural disasters in this area. This problem is further intensified by the deterioration of water quality, which is a side effect of economic development, industrial expansion, and agricultural growth. These problems interconnect with each other, leading to a variety of conflicts over water resources. To resolve such conflicts, a two-person game-theory based model is developed for water resource management in the trans-boundary regions of a river basin. The model is tested on a river basin within this area of China in which water resources are contested. Through the developed model, potential policy alternatives can be generated for water resource management and pollution control within the context of a river basin. Cooperative and/or competitive mechanisms can thus be formulated through competition and cooperation over issues of water quantity and quality among different levels of jurisdictions in the same river basin. Then, Nash Solution is used to solve scenarios of competition and cooperation between the upstream and downstream based on the integration of optimization and simulation models. The developed model and the solution process is then applied to resolve water conflicts between Beijing and Hebei province, which are in the upstream and downstream of the Guanting Reservoir Basin (GRB), respectively, representing two major stakeholders in the contest over water quantity and quality. Non-cooperative and cooperative scenarios for the two-player game scenarios are developed to investigate and compare potential economic benefits of these scenarios. Results indicate that the developed model can be effectively and flexibly employed to resolve water conflicts between upstream and downstream stakeholders in the same river basin. The results can also be used to help formulate economic compensation mechanisms within these jurisdictions.

The Longyangxia Reservoir lies in Qinghai-Tibet Platean, upstream of Yellow River. Some ecological problems, such as drought, water shortage; soil erosion and meadow degeneration are major factors to threaten the safety of water resources. The reservoir usually contained great area and its effects are complex and uncertain. Some effects can be shown after a few years. Thus, it is very hard to assess the eco-environmental impacts of large reservoir. Fortunately, there are few human activities in this area except for reservoir building. The changes of ecosystem are mainly caused by the water resources development. By analyzing these changes, the eco-environmental impacts of the reservoir can be assessed clearly. In this study, the measures of landscape ecology and remoting sense are used to analyze the structural changes of local ecosystem and some factors, which impact the stability of ecosystem, are discussed. Based on the interpretation of the Landsat-TM data around the Longyangxia Reservoir, which includes two digital images taken in 1986 and 2001, the landscape structure was studied by analyzing landscape pattern indices and transfer matrix of landscape. The results showed that grassland and unused lands were the main landscape types around the Longyangxia Reservoir, and the proportion of rivers, ponds and lakes, farmland and building zone had increased as the construction of the Longyangxia Reservoir and other artificial disturbance from 1986 to 2001. The landscape pattern had a trend of heterogeneity and had not a trend of fragmentation. From 1986 to 2001, parts of woodland had changed to grassland and parts of grassland and farmland had changed to unused lands, which indicate there is a trend of desertification. In this way, the environmental impacts of Longyangxia reservoir are assessed. Finally, the possibility of this method to use in eco-environmental impact assessment of reservoir is also discussed in this paper