Shandong Academy of Sciences

Active learning (AL) attempts to maximize a model’s performance gain while annotating the fewest samples possible. Deep learning (DL) is greedy for data and requires a large amount of data supply to optimize a massive number of parameters if the model is to learn how to extract high-quality features. In recent years, due to the rapid development of internet technology, we have entered an era of information abundance characterized by massive amounts of available data. As a result, DL has attracted significant attention from researchers and has been rapidly developed. Compared with DL, however, researchers have a relatively low interest in AL. This is mainly because before the rise of DL, traditional machine learning requires relatively few labeled samples, meaning that early AL is rarely according the value it deserves. Although DL has made breakthroughs in various fields, most of this success is due to a large number of publicly available annotated datasets. However, the acquisition of a large number of high-quality annotated datasets consumes a lot of manpower, making it unfeasible in fields that require high levels of expertise (such as speech recognition, information extraction, medical images, etc.). Therefore, AL is gradually coming to receive the attention it is due. It is therefore natural to investigate whether AL can be used to reduce the cost of sample annotation while retaining the powerful learning capabilities of DL. As a result of such investigations, deep active learning (DeepAL) has emerged. Although research on this topic is quite abundant, there has not yet been a comprehensive survey of DeepAL-related works; accordingly, this article aims to fill this gap. We provide a formal classification method for the existing work, along with a comprehensive and systematic overview. In addition, we also analyze and summarize the development of DeepAL from an application perspective. Finally, we discuss the confusion and problems associated with DeepAL and provide some possible development directions.

In modern society, traffic and transportation and the manufacturing industry and construction industries continuously release large amounts of dust and particles into the atmosphere, which can cause heavy air pollution, leading to health hazards. The haze disaster, a serious problem in developing countries such as China and India, has become one of the main issues of global environmental pollution in recent decades. Many air filtration technologies have been developed. Air filtration using electrospun fibers that intercept fine particles/volatile organic gases/bacterium is a relatively new, but highly promising, technique. Due to their interconnected nanoscale pore structures, highly specific surface areas, fine diameters, and porous structure as well as their ability to incorporate active chemistry on a nanoscale surface, electrospun fibers are becoming a promising versatile platform for air filtration. In this review, following a short introduction concerning the need for air filtration and filtration theory and mechanism, electrospun nanofibers membranes for air filtration have been highlighted, including the preparation (electrospinning process) and the parameters relevant to filtration efficacy. Additionally, various types (function) of the electrospun air filtration membranes have been classified in detail. Furthermore, their potential in the filtration of fine particles and chemical pollutants has been discussed. Finally, the challenges of their practical application and the future prospects have been summarized. Given that some advanced electrospun air filtration nanofibrous membranes exist for treating different contaminants from various types of polluted atmosphere, it is believed that they should make a significant contribution in protection against air pollution.

Deep learning has made substantial breakthroughs in many fields due to its powerful automatic representation capabilities. It has been proven that neural architecture design is crucial to the feature representation of data and the final performance. However, the design of the neural architecture heavily relies on the researchers’ prior knowledge and experience. And due to the limitations of humans’ inherent knowledge, it is difficult for people to jump out of their original thinking paradigm and design an optimal model. Therefore, an intuitive idea would be to reduce human intervention as much as possible and let the algorithm automatically design the neural architecture. Neural Architecture Search ( NAS ) is just such a revolutionary algorithm, and the related research work is complicated and rich. Therefore, a comprehensive and systematic survey on the NAS is essential. Previously related surveys have begun to classify existing work mainly based on the key components of NAS: search space, search strategy, and evaluation strategy. While this classification method is more intuitive, it is difficult for readers to grasp the challenges and the landmark work involved. Therefore, in this survey, we provide a new perspective: beginning with an overview of the characteristics of the earliest NAS algorithms, summarizing the problems in these early NAS algorithms, and then providing solutions for subsequent related research work. In addition, we conduct a detailed and comprehensive analysis, comparison, and summary of these works. Finally, we provide some possible future research directions.

Exosomes, a nano-sized subtype of extracellular vesicles secreted from almost all living cells, are capable of transferring cell-specific constituents of the source cell to the recipient cell. Cumulative evidence has revealed exosomes play an irreplaceable role in prognostic, diagnostic, and even therapeutic aspects. A method that can efficiently provide intact and pure exosomes samples is the first step to both exosome-based liquid biopsies and therapeutics. Unfortunately, common exosomal separation techniques suffer from operation complexity, time consumption, large sample volumes and low purity, posing significant challenges for exosomal downstream analysis. Efficient, simple, and affordable methods to isolate exosomes are crucial to carrying out relevant researches. In the last decade, emerging technologies, especially microfluidic chips, have proposed superior strategies for exosome isolation and exhibited fascinating performances. While many excellent reviews have overviewed various methods, a compressive review including updated/improved methods for exosomal isolation is indispensable. Herein, we first overview exosomal properties, biogenesis, contents, and functions. Then, we briefly outline the conventional technologies and discuss the challenges of clinical applications of these technologies. Finally, we review emerging exosomal isolation strategies and large-scale GMP production of engineered exosomes to open up future perspectives of next-generation Exo-devices for cancer diagnosis and treatment.

The rapid developments of the Internet of Things (IoT) and smart mobile devices in recent years have been dramatically incentivizing the advancement of edge computing. On the one hand, edge computing has provided a great assistance for lightweight devices to accomplish complicated tasks in an efficient way; on the other hand, its hasty development leads to the neglection of security threats to a large extent in edge computing platforms and their enabled applications. In this paper, we provide a comprehensive survey on the most influential and basic attacks as well as the corresponding defense mechanisms that have edge computing specific characteristics and can be practically applied to real-world edge computing systems. More specifically, we focus on the following four types of attacks that account for 82% of the edge computing attacks recently reported by Statista: distributed denial of service attacks, side-channel attacks, malware injection attacks, and authentication and authorization attacks. We also analyze the root causes of these attacks, present the status quo and grand challenges in edge computing security, and propose future research directions.

Abstract In recent decades, flexible and wearable devices have been extensively investigated due to their promising applications in portable mobile electronics and human motion monitoring. MXene, a novel growing family of 2D nanomaterials, demonstrates superiorities such as outstanding electrical conductivity, abundant terminal groups, unique layered‐structure, large surface area, and hydrophilicity, making it to be a potential candidate material for flexible and wearable devices. Numerous pioneering works are devoted to develop flexible MXene‐based composites with various functions and designed structures. Therefore, the latest progress of the flexible MXene‐based composites for wearable devices is summarized in this review, focusing on the preparation strategies, working mechanisms, performances, and applications in sensors, supercapacitors, and electromagnetic interference shielding materials. Moreover, the current challenges and future outlooks are also discussed.

Traditional camera sensors rely on human eyes for observation. However, human eyes are prone to fatigue when observing objects of different sizes for a long time in complex scenes, and human cognition is limited, which often leads to judgment errors and greatly reduces efficiency. Object recognition technology is an important technology used to judge the object’s category on a camera sensor. In order to solve this problem, a small-size object detection algorithm for special scenarios was proposed in this paper. The advantage of this algorithm is that it not only has higher precision for small-size object detection but also can ensure that the detection accuracy for each size is not lower than that of the existing algorithm. There are three main innovations in this paper, as follows: (1) A new downsampling method which could better preserve the context feature information is proposed. (2) The feature fusion network is improved to effectively combine shallow information and deep information. (3) A new network structure is proposed to effectively improve the detection accuracy of the model. From the point of view of detection accuracy, it is better than YOLOX, YOLOR, YOLOv3, scaled YOLOv5, YOLOv7-Tiny, and YOLOv8. Three authoritative public datasets are used in these experiments: (a) In the Visdron dataset (small-size objects), the map, precision, and recall ratios of DC-YOLOv8 are 2.5%, 1.9%, and 2.1% higher than those of YOLOv8s, respectively. (b) On the Tinyperson dataset (minimal-size objects), the map, precision, and recall ratios of DC-YOLOv8 are 1%, 0.2%, and 1.2% higher than those of YOLOv8s, respectively. (c) On the PASCAL VOC2007 dataset (normal-size objects), the map, precision, and recall ratios of DC-YOLOv8 are 0.5%, 0.3%, and 0.4% higher than those of YOLOv8s, respectively.

The degree of crystallinity in cellulose significantly affects the physical, mechanical, and chemical properties of cellulosic materials, their processing, and their final application. Measuring the crystalline structures of cellulose is a challenging task due to inadequate consistency among the variety of analytical techniques available and the lack of absolute crystalline and amorphous standards. Our article reviews the primary methods for estimating the crystallinity of cellulose, namely, X-ray diffraction (XRD), nuclear magnetic resonance (NMR), Raman and Fourier-transform infrared (FTIR) spectroscopy, sum-frequency generation vibrational spectroscopy (SFG), as well as differential scanning calorimetry (DSC), and evolving biochemical methods using cellulose binding molecules (CBMs). The techniques are compared to better interrogate not only the requirements of each method, but also their differences, synergies, and limitations. The article highlights fundamental principles to guide the general community to initiate studies of the crystallinity of cellulosic materials.

Differential scanning calorimetry (DSC) is a powerful tool to address some of the most challenging issues in glass science and technology, such as the nonequilibrium nature of the glassy state and the detailed thermodynamics and kinetics of glass-forming systems during glass transition, relaxation, rejuvenation, polyamorphic transition, and crystallization. The utility of the DSC technique spans across all glass-forming chemistries, including oxide, chalcogenide, metallic, and organic systems, as well as recently discovered metal-organic framework glass-forming systems. Here we present a comprehensive review of the many applications of DSC in glass science with focus on glass transition, relaxation, polyamorphism, and crystallization phenomena. We also emphasize recent advances in DSC characterization technology, including flash DSC and temperature-modulated DSC. This review demonstrates how DSC studies have led to a multitude of relevant advances in the understanding of glass physics, chemistry, and even technology.

Emotion recognition technology through analyzing the EEG signal is currently an essential concept in Artificial Intelligence and holds great potential in emotional health care, human-computer interaction, multimedia content recommendation, etc. Though there have been several works devoted to reviewing EEG-based emotion recognition, the content of these reviews needs to be updated. In addition, those works are either fragmented in content or only focus on specific techniques adopted in this area but neglect the holistic perspective of the entire technical routes. Hence, in this paper, we review from the perspective of researchers who try to take the first step on this topic. We review the recent representative works in the EEG-based emotion recognition research and provide a tutorial to guide the researchers to start from the beginning. The scientific basis of EEG-based emotion recognition in the psychological and physiological levels is introduced. Further, we categorize these reviewed works into different technical routes and illustrate the theoretical basis and the research motivation, which will help the readers better understand why those techniques are studied and employed. At last, existing challenges and future investigations are also discussed in this paper, which guides the researchers to decide potential future research directions.

Besides genome editing, CRISPR-Cas12a has recently been used for DNA detection applications with attomolar sensitivity but, to our knowledge, it has not been used for the detection of small molecules. Bacterial allosteric transcription factors (aTFs) have evolved to sense and respond sensitively to a variety of small molecules to benefit bacterial survival. By combining the single-stranded DNA cleavage ability of CRISPR-Cas12a and the competitive binding activities of aTFs for small molecules and double-stranded DNA, here we develop a simple, supersensitive, fast and high-throughput platform for the detection of small molecules, designated CaT-SMelor (CRISPR-Cas12a- and aTF-mediated small molecule detector). CaT-SMelor is successfully evaluated by detecting nanomolar levels of various small molecules, including uric acid and p-hydroxybenzoic acid among their structurally similar analogues. We also demonstrate that our CaT-SMelor directly measured the uric acid concentration in clinical human blood samples, indicating a great potential of CaT-SMelor in the detection of small molecules.

Fused filament fabrication (FFF), also known as fused deposition modeling (FDM™), is considered one of the most promising additive manufacturing (AM) methods for its versatility, reliability and affordability. First adopted by industries for professional uses such as rapid prototyping, then by the general public in recent years, FFF has gathered itself considerable attention. Nevertheless, despite key advancements in printer technologies and filament materials, the fabrication of robust, performing and functional parts for high-demanding practical applications remains a significant challenge. Due to intrinsic deficiencies, such as the presence of voids and weak layer-to-layer adhesion, FFF-printed parts are plagued by weak and anisotropic mechanical properties in contrast to their conventionally manufactured counterparts. With the increasing demand for designable porous structures in the fields of biomedicine, 4D printing and lightweight cellular composites, understanding the challenges presented by void presence has become more relevant than ever. As existing literature has reviewed the significance of interlayer bonding, this review focuses on documenting recent insights on the formation of voids by its categorization, research method and mechanism. The primary objective is to provide a comprehensive understanding of the two current primary methods of void research—quantitative analysis and imaging. Detailed discussions on the effects of feedstock and printing parameters on void formation are also presented. Lastly, this review discusses gaps in the current research and outlines unaddressed challenges regarding void formation and its relation with the mechanical performance of FFF parts.

Abstract Multifunctional architecture with intriguing structural design is highly desired for realizing the promising performances in wearable sensors and flexible energy storage devices. Cellulose nanofiber (CNF) is employed for assisting in building conductive, hyperelastic, and ultralight Ti 3 C 2 T x MXene hybrid aerogels with oriented tracheid-like texture. The biomimetic hybrid aerogels are constructed by a facile bidirectional freezing strategy with CNF, carbon nanotube (CNT), and MXene based on synergistic electrostatic interaction and hydrogen bonding. Entangled CNF and CNT “mortars” bonded with MXene “bricks” of the tracheid structure produce good interfacial binding, and superior mechanical strength (up to 80% compressibility and extraordinary fatigue resistance of 1000 cycles at 50% strain). Benefiting from the biomimetic texture, CNF/CNT/MXene aerogel shows ultralow density of 7.48 mg cm −3 and excellent electrical conductivity (~ 2400 S m −1 ). Used as pressure sensors, such aerogels exhibit appealing sensitivity performance with the linear sensitivity up to 817.3 kPa −1 , which affords their application in monitoring body surface information and detecting human motion. Furthermore, the aerogels can also act as electrode materials of compressive solid-state supercapacitors that reveal satisfactory electrochemical performance (849.2 mF cm −2 at 0.8 mA cm −2 ) and superior long cycle compression performance (88% after 10,000 cycles at a compressive strain of 30%).

Chitin, a fascinating biopolymer found in living organisms, fulfills current demands of availability, sustainability, biocompatibility, biodegradability, functionality, and renewability. A feature of chitin is its ability to structure into hierarchical assemblies, spanning the nano- and macroscales, imparting toughness and resistance (chemical, biological, among others) to multicomponent materials as well as adding adaptability, tunability, and versatility. Retaining the inherent structural characteristics of chitin and its colloidal features in dispersed media has been central to its use, considering it as a building block for the construction of emerging materials. Top-down chitin designs have been reported and differentiate from the traditional molecular-level, bottom-up synthesis and assembly for material development. Such topics are the focus of this Review, which also covers the origins and biological characteristics of chitin and their influence on the morphological and physical-chemical properties. We discuss recent achievements in the isolation, deconstruction, and fractionation of chitin nanostructures of varying axial aspects (nanofibrils and nanorods) along with methods for their modification and assembly into functional materials. We highlight the role of nanochitin in its native architecture and as a component of materials subjected to multiscale interactions, leading to highly dynamic and functional structures. We introduce the most recent advances in the applications of nanochitin-derived materials and industrialization efforts, following green manufacturing principles. Finally, we offer a critical perspective about the adoption of nanochitin in the context of advanced, sustainable materials.

Abstract Developing a scalable approach to construct efficient and multifunctional electrodes for the hydrogen evolution reaction (HER), oxygen evolution reaction (OER), and oxygen reduction reaction (ORR) is an urgent need for overall water splitting and zinc–air batteries. In this work, a freestanding 3D heterostructure film is synthesized from a Ni‐centered metal−organic framework (MOF)/graphene oxide. During the pyrolysis process, 1D carbon nanotubes formed from the MOF link with the 2D reduced graphene oxide sheets to stitch the 3D freestanding film. The results of the experiments and theoretical calculations show that the synergistic effect of the N‐doped carbon shell and Ni nanoparticles leads to an optimized film with excellent electrocatalytic activity. Low overpotentials of 95 and 260 mV are merely needed for HER and OER, respectively, to reach a current density of 10 mA cm −2 . In addition, a high half‐wave potential of 0.875 V is obtained for the ORR, which is comparable to that of Pt/RuO 2 and ranks among the top of non‐noble‐metal catalysts. The use of an “all‐in‐one” film as the electrode leads to excellent performance of the homemade water electrolyzer and zinc–air battery, indicating the potential of the film for practical applications.

This study focused on the distribution, combined pollution, potential source and risk assessment of 17 antibiotics in an aquaculture ecosystem surrounding the Yellow sea, North China. Antibiotics were detected in various matrices (seawater, sediment/biofilm, organism and feed) in different aquaculture modes (greenhouse and outdoor aquaculture) during the wet and dry seasons in coastal areas of Shandong province. The innovation points of the study were as follows: (1) To the best of our knowledge, this study was one of the few to investigate the occurrence and distribution of antibiotics in mariculture environments along the Yellow Sea coast; (2) Biofilms, a focus of the study, might act as a sink for antibiotics in the aquaculture ecosystem; and (3) The correlation of heavy metals and antibiotic concentrations was proved, which could correspondingly be used as an indicator for antibiotic concentrations in the studied area. The levels of antibiotics in water were observed to be relatively low, at the ng/L level. Trimethoprim was the most prevalent antibiotic, and was detected in all water samples. Oxytetracycline was detected at high concentrations in biofilms (up to 1478.29 ng/g). Moreover, biofilms exhibited a higher antibiotic accumulation capacity compared to sediments. Concentrations of oxytetracycline and doxycycline were high in feed, while other antibiotics were almost undetected. Tetracycline was widely detected and the concentration of enrofloxacin was highest in organisms. Correlation analysis demonstrated that environmental parameters and other coexisting contaminants (e.g. heavy metals) significantly affected antibiotic concentrations. In addition, the concentration of Zn was significantly correlated with the total antibiotic concentration and was proportional to several antibiotics in water and sediment (biofilm) samples (p < 0.01). High Mn concentrations were closely related to total and individual (e.g. sulfadiazine, sulfamethazine and enrofloxacin) antibiotic levels, which may result in the combined contamination of the environment. Antibiotics in estuaries and groundwater generally originated from aquaculture wastewater and untreated/treated domestic sewage. Most of the detected antibiotics posed no risk to the environment. Ciprofloxacin and enrofloxacin found in water may present high ecological and resistance risks, while the two antibiotics observed to accumulate in fish may pose a considerable risk to human health through diet consumption. All antibiotics detected in seafood were lower than the respective maximum residue limits. This study can act as a reference for the government for the determination of antibiotic discharge standards in aquaculture wastewater and the establishment of a standardized antibiotic monitoring and management system.

The successful use of electrochemiluminescence (ECL) in immunoassay for clinical diagnosis requires development of novel ECL signal probes. Herein, we report lanthanide (Ln) metal–organic frameworks (LMOFs) as ECL signal emitters in the ECL immunoassay. The LMOFs were prepared from precursors containing Eu (III) ions and 5-boronoisophthalic acid (5-bop), which could be utilized to adjust optical properties. Investigations of ECL emission mechanisms revealed that 5-bop was excited with ultraviolet photons to generate a triplet-state, which then triggered Eu (III) ions for red emission. The electron-deficient boric acid decreased the energy-transfer efficiency from the triplet-state of 5-bop to Eu (III) ions; consequently, both were excited with high-efficiency at single excitation. In addition, by progressively tailoring the atomic ratios of Ni/Fe, NiFe composites (Ni/Fe 1:1) were synthesized with more available active sites, enhanced stability, and excellent conductivity. As a result, the self-luminescent europium LMOFs displayed excellent performance characteristics in an ECL immunoassay with a minimum detectable limit of 0.126 pg mL–1, using Cytokeratins21-1 (cyfra21-1) as the target detection model. The probability of false positive/false negative was reduced dramatically by using LMOFs as signal probes. This proposed strategy provides more possibilities for the application of lanthanide metals in analytical chemistry, especially in the detection of other disease markers.

Abstract Lithium–sulfur (Li–S) batteries with a theoretical energy density of 2567 Wh kg −1 are very promising next‐generation energy storage systems, but suffer from the insulativity of sulfur and Li 2 S, the shuttle effect due to the dissolution and migration of polysulfides, and the lithium dendrite issue. MXenes, a family of 2D transition metal carbides/nitrides, which have metallic conductivity, structural variety, strong chemical adsorption ability to polysulfides, effective catalytic effect for fast kinetics, and inducing effect for uniform growth of Li, exhibit promising potential for high‐performance Li–S batteries. In this review, the recent progress and achievements of MXene‐based Li–S batteries are summarized, including the use of MXenes in sulfur cathode, interlayer between cathode and separator, and Li anode. The architecture construction and chemical modification of MXenes, as well as hybridization with other materials are demonstrated. The enhancement on electrochemical performance and the related mechanisms of MXenes and MXene‐based composites are discussed. Finally, challenges and perspectives of MXenes for Li–S battery application are also given.

Microbial fuel cell (MFC) is an environmentally friendly technology for electricity harvesting from a variety of substrates. Microorganisms used as catalysts in the anodic chamber, which are termed as electricigens, play a major role in the operation of MFCs. This review provides an introduction to the currently identified electricigens on their taxonomical groups and electricity producing abilities. The mechanism of electron transfer from electricigens to electrode is highlighted. The performances of pure culture and mixed communities are compared particularly. It has been proved that the electricity generation capacity and the ability to adapt to the complex environment of MFC systems constructed by pure microbial cultures are less than the systems constructed by miscellaneous consortia. However, pure cultures are useful to clarify the electron transfer mechanism at the microbiological level and further reduce the complexity of mixed communities. Future research trends of electricigens in MFCs should be focused on screening, domestication, modification and optimization of multi-strains to improve their electrochemical activities. Although the MFC techniques have been greatly advanced during the past few years, the present state of this technology still requires to be combined with other processes for cost reduction.

Abstract The synthesis of fully conjugated sp 2 ‐carbon covalent organic frameworks (COF) is extremely challenging given the difficulty of the formation of very stable carbon‐carbon double bonds (‐C=C‐). Here, we report the successful preparation of a 2D COF (TP‐COF) based on triazine as central planar units bridged by sp 2 ‐carbon linkers through the ‐C=C‐ condensation reaction. High‐resolution‐transmission electron microscopy (HRTEM) clearly confirmed the tessellated hexagonal pore structure with a pore center‐to‐center distance of 2 nm. Powder X‐ray diffraction (PXRD) together with structural simulations revealed an AA stacking mode of the obtained layered structure. TP‐COF turned out to be an excellent semiconductor material with a LUMO energy of −3.23 eV and a band gap of 2.36 eV. Excitingly, this novel sp 2 ‐carbon conjugated TP‐COF exhibited unprecedented coenzyme regeneration efficiency and can significantly boost the coenzyme‐assisted synthesis of l ‐glutamate to a record‐breaking 97 % yield within 12 minutes.