Search

Abstract Lateral interaction in the biological brain is a key mechanism that underlies higher cognitive functions. Linear self‐organising map (SOM) introduces lateral interaction in a general form in which signals of any modality can be used. Some approaches directly incorporate SOM learning rules into neural networks, but incur complex operations and poor extendibility. The efficient way to implement lateral interaction in deep neural networks is not well established. The use of Laplacian Matrix‐based Smoothing (LS) regularisation is proposed for implementing lateral interaction in a concise form. The authors’ derivation and experiments show that lateral interaction implemented by SOM model is a special case of LS‐regulated k‐means, and they both show the topology‐preserving capability. The authors also verify that LS‐regularisation can be used in conjunction with the end‐to‐end training paradigm in deep auto‐encoders. Additionally, the benefits of LS‐regularisation in relaxing the requirement of parameter initialisation in various models and improving the classification performance of prototype classifiers are evaluated. Furthermore, the topologically ordered structure introduced by LS‐regularisation in feature extractor can improve the generalisation performance on classification tasks. Overall, LS‐regularisation is an effective and efficient way to implement lateral interaction and can be easily extended to different models.

Abstract Ligand-induced receptor dimerization or oligomerization is a widespread mechanism for ensuring communication specificity, safeguarding receptor activation, and facilitating amplification of signal transduction across the cellular membrane. However, cell-surface antigen-induced multimerization (dubbed AIM herein) has not yet been consciously leveraged in chimeric antigen receptor (CAR) engineering for enriching T cell-based therapies. We co-developed ciltacabtagene autoleucel (cilta-cel), whose CAR incorporates two B-cell maturation antigen (BCMA)-targeted nanobodies in tandem, for treating multiple myeloma. Here we elucidated a structural and functional model in which BCMA-induced cilta-cel CAR multimerization amplifies myeloma-targeted T cell-mediated cytotoxicity. Crystallographic analysis of BCMA–nanobody complexes revealed atomic details of antigen–antibody hetero-multimerization whilst analytical ultracentrifugation and small-angle X-ray scattering characterized interdependent BCMA apposition and CAR juxtaposition in solution. BCMA-induced nanobody CAR multimerization enhanced cytotoxicity, alongside elevated immune synapse formation and cytotoxicity-mediating cytokine release, towards myeloma-derived cells. Our results provide a framework for contemplating the AIM approach in designing next-generation CARs.

Average-rendezvous problem with connectivity preservation of constant reference signals is investigated for first-order multiagent systems. According to a constructive potential function, a proportional–integral average-rendezvous algorithm is proposed to make agents converge to the average value of constant reference signals as well as the network connectivity is preserved all the time. Sufficient convergence conditions are obtained for the algorithm by designing the proper Lyapunov functions. Numerical examples illustrate the correctness of theoretical results.

Conductivity of molten alkali metal chlorides is the foundation of molten-salt electrolysis process. In the study, equilibrium molecular dynamics was performed to investigate the ionic conductivities of the molten alkali metal chloride binary systems. The calculated values were in good agreement with the experimental results. Simple databases for conductivity have been built based on the verified algorithm. Mathematic expressions of conductivity vs. temperature & composition have also been obtained according to calculated data. Ionic conductivities of these melts exhibited positive dependences on temperature. Adding a smaller alkali metal cation into molten alkali chlorides could improve the conductive performance. With the gradual addition of LiCl, ionic conductivities of molten LiCl-NaCl, LiCl-KCl, LiCl-RbCl, and LiCl-CsCl all increased monotonously. The conductivities of molten NaCl-KCl, NaCl-RbCl, and NaCl-CsCl also increased with the continuous addition of NaCl. The improvement effects of LiCl on ionic conductivities of molten KCl, RbCl, and CsCl are more significant than those of NaCl.

ABSTRACT: In the greater inland Jianghan Basin of South China, three salt depressions are lacking accurate geological times, of which Jiangling Depression is the largest. Evaporites are important records of paleoclimate, however, the geological ages of evaporates are very difficult to be determined because often evaporates have scare macrofossils and microfossils. Nonmarine Cretaceous to Tertiary halite deposits interbedded with mudstones are widely distributed in China. Paleocene-Eocene Thermal Maximum had very high temperatures and attracted strong interests of geologists because these times can be compared with future climate change because of global warming. However, previous studies focused on marine sediments found that during the Paleocene-Early Eocene, massive evaporate deposits formed in Jiangling depression of the Jianghan Basin. In this paper, the authors show that the Shashi Formation halite deposits formed in the Paleocene according to palynology. Most of these palynology fossils are arid types, so the massive evaporites in the Jiangling depression could be closely related to the hot Paleocene climate. High temperatures during the Paleocene contributed to the formation of the massive evaporates in the Jiangling Depression, until sylvite was the result.

Colorectal cancer (CRC) is the third most common cancer worldwide, and liver metastasis presents a major cause of CRC-associated death. Extensive genomic analysis has provided valuable insight into the pathogenesis and progression of CRC; however, a comprehensive proteogenomic characterization of CRC liver metastasis (CLM) has yet to be reported. Here, we analyzed the proteomes of 44 paired normal colorectal tissues and CRC tissues with or without liver metastasis, as well as analyzed genomics of CRC characterized previously by The Cancer Genome Atlas (TCGA) to conduct integrated proteogenomic analyses. We identified a total of 2,170 significantly deregulated proteins associated with CLM, 14.88% of which were involved in metabolic pathways. The mutated peptide number was found to have potential prognosis value, and somatic variants revealed two metabolism-related genes UQCR5 and FDFT1 that frequently mutated only in the liver metastatic cohort and displayed dysregulated protein abundance with biological function and clinical significance in CLM. Proteogenomic characterization and integrative and comparative genomic analysis provides functional context and prognostic value to annotate genomic abnormalities and affords a new paradigm for understanding human colon and rectal cancer liver metastasis. Keywords: CRC, liver metastasis, proteomics, genomics, integrated analysis

There are many decisions which are usually made heuristically both in single object tracking (SOT) and multiple object tracking (MOT). Existing methods focus on tackling decision-making problems on special tasks in tracking without a unified framework. In this paper, we propose a decision controller (DC) which is generally applicable to both SOT and MOT tasks. The controller learns an optimal decision-making policy with a deep reinforcement learning algorithm that maximizes long term tracking performance without supervision. To prove the generalization ability of DC, we apply it to the challenging ensemble problem in SOT and tracker-detector switching problem in MOT. In the tracker ensemble experiment, our ensemble-based tracker can achieve leading performance in VOT2016 challenge and the light version can also get a state-of-the-art result at 50 FPS. In the MOT experiment, we utilize the tracker-detector switching controller to enable real-time online tracking with competitive performance and 10× speed up.

Based on an electrochemical multiphysical simulation, a method for analysing electrolysis efficiency has been presented that considers the energy consumption required to produce a single kilogram of lithium and for the production of lithium, rather than the voltage in various parts. By adopting them as the criteria for analysing electrolysis efficiency in the lithium cell, several structural parameters have been optimized, such as the anode radius and anode–cathode distance. These parameters strongly affect the cell voltage and the velocity field distribution, which has a significant impact on the concentration distribution. By integrating the concentration distribution, the lithium production and energy consumption per kilogram, lithium is computed. By appointing the minimum of the chlorine and lithium concentration as the secondary reaction intensity, it is clear where the secondary reaction intensity is strong in the cell. The structure of a lithium electrolysis cell has been optimized by applying an orthogonal design approach, with the energy consumption notably decreasing from 35.0 to 28.3 kWh (kg Li)−1 and the lithium production successfully increasing by 0.17 mol.

ABSTRACT: With the technological development of exploitation and separation, the primary sources of lithium have gradually changed from ore to brine, which has become the main raw material, accounting for more than 80% of the total production. Resources of lithium-bearing brine are abundant in China. This paper has summarized the spatial and temporal distribution, characteristics, and formation mechanism of the lithium-rich brine in China, aiming to provide a comprehensive set of guidelines for future lithium exploitation from brines. Lithium-rich brines usually exist in modern saline lakes and deep underground sedimentary rocks as subsurface brines. The metallogenic epoch of China’s lithium-rich brine spans from the Triassic to the Quaternary, and these brines exhibit obvious regional distribution characteristics. Modern lithium-rich saline lakes are predominately located in the Qinghai-Tibet Plateau. In comparison, the subsurface lithium-rich brines are mainly distributed in the sedimentary basins of Sichuan, Hubei, Jiangxi provinces and so on in south Block of China, and some are in the western part of the Qaidam Basin in Qinghai province in northwestern China. Lithium-rich saline lakes are belonging to chloride-enriched, sulfate-enriched, and carbonate-enriched, while the deep lithium-rich brines are mainly chloride-enriched in classification. On the whole, the value of Mg/Li in deep brine is generally lower than that of brine in saline lakes. The genesis of lithium-rich brines in China is not uniform, generally there are two processes, which are respectively suitable for salt lakes and deep brine.

Event-triggered consensus problem is studied for a class of heterogeneous first-order multi-agent system, which contains multiple single integrators with different nominal velocities, and an adaptive consensus algorithm in the leader-following structure is adopted. Under the event-triggered mechanism, the agents' inputs are only updated at discrete times determined by a centralized event-triggered function. By constructing proper Lyapunov function, sufficient consensus conditions are gained for the asymptotic consensus convergence. Furthermore, the consensus problem subject to identical input delay is considered, and consensus conditions in the form of linear matrix inequalities are obtained. Numerical simulations show the effectiveness of the event-triggered consensus control strategy.

To effectively mine the contents embedded in web images, it is useful to classify the images into different types so that they can be fed to different procedures for detailed analysis. The authors herein propose a hierarchical algorithm for efficiently classifying web images into four classes. Their algorithm consists of two stages: the first stage extracts global features reflecting the distributions of color, edge and gradient, and uses a support vector machine (SVM) classifier for preliminary classification. Images assigned low confidence by the first stage classifier are processed by the second stage, which further extracts local texture features represented in the bag-of-words framework and uses another SVM classifier for final classification. In addition, they design two fusion strategies to train the second-stage classifier and generate the final prediction depending on the usage of local features in the second stage. To validate the effectiveness of proposed method, they built a database containing more than 55,000 images from various sources. On their test image set, they obtained an overall classification accuracy of 98.4% and the processing speed is over 27 fps on an Intel(R) Xeon(R) central processing unit (2.90 GHz).

This paper investigates the distributed filtering for discrete-time-invariant systems in sensor networks where each sensor’s measuring system may not be observable, and each sensor can just obtain partial system parameters with unknown coefficients which are modeled by Gaussian white noises. A fully distributed robust Kalman filtering algorithm consisting of two parts is proposed. One is a consensus Kalman filter to estimate the system parameters. It is proved that the mean square estimation errors for the system parameters converge to zero if and only if, for any one system parameter, its accessible node subset is globally reachable. The other is a consensus robust Kalman filter to estimate the system state based on the system matrix estimations and covariances. It is proved that the mean square estimation error of each sensor is upper-bounded by the trace of its covariance. An explicit sufficient stability condition of the algorithm is further provided. A numerical simulation is given to illustrate the results.

This paper investigates the formation-containment control of second-order multiagent systems with intermittent communication. Distributed coordination control algorithms are proposed under aperiodic intermittent communication, where each agent only communicates with its neighboring agents on some disconnected time intervals. By means of constructing Lyapunov functions, sufficient convergence conditions are obtained for the leaders reaching a prescribed formation asymptotically and the followers converging into the convex hull formed by leaders asymptotically, respectively. Besides, sufficient convergence conditions are also provided for second-order multiagent systems converging to the desired formation-containment under time-varying communication delay and intermittent communication. Finally, the validity of theoretical results is illustrated by numerical simulations.

Gas-evolving vertical electrode system is a typical electrochemical industrial reactor. Gas bubbles are released from the surfaces of the anode and affect the electrolyte flow pattern and even the cell performance. In the current work, the hydrodynamics induced by the air bubbles in a cold model was experimentally and numerically investigated. Particle image velocimetry and volumetric three-component velocimetry techniques were applied to experimentally visualize the hydrodynamics characteristics and flow fields in a two-dimensional (2D) plane and a three-dimensional (3D) space, respectively. Measurements were performed at different gas rates. Furthermore, the corresponding mathematical model was developed under identical conditions for the qualitative and quantitative analyses. The experimental measurements were compared with the numerical results based on the mathematical model. The study of the time-averaged flow field, three velocity components, instantaneous velocity and turbulent intensity indicate that the numerical model qualitatively reproduces liquid motion. The 3D model predictions capture the flow behaviour more accurately than the 2D model in this study.

In this paper, the leader-following consensus algorithm, which is accompanied with compensations related to neighboring agents’ delayed states, is constructed for second-order multi-agent systems with communication delay. Using frequency-domain analysis, delay-independent and delay-dependent consensus conditions are obtained for second-order agents respectively to converge to the dynamical leader’s states asymptotically. Simulation illustrates the correctness of the results.

This paper deals with the consensus tracking problem of heterogeneous linear multiagent systems under the repeatable operation environment, and adopts a proportional differential (PD)-type iterative learning control (ILC) algorithm based on the fractional-power tracking error. According to graph theory and operator theory, convergence condition is obtained for the systems under the interconnection topology that contains a spanning tree rooted at the reference trajectory named as the leader. Our algorithm based on fractional-power tracking error achieves a faster convergence rate than the usual PD-type ILC algorithm based on the integer-order tracking error. Simulation examples illustrate the correctness of our proposed algorithm.

The evaluation of the trench stability under the action of ocean waves is an important issue in the construction of an immersed tunnel. In this study, a two-dimensional coupling model of a wave-seabed-immersed tunnel is proposed for the dynamic responses of a trench under wave action in the immersing process of tunnel elements. The porous seabed is characterized by Biot consolidation equations. The k − ε model and RANS equation are adopted to achieve the flow field simulation, and the level set method (LSM) is used to capture the free surface between the water and air. The proposed numerical model is verified using the experimental data and analytical results. Then, the transient liquefaction and shear failure in the vicinity of the trench are discussed at two different conditions, namely, after the foundation groove is excavated and after the tunnel element is placed. The pore pressure amplitude on the weather side slope is demonstrated to be significantly smaller than that on the lee side slope. Also, the distribution of the surrounding flow field and pressure field change dramatically after the tunnel element is settled, leading to the significant changes of seabed stability.

General asynchronously coupled consensus algorithm associated with adjacent compensations, is proposed to solve the dynamical consensus problem of second-order multiagent network with communication delay under leader-following coordination control framework. Based on frequency-domain analysis, firstly, delay-independent consensus convergence is proved for the second-order multiagent systems with a spanning tree topology that has the leader root and then delay-dependent consensus condition is obtained for the multiagent systems with communication delay under a general leader-following interconnection topology. Simulation illustrates the correctness of the results.