Abstract The agricultural industry significantly relies on autonomous systems for detecting and analyzing rice diseases to minimize financial and resource losses, reduce yield reductions, improve processing efficiency, and ensure healthy crop production. Advances in deep learning have greatly enhanced disease diagnostic techniques in agriculture. Accurate identification of rice plant diseases is crucial to preventing the severe consequences these diseases can have on crop yield. Current methods often struggle with reliably diagnosing conditions and detecting issues in leaf images. Previously, leaf segmentation posed challenges, and while analyzing complex disease stages can be effective, it is computationally intensive. Therefore, segmentation methods need to be more accurate, cost-effective, and reliable. To address these challenges, we propose a hybrid bio-inspired algorithm, named the Hybrid WOA_APSO algorithm, which merges Adaptive Particle Swarm Optimization (APSO) with the Whale Optimization Algorithm (WOA). For disease classification in rice crops, we utilize a Convolutional Neural Network (CNN). Multiple experiments are conducted to evaluate the performance of the proposed model using benchmark datasets (Plantvillage), with a focus on feature extraction, segmentation, and preprocessing. Optimizing feature selection is a critical factor in enhancing the classification algorithm’s accuracy. We compare the accuracy, sensitivity, and specificity of our model against industry-standard techniques such as Support Vector Machine (SVM), Artificial Neural Network (ANN), and conventional CNN models. The experimental results indicate that the proposed hybrid approach achieves an impressive accuracy of 97.5% (Refer Table 8), which could inspire further research in this field.
Abstract Mapping a sentence into a two-dimensional (2D) representation can flatten nested semantic structures and build multi-granular span dependencies in named entity recognition. Existing approaches to recognizing named entities often classify each entity span independently, which ignores the spatial structures between neighboring spans. To address this issue, we propose a Span Spatial Attention Model (SSAM) that consists of a token encoder, a span generation module, and a 2D spatial attention network. The SSAM employs a two-channel span generation strategy to capture multi-granular features. Unlike traditional attention implemented on a sequential sentence representation, spatial attention is applied to a 2D sentence representation, enabling the model to learn the spatial structures of the sentence. This allows the SSAM to adaptively encode important features and suppress non-essential information in the 2D sentence representation. Experimental results on the GENIA, ACE2005, and ACE2004 datasets demonstrate that our proposed model achieves state-of-the-art performance, with F1-scores of 81.82%, 89.04%, and 89.24%, respectively. The code is available at https://github.com/Gzuwkj/SpatialAttentionForNer .
Abstract The pursuit of obtaining enzymes with high activity and stability remains a grail in enzyme evolution due to the stability-activity trade-off. Here, we develop an isothermal compressibility-assisted dynamic squeezing index perturbation engineering (iCASE) strategy to construct hierarchical modular networks for enzymes of varying complexity. Molecular mechanism analysis elucidates that the peak of adaptive evolution is reached through a structural response mechanism among variants. Furthermore, this dynamic response predictive model using structure-based supervised machine learning is established to predict enzyme function and fitness, demonstrating robust performance across different datasets and reliable prediction for epistasis. The universality of the iCASE strategy is validated by four sorts of enzymes with different structures and catalytic types. This machine learning-based iCASE strategy provides guidance for future research on the fitness evolution of enzymes.
Emmanuel Gottlob, Dan S. Borgnia, Robert-Jan Slager
et al.
The robustness of topological properties, such as quantized currents, generally depends on the existence of gaps surrounding the relevant energy levels or on symmetry-forbidden transitions. Here, we observe quantized currents that survive the addition of bounded local disorder beyond the closing of the relevant instantaneous energy gaps in a driven Aubry-André-Harper chain, a prototypical model of quasiperiodic systems. We explain the robustness using a local picture in configuration space based on Landau-Zener transitions, which rests on the Anderson localization of the eigenstates. Moreover, we propose a protocol, directly realizable in, for instance, cold atoms or photonic experiments, that leverages this stability to prepare topological many-body states with high Chern numbers and opens new experimental avenues for the study of both the integer and fractional quantum Hall effects.
Abstract We give a causal inference scheme using quantum observations alone for a case with both temporal and spatial correlations: a bipartite quantum system with measurements at two times. The protocol determines compatibility with five causal structures distinguished by the direction of causal influence and whether there are initial correlations. We derive and exploit a closed-form expression for the spacetime pseudo-density matrix (PDM) for many times and qubits. This PDM can be determined by light-touch coarse-grained measurements alone. We prove that if there is no signalling between two subsystems, the reduced state of the PDM cannot have negativity, regardless of initial spatial correlations. In addition, the protocol exploits the time asymmetry of the PDM to determine the temporal order. The protocol succeeds for a state with coherence undergoing a fully decohering channel. Thus coherence in the channel is not necessary for the quantum advantage of causal inference from observations alone.
Oishi Jyoti, Md. Samiul Habib, Nguyen Hoang Hai
et al.
We present a tunable broadband terahertz (THz) metamaterial absorber with a structurally simple, single-layered vanadium dioxide (VO2) elliptical ring resonator. This design achieves a wide, near-perfect absorption band (3.5–5 THz) without the need for complex multi-layer stacks or hybrid-patterned alternatives. Full-wave simulations demonstrate that VO2’s insulator-to-metal transition dynamically enhances absorption, while structural parameters—ring width, ellipticity ratio, and dielectric thickness—precisely control bandwidth and spectral response, as explained by impedance matching theory and electric field distributions. Furthermore, we explore the impact of varying the angle of incidence, highlighting the angular sensitivity of the structure. Beyond conventional parametric sweeps, we implement a targeted machine learning (ML) strategy for inverse design. Our models, trained on augmented data, show that Random Forest Regressor excels in predicting multiple geometric parameters simultaneously, while CatBoost is optimal for single-target prediction. The predicted geometric parameters are validated through simulation; this ML-guided approach, tailored to different design goals, combines physics-based modeling with data-driven optimization, offering a robust and efficient framework for designing next-generation broadband THz absorbers.
Solving long-horizon, temporally-extended tasks using Reinforcement Learning (RL) is challenging, compounded by the common practice of learning without prior knowledge (or tabula rasa learning). Humans can generate and execute plans with temporally-extended actions and quickly learn to perform new tasks because we almost never solve problems from scratch. We want autonomous agents to have this same ability. Recently, LLMs have been shown to encode a tremendous amount of knowledge about the world and to perform impressive in-context learning and reasoning. However, using LLMs to solve real world problems is hard because they are not grounded in the current task. In this paper we exploit the planning capabilities of LLMs while using RL to provide learning from the environment, resulting in a hierarchical agent that uses LLMs to solve long-horizon tasks. Instead of completely relying on LLMs, they guide a high-level policy, making learning significantly more sample efficient. This approach is evaluated in simulation environments such as MiniGrid, SkillHack, and Crafter, and on a real robot arm in block manipulation tasks. We show that agents trained using our approach outperform other baselines methods and, once trained, don't need access to LLMs during deployment.
Social network research analyzes the relationships between interactions, people, organizations, and entities. With the developing reputation of social media, community detection is drawing the attention of researchers. The purpose of community detection is to divide social networks into groups. These communities are made of entities that are very closely related. Communities are defined as groups of nodes or summits that have strong relationships among themselves rather than between themselves. The clustering of social networks is important for revealing the basic structures of social networks and discovering the hyperlink of systems on human beings and their interactions. Social networks can be represented by graphs where users are shown with the nodes of the graph and the relationships between the users are shown with the edges. Communities are detected through clustering algorithms. In this paper, we proposed a new clustering algorithm that takes into account the extent of relationships among people. Outcomes from particular data suggest that taking into account the profundity of people-to-people relationships increases the correctness of the aggregation methods.
XU Le, AN Hong, CHEN Junshi, ZHANG Pengfei, WU Zheng
The performance of unstructured grid computing on Sunway TaihuLight, a domestic heterogeneous many-core platform, is limited by sparse storage, discrete memory access, and data dependency.To relieve the sparse storage and discrete memory access problems, this paper proposes an N-order diagonal coloring algorithm, which effectively balances the computing between Management Processing Element (MPE) and Computing Processing Elements (CPEs) and convert global memory access to Local Device Memory (LDM) access using CPEs.To solve the computing competition caused by data dependence, this paper presents an adaptive and independent blocking method to avoid data conflicts in parallel computing.Furthermore, various optimizations are employed to overcome the performance bottlenecks:1.To leverage hardware resources, the authors use asynchronous parallelism between MPE and CPEs.2.To reduce synchronization costs, they avoid register communication, which increases the scalability of the next-generation Sunway platform.3.To hide the memory access latency, the authors overlap memory access with computing.The SpMV, Integration, and calcLudsFcc operations are generally used to verify the validity of the algorithm, and the results show that our algorithm achieves an average speedup of about 10 times and up to 24 times higher than that of the MPE implementation.Moreover, the N-order diagonal coloring algorithm has a 5.8 times higher speedup than that of the non-coloring blocking algorithm, which effectively improves data locality and computational parallelism.The algorithm also has good acceleration performance for dependent conflict operators, which verifies the effectiveness of adaptive and independent task partitioning methods.
In order to study the basic principles of vibration-excited liquid metal nucleation technology, a coupled model to connect the temperature field calculated by ANSYS Fluent and the dendritic growth simulated by cellular automaton (CA) algorithm was proposed. A two-dimensional CA model for dendrite growth controlled by solute diffusion and local curvature effects with random zigzag capture rule was developed. The proposed model was applied to simulate the temporal evolution of solidification microstructures under different degrees of surface undercooling and vibration frequency of the crystal nucleus generator conditions. The simulation results showed that the predicted columnar dendrites regions were more developed, the ratio of interior equiaxed dendrite reduced and the size of dendrites increased with the increase of the surface undercooling degrees on the crystal nucleus generator. It was caused by a large temperature gradient formed in the melt. The columnar-to-equiaxed transition (CET) was promoted, and the refined grains and homogenized microstructure were also achieved at the high vibration frequency of the crystal nucleus generator. The influences of the different process parameters on the temperature gradient and cooling rates in the mushy zone were investigated in detail. A lower cooling intensity and a uniform temperature gradient distribution could promote nucleation and refine grains. The present research has guiding significance for the process parameter selection in the actual experimental.
Schistosomiasis is a neglected tropical disease affecting communities surrounded by water bodies where fishing activities take place or people go to swim, wash and cultivate crops. It poses a great risk to the health and economic life of inhabitants of the area. This study was carried out to evaluate the impact of public health education and snail control measures on the incidence of schistosomiasis. A model was developed with attention given to the snail and human populations that are the hosts of the cercariae and miracidia respectively. The existence and stability of disease-free and endemic equilibrium states were established. The disease-free and endemic equilibrium states were shown to be locally asymptotically stable whenever the basic reproduction number was less than unity. Numerical simulations of the model were carried out to evaluate the impact of interventions (public health education and snail control measures) on schistosomiasis transmission. It was observed that the implementation of low coverage snail control with highly efficacious molluscicide and massive public health education will make the basic reproduction number smaller than unity, which implies the eradication of schistosomiasis in the population.