Rice, a staple food crop worldwide, is pivotal in agricultural productivity and public health. Automatic classification of typical rice pests diseases crucial for optimizing yield quality practical production. However, infrequent occurrences specific lead to uneven dataset samples similar early-stage symptoms, posing challenges effective identification methods. In this study, we employ four image enhancement techniques—flipping, modifying saturation, contrast, adding blur—to balance...

We investigate the channel assignment problem in a cluster-based multi-channel cognitive radio sensor network this paper. Due to inherent power and resource constraints of networks, energy efficiency is primary concern for design. An R-coefficient developed estimate predicted residual using information (current expected consumption) conditions (primary user behavior). examine three approaches: Random pairing, Greedy search Optimization-based assignment. The last two exploit obtain aware...

Large language models (LLMs) have shown remarkable capabilities in commonsense reasoning; however, some variations questions can trigger incorrect responses. Do these truly understand knowledge, or just memorize expression patterns? To investigate this question, we present the first extensive robustness evaluation of LLMs reasoning. We introduce HellaSwag-Pro, a large-scale bilingual benchmark consisting 11,200 cases, by designing and compiling seven types question variants. construct...

Modern General-Purpose computation on Graphics Processing Units (GPGPUs) explore parallelism in applications by building massively parallel architecture and apply multithreading technology to hide the instruction memory latencies. Such architectures become increasingly popular for using CUDA/OpenCL programming languages. In this paper, we investigate thread scheduling algorithms such highly-threaded GPGPUs. The traditional round-robin schemes are inefficient handling execution accesses with...

Abstract This paper aims to enhance the resilience of financial enterprises against environmental risks by leveraging data analysis tools. The approach involves designing risk assessment indicators and rating criteria. study utilizes a convolutional neural network model extended multi-scale feature fusion module analyze information in industry. proposed achieves impressive results with accuracy (Acc), precision ( P ), recall R F1 scores reaching 99.09, 96.31, 95.32, 95.64, respectively....

The primary user (PU)'s busy/idle behavior in a cognitive radio network is conventionally modeled using two-state Markov chain. Maximum likelihood (ML) estimation widely applied to estimate the state transition probabilities. This letter derives precise expression of probability mass function (PMF) for ML estimator, which has not been reported literature. By leveraging exact PMF expression, essential relation among number samples, probabilities, and accuracy revealed.

Perspective distortion often appears in the image documents which were token by digital camera. This phenomenon will lead to recognition errors or failures. Therefore, a new correction algorithm is proposed this paper for perspective images of Chinese documents. The makes use document image's horizontal characteristics text line and characterpsilas features vertical strokes, find information then rectify image. method does not require imagepsilas edge paragraphpsilas format. It has good...

Aiming at the low efficiency of path planning RRT-Connect algorithm and random sampling, an improved based on is proposed. The introduces a target bias strategy original algorithm, guides sampling point to expand in direction point, changes length step during tree expansion process, thereby increasing speed exploring space, finally adopts greedy Prune achieve smoothing. compared with RRT algorithms respectively, results show that proposed this paper significantly better than terms execution time.

Modeling and estimating the primary user (PU) behavior is critical to implement dynamic spectrum access in cognitive radio networks. In this paper, we investigate estimation accuracy of PU based on Markov model. Maximum Likelihood (ML) employed estimate transition probabilities model sample sequence idle/busy states. An approximate distribution ML estimator derived evaluate specified by confidence level. To meet requirement while reducing unnecessary sensing time, propose a learning...

Synthetic aperture radar (SAR) image segmentation is a challenging problem in recent years because of the speckle noise. An unsupervised SAR with superpixels by independent component analysis (ICA) proposed. ICA space proposed to represent images for feature extraction effectively. First, divided into small regions mean-shift algorithm and then those are merged region adjacent graph full-connected based on Mining Spanning Tree theory, which balances speed quality segmentation. Finally,...

The rapid advancement of Large Language Models (LLMs) in the realm mathematical reasoning necessitates comprehensive evaluations to gauge progress and inspire future directions. Existing assessments predominantly focus on problem-solving from examinee perspective, overlooking a dual perspective examiner regarding error identification correction. From we define four evaluation tasks for correction along with new dataset annotated types steps. We also design diverse prompts thoroughly evaluate...

With the continuous development of economy in China, manual production has long been unable to meet current social needs, and more companies have begun use robots instead operations carry out changes[1]. In manufacturing field, human-machine collaborative can effectively combine advantages industrial workers, reduce costs, improve product quality, make flexible individualized, develop direction high speed, precision efficiency. A key factor collaboration is safety security, namely operate...