Due to the complexity of the underwater environment, underwater acoustic target recognition is more challenging than ordinary target recognition, and has become a hot topic in the field of underwater acoustics researc...
详细信息
Deep learning has gained tremendous success in various fields while training deep neural networks(DNNs) is very compute-intensive, which results in numerous deep learning frameworks that aim to offer better usability ...
详细信息
Deep learning has gained tremendous success in various fields while training deep neural networks(DNNs) is very compute-intensive, which results in numerous deep learning frameworks that aim to offer better usability and higher performance to deep learning practitioners. Tensor Flow and Py Torch are the two most popular frameworks. Tensor Flow is more promising within the industry context, while Py Torch is more appealing in academia. However, these two frameworks differ much owing to the opposite design philosophy:static vs dynamic computation graph. Tensor Flow is regarded as being more performance-friendly as it has more opportunities to perform optimizations with the full view of the computation graph. However, there are also claims that Py Torch is faster than Tensor Flow sometimes, which confuses the end-users on the choice between them. In this paper, we carry out the analytical and experimental analysis to unravel the mystery of comparison in training speed on single-GPU between Tensor Flow and Py Torch. To ensure that our investigation is as comprehensive as possible, we carefully select seven popular neural networks, which cover computer vision, speech recognition, and natural language processing(NLP). The contributions of this work are two-fold. First, we conduct the detailed benchmarking experiments on Tensor Flow and Py Torch and analyze the reasons for their performance difference. This work provides the guidance for the end-users to choose between these two frameworks. Second, we identify some key factors that affect the performance,which can direct the end-users to write their models more efficiently.
Existing long-tail classification (LT) methods ignore attribute class imbalance and focus only on addressing class imbalance where the head class has more samples than the tail class. In fact, even if the classes are ...
详细信息
3D human pose and body shape estimation is a hot research topic in computer vision, and there are many optimization-based methods that rely only on the input parameters to obtain the optimal solutions for pose and bod...
详细信息
Over the past few years,deep reinforcement learning(RL)has made remarkable progress in a range of applications,including Go games,vision-based control,and generative dialogue *** error-and-trial mechanisms,deep RL ena...
详细信息
Over the past few years,deep reinforcement learning(RL)has made remarkable progress in a range of applications,including Go games,vision-based control,and generative dialogue *** error-and-trial mechanisms,deep RL enables data-driven optimization and sequential decision-making in uncertain *** to traditional programming or heuristic optimization methods,deep RL can elegantly balance exploration and exploitation and handle environmental *** a result,this learning paradigm has attracted increasing attention from both academia and industry and is paving a new path for largescale complex decision-making applications.
In this research,activated carbon from mangosteen peel has been synthesized using sulfuric acid as an *** adsorption performance of the activated carbon was optimized using malachite green dye as ***-chite green dye w...
详细信息
In this research,activated carbon from mangosteen peel has been synthesized using sulfuric acid as an *** adsorption performance of the activated carbon was optimized using malachite green dye as ***-chite green dye waste is a toxic and non-biodegradable material that damages the *** of adsorption processes was carried out using Response Surface Methodology(RSM)with a Box-Behnken Design(BBD).The synthesized activated carbon was characterized using FTIR and SEM *** FTIR spectra confirmed the presence of a sulfonate group(-SO_(3)H)in the activated carbon,indicating that the activation pro-cess using sulfuric acid was *** characterization shows that activated carbon has porous *** was carried out for three adsorption parameters,namely contact time(20,60,and 120 min),adsor-bent mass(0.005,0.025,and 0.05 g),and initial concentration of malachite green solution(5,50,and 100 mg·L^(-1)).The concentration of the malachite green solution was determined using a UV-Vis spectrophotometer at the max-imum wavelength of malachite green,618 *** optimum of malachite green adsorption using mangosteen peel activated carbon was obtained at a contact time of 80 min,an adsorbent mass of 0.032 g,and malachite green initial concentration of 25 mg·L^(-1),with a maximum removal percentage and maximum adsorption capacity of 93.66%and 19.345 mg·g^(-1),respectively.
Rough set theory places great importance on approximation accuracy,which is used to gauge how well a rough set model describes a target ***,traditional approximation accuracy has limitations since it varies with chang...
详细信息
Rough set theory places great importance on approximation accuracy,which is used to gauge how well a rough set model describes a target ***,traditional approximation accuracy has limitations since it varies with changes in the target concept and cannot evaluate the overall descriptive ability of a rough set *** overcome this,two types of average approximation accuracy that objectively assess a rough set model’s ability to approximate all information granules is *** first is the relative average approximation accuracy,which is based on all sets in the universe and has several basic *** second is the absolute average approximation accuracy,which is based on undefinable sets and has yielded significant *** also explore the relationship between these two types of average approximation ***,the average approximation accuracy has practical applications in addressing missing attribute values in incomplete information tables.
Group testing involves discovering a small subset of distinguished subjects from a large population while efficiently reducing the total number of *** has been widely used for industrial testing,information technology...
详细信息
Group testing involves discovering a small subset of distinguished subjects from a large population while efficiently reducing the total number of *** has been widely used for industrial testing,information technology,and biology,especially epidemic ***,in reality,are noisy for the presence of false *** tests are accurate but time-consuming,while others are cheaper but less *** which test to use is constrained by various considerations,such as availability,cost,accuracy,and *** this paper,we propose flexible,efficient,and accurate tests(FEATs).FEATs are based on group testing with simple but careful designs by incorporating ideas such as close contact cliques and repeated *** could dramatically improve the efficiency or accuracy of existing *** example,for accurate but slow tests,the FEAT can improve efficiency multiple times without compromising *** the other hand,for fast but inaccurate tests,the FEAT can sharply reduce the false-negative rate(FNR)and significantly increase *** justifications are *** point out some scenarios where the FEAT can be effectively employed.
An improved algorithm is proposed for the omission and re-detection problems in the point cloud object detection method CenterPoint. The algorithm firstly adds Focal sparse convolution module to the feature extraction...
详细信息
Generating molecules with desired properties is an important task in chemistry and *** efficient method may have a positive impact on finding drugs to treat diseases like *** mining and artificial intelligence may be ...
详细信息
Generating molecules with desired properties is an important task in chemistry and *** efficient method may have a positive impact on finding drugs to treat diseases like *** mining and artificial intelligence may be good ways to find an efficient ***,both the generative models based on deep learning and the work based on genetic algorithms have made some progress in generating molecules and optimizing the molecule’s ***,existing methods have defects in the experimental evaluation *** methods also need to be improved in efficiency and *** solve these problems,we propose a method named the Chemical Genetic Algorithm for Large Molecular Space(CALM).Specifically,CALM employs a scalable and efficient molecular representation called molecular *** we design corresponding crossover,mutation,and mask operators inspired by domain knowledge and previous *** apply our genetic algorithm to several tasks related to molecular property optimization and constraint molecular *** results of these tasks show that our approach outperforms the other state-of-the-art deep learning and genetic algorithm methods,where the z tests performed on the results of several experiments show that our method is more than 99%likely to be *** the same time,based on the experimental results,we point out the defects in the experimental evaluation standard which affects the fair evaluation of all previous *** these defects helps to objectively evaluate the performance of all work.
暂无评论