检索结果-内蒙古大学图书馆

A comparison of statistical learning of naturalistic textures between DCNNs and the human visual hierarchy

science China(Technological sciences) 2024年第8期67卷 2310-2318页

作者： LU XinCheng YUAN ZiQi ZHANG YiChi AI HaiLin CHENG SiYuan GE YiRan FANG Fang CHEN NiHong Department of Psychological and Cognitive Sciences Tsinghua UniversityBeijing 100084China IDG/McGovern Institute for Brain Research Tsinghua UniversityBeijing 100084China Department of Computer Science and Technology Tsinghua UniversityBeijing 100084China School of Psychological and Cognitive Sciences and Beijing Key Laboratory of Behavior and Mental Health Peking UniversityBeijing 100871China Key Laboratory of Machine Perception(Ministry of Education) Peking UniversityBeijing 100871China Peking-Tsinghua Center for Life Sciences Peking UniversityBeijing 100871China IDG/McGovern Institute for Brain Research Peking UniversityBeijing 100871China State Key Laboratory of Brain and Cognitive Science Institute of PsychologyChinese Academy of SciencesBeijing 100101China

The visual system continuously adapts to the statistical properties of the environment. Existing evidence shows a close resemblance between deep convolutional neural networks(CNNs) and primate visual stream in neural selectivity to naturalistic textures above the primary visual processing stage. This study delves into the mechanisms of perceptual learning in CNNs,focusing on how they assimilate the high-order statistics of natural textures. Our results show that a CNN model achieves a similar performance improvement as humans, as manifested in the learning pattern across different types of high-order image statistics. While L2 was the first stage exhibiting texture selectivity, we found that stages beyond L2 were critically involved in learning. The significant contribution of L4 to learning was manifested both in the modulations of texture-selective responses and in the consequences of training with frozen connection weights. Our findings highlight learning-dependent plasticity in the mid-to-high-level areas of the visual hierarchy. This research introduces an AI-inspired approach for studying learning-induced cortical plasticity, utilizing DCNNs as an experimental framework to formulate testable predictions for empirical brain studies.

关键词： CNN perceptual learning naturalistic texture psychophysics

来源：评论

学校读者我要写书评

暂无评论

Literature Review: Analysis of Essential Metadata Attributes of Digital Goods for Transaction in The Virtual World 9

Literature Review: Analysis of Essential Metadata Attributes...

引用

9th International Conference on Information Management and Technology, ICIMTech 2024

作者： Martha, Gede Indra Raditya Prabowo, Harjanto Gaol, Ford Lumban Wiputra, Richard BINUS Graduate Program - Doctor of Computer Science Bina Nusantara University Computer Science Department Jakarta11480 Indonesia BINUS Business School Undergraduate Program Bina Nusantara University Management Department Jakarta11480 Indonesia School of Information Systems Bina Nusantara University Information Systems Department Jakarta11480 Indonesia

ISBN: (纸本)9798350390025

In online-based virtual worlds such as Metaverse, Online Games, and other online digital spaces, the virtual/digital goods (digital items / digital assets) are fundamental things that must be available to be able to do business & create an economy in the virtual world through monetization & transactions from these digital goods. This research focuses on summarizing and exploring the attributes used in digital goods in the virtual world by extracting attributes in digital goods from other related studies using the Systematic Literature Review methodology & PRISMA Framework comprehensively so that it can be a reference for developers in designing the properties and functions of digital goods in the virtual world that are being developed, which in this case will be in the form of an NFT in the future. Based on research that has been carried out, 3 essential metadata attributes have the most influence on transactions, number 1 Visual Design / Aesthetic (Representative), number 2 Effect (Utilities), and number 3 Statistics / Performance (Utilities) which must be considered in developing/making digital goods in the form of NFTs in the virtual world so that the value of digital goods can attract users to buy or transact them in the virtual world. © 2024 IEEE.

关键词： Virtual environments

来源：评论

学校读者我要写书评

暂无评论

DOMAIN RANDOMIZATION VIA ENTROPY MAXIMIZATION 12

DOMAIN RANDOMIZATION VIA ENTROPY MAXIMIZATION

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Tiboni, Gabriele Klink, Pascal Peters, Jan Tommasi, Tatiana D'Eramo, Carlo Chalvatzaki, Georgia Department of Control and Computer Engineering Politecnico di Torino Italy Department of Computer Science Technische Universität Darmstadt Germany Center for Artificial Intelligence and Data Science University of Würzburg Germany Darmstadt Germany Centre for Cognitive Science TU Darmstadt Germany Germany Center for Mind Brain and Behavior Uni. Marburg and JLU Giessen Germany

Varying dynamics parameters in simulation is a popular Domain Randomization (DR) approach for overcoming the reality gap in Reinforcement Learning (RL). Nevertheless, DR heavily hinges on the choice of the sampling distribution of the dynamics parameters, since high variability is crucial to regularize the agent's behavior but notoriously leads to overly conservative policies when randomizing excessively. In this paper, we propose a novel approach to address sim-to-real transfer, which automatically shapes dynamics distributions during training in simulation without requiring real-world data. We introduce DOmain RAndomization via Entropy MaximizatiON (DORAEMON), a constrained optimization problem that directly maximizes the entropy of the training distribution while retaining generalization capabilities. In achieving this, DORAEMON gradually increases the diversity of sampled dynamics parameters as long as the probability of success of the current policy is sufficiently high. We empirically validate the consistent benefits of DORAEMON in obtaining highly adaptive and generalizable policies, i.e. solving the task at hand across the widest range of dynamics parameters, as opposed to representative baselines from the DR literature. Notably, we also demonstrate the Sim2Real applicability of DORAEMON through its successful zero-shot transfer in a robotic manipulation setup under unknown real-world parameters. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Massively Scaling Explicit Policy-conditioned Value Functions

arXiv

引用

arXiv 2025年

作者： Bohlinger, Nico Peters, Jan Department of Computer Science Technical University of Darmstadt Germany Hessian.AI Centre for Cognitive Science Department of Computer Science Technical University of Darmstadt Germany

We introduce a scaling strategy for Explicit Policy-Conditioned Value Functions (EPVFs) that significantly improves performance on challenging continuous-control tasks. EPVFs learn a value function V (θ) that is explicitly conditioned on the policy parameters, enabling direct gradient-based updates to the parameters of any policy. However, EPVFs at scale struggle with unrestricted parameter growth and efficient exploration in the policy parameter space. To address these issues, we utilize massive parallelization with GPU-based simulators, big batch sizes, weight clipping and scaled peturbations. Our results show that EPVFs can be scaled to solve complex tasks, such as a custom Ant environment, and can compete with state-of-the-art Deep Reinforcement Learning (DRL) baselines like Proximal Policy Optimization (PPO) and Soft Actor-Critic (SAC). We further explore action-based policy parameter representations from previous work and specialized neural network architectures to efficiently handle weight-space features, which have not been used in the context of DRL before. Copyright © 2025, The Authors. All rights reserved.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Fuzzy Decision Support Model on Virtual Plant Computational Model of Water Spinach (Ipomoea aquatica. Forssk) in Farming Business 6

Fuzzy Decision Support Model on Virtual Plant Computational ...

引用

6th International Conference on Information Technology, Information Systems and Electrical Engineering, ICITISEE 2022

作者： Utama, Ditdit Nugeraha Agustia Rachmawati, Ika Diah Purwantono, Heraldo Yusron Bina Nusantara University Binus Graduate Program - Master of Computer Science Computer Science Department Jakarta11480 Indonesia Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia

ISBN: (纸本)9798350399615

Using soil as a planting medium (conventional system) raises several problems, such as the need for large agricultural land, but the available land is limited. This problem is triggered by an increase in demand for nutrients needed by the body to support daily activities, one of which can be met by consuming green vegetables, such as Water Spinach. To solve this problem, a hydroponic wick system or a wick system can use that is simple and easy to care for. Water Spinach (Ipomoea aquatica. Forssk) vegetables are easy to grow and care for. Nutrients contained in 100 grams of Water Spinach, energy: 29 kcal, protein: 3 grams, Fat: 0.3 grams, carbohydrates: 5.4 grams, calcium: 73 mg, phosphorus: 50 mg, iron: 3 mg, vitamin A: 6300IU,vitamin B1: 0.07 mg, vitamin C: 32 mg. In addition, a combination of disciplines such as mathematics, computer science, plant science, biology, and statistics can be used. This combination of knowledge is known as Plant Computational Modeling (PCM), which uses the Growth Grammar Interactive Modeling Platform (GroIMP) with the Functional Structural Plant Modeling (FSPM) method. The virtual plant model was successfully built, the results were divided into two parts, such as single plant and multi-plant. Plant computational modeling and DSM can help deal with agricultural problems. By using a plant dataset, the GroIMP-FSPM platform, and fuzzy logic, the model could help researchers in making a better decision based on BEP evaluation. The minimum numbers of plants cultivated were 650 plants in one harvest period in order to gain a desirable profit. We recommend to cultivate even more plants in one harvest period to gain even higher profits. © 2022 IEEE.

关键词： Decision Support Model Functional Structural Plant Modelling Fuzzy Logic Growth-Grammar Interactive Modelling Platform Ipomoea aquatica Forssk Plant Computational Modelling

来源：评论

学校读者我要写书评

暂无评论

Mapping the Scientific Landscape of Integrated Scheduling Towards Industry 5.0 with A Bibliometric Analysis 6

Mapping the Scientific Landscape of Integrated Scheduling To...

引用

6th International Conference on Industry 4.0 and Smart Manufacturing, ISM 2024

作者： Putri Ardyanti, Anak Agung Ayu Xie, Zhiqiang Karnyoto, Andrea Stevens Pardamean, Bens Estiyanti, Ni Made School of Computer Science and Technology Harbin University of Science and Technology Harbin150000 China Faculty of Information Technology and Design Primakara University Bali 80226 Indonesia Bioinformatics and Data Science Research Center Bina Nusantara University Jakarta11480 Indonesia Computer Science Department BINUS Graduate Program Master of Computer Science Bina Nusantara University Jakarta 11480 Indonesia School of Management and Business University of Debrecen Debrecen Hungary

The industry is rapidly transitioning from the 4.0 era to the 5.0 era, prompting renewed interest among scholars in scheduling problems. They allow operations to process and assemble various components simultaneously. This bibliometric analysis provides a comprehensive understanding of the diverse research perspectives on the potential challenges of integrated scheduling within the framework. We identified 357 Scopus articles and analysed co-occurrence keyword (CNK) by VOSViewer tools and Global Citation Score (GCS) by Scopus tools. The research data obtained from 1993 to 2023 mainly in the subject areas of Engineering, computer science, Maths, and others. CNK identified six cluster keywords, and nine most cited articles based on normalised GCS. These results contribute to our understanding of the development of scheduling algorithms, shifting from an emphasis on manufacturing transfers to contemporary demand-centric research directions. Furthermore, this research provides a valuable perspective for policymakers, plant personnel, and manufacturing managers to make informed decisions. © 2025 The Authors. Published by Elsevier B.V.

关键词： Job shop scheduling

来源：评论

学校读者我要写书评

暂无评论

The EGY-DRiVeS Dataset: Illuminating the Dynamics of Egyptian Road Traffic Behaviors

The EGY-DRiVeS Dataset: Illuminating the Dynamics of Egyptia...

引用

2024 International Conference on computer and Applications, ICCA 2024

作者： Saadawy, Zeyad H. Elfar, Mohamed A. Aly, Menna M. Adel, Momen Bahy, Mohamed A. Elias, Catherine M. C-DRiVeS Lab: Cognitive Driving Research in Vehicular Systems Cairo Egypt German University in Cairo Faculty of Media Engineering and Technology Computer Science and Engineering Department Egypt

ISBN: (纸本)9798350367560

This paper presents a comprehensive dataset of Egyptian roads integrated with chaotic scenarios captured by EGY-DRiVeS' lab golf car to aid in developing the autonomous driving experience. Using Velodyne 3D laser scanners, C922 Pro HD webcams, and iPhone 11 Pro cameras, a cumulative recording time of 3 hours and 48 minutes was documented. This included 1 hour and 15 minutes of LiDAR recordings, 1 hour and 45 minutes of camera recordings, and 48 minutes and 13 seconds of synced recordings. The dataset encompasses diverse scenarios and objects, ranging from urban environments with dynamic and static objects to rural highways. In-campus roads and intersections are also included, providing real-world scenarios such as vehicular and pedestrian behavior within campus premises. Each part of the dataset is annotated with 2D bounding boxes offering detailed information about vehicles, pedestrians and road objects. The provided data are calibrated, synchronized, and timestamped. Alongside the dataset, data format specifications, annotated data samples, and all supplementary recordings are provided upon request. © 2024 IEEE.

关键词： Network security

来源：评论

学校读者我要写书评

暂无评论

AlexNet Architecture Based Convolution Neural Network for Realtime Audio to Text Translator of Bisindo Hand Sign

AlexNet Architecture Based Convolution Neural Network for Re...

引用

2023 International Seminar on Application for Technology of Information and Communication, iSemantic 2023

作者： Sujatmiko, Dhiki Sari, Christy Atika Rachmawanto, Eko Hari Krismawan, Andi Danang Altamer, Bilal R. Alkhafaji, Mohamed Ayad University of Dian Nuswantoro Study Program in Informatics Engineering Semarang Indonesia University of Dian Nuswantoro Study Program in Animation Semarang Indonesia Mosul University Presidency University of Mosul Department of Computer Science Mosul Iraq National University of Science and Technology Department of Computer Science DhiQar Iraq

ISBN: (纸本)9798350339215

Deafness is a condition that results in the loss of hearing function, hindering the reception of information such as oral communication that relies on auditory senses. Consequently, individuals with hearing impairment experience communication barriers and may have limited or no ability to respond. One solution is the use of sign language. In Indonesia, there are two known sign languages: Sibi and Bisindo. Both serve the same function but differ in their style of movement and expression. Bisindo is considered more flexible as it conveys meaning based on the Indonesian language. However, the universal understanding of this language solution is still limited among many people. Therefore, a program is needed to facilitate translation between deaf individuals who use sign language and their counterparts who do not communicate through sign language. CNN (Convolutional Neural Network) is a deep learning algorithm used for training visual input data recognition by computer systems. There are various CNN-based architectures, and one of them is AlexNet. Based on the author's testing, the AlexNet architecture proves to be suitable for real-time sign language translation. The evaluation of the system involved 7,800 datasets and 520 testing instances, with an average accuracy of 468 correct translations. When averaged, the system achieved a 90% accuracy rate, representing a 100% increase in accuracy compared to previous approaches. © 2023 IEEE.

关键词： Audition

来源：评论

学校读者我要写书评

暂无评论

Visual Grounding Helps Learn Word Meanings in Low-Data Regimes

Visual Grounding Helps Learn Word Meanings in Low-Data Regim...

引用

2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2024

作者： Zhuang, Chengxu Fedorenko, Evelina Andreas, Jacob McGovern Institute for Brain Research MIT United States Department of Brain and Cognitive Sciences MIT United States The Program in Speech and Hearing Bioscience and Technology Harvard University United States Computer Science and Artificial Intelligence Laboratory MIT United States

ISBN: (纸本)9798891761148

Modern neural language models (LMs) are powerful tools for modeling human sentence production and comprehension, and their internal representations are remarkably well-aligned with representations of language in the human brain. But to achieve these results, LMs must be trained in distinctly un-human-like ways-requiring orders of magnitude more language data than children receive during development, and without perceptual or social context. Do models trained more naturalistically-with grounded supervision-exhibit more human-like language learning? We investigate this question in the context of word learning, a key sub-task in language acquisition. We train a diverse set of LM architectures, with and without auxiliary visual supervision, on datasets of varying scales. We then evaluate these models’ learning of syntactic categories, lexical relations, semantic features, word similarity, and alignment with human neural representations. We find that visual supervision can indeed improve the efficiency of word learning. However, these improvements are limited: they are present almost exclusively in the low-data regime, and sometimes canceled out by the inclusion of rich distributional signals from text. The information conveyed by text and images is not redundant-models mainly driven by visual information yield qualitatively different from those mainly driven by word co-occurrences. However, our results suggest that current multimodal modeling approaches fail to effectively leverage visual information to build human-like word representations from human-scale data. © 2024 Association for Computational Linguistics.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Learning the Jawa Culture by using Android Smartphone 2

Learning the Jawa Culture by using Android Smartphone

引用

2nd International Conference on Artificial Intelligence and Smart Energy, ICAIS 2022

作者： Antonio, Andrew Kezyanto, Julio Warnars, Harco Leslie Hendric Spits Bina Nusantara University School of Computer Science Computer Science Department Jakarta11480 Indonesia Bina Nusantara University Binus Graduate Program - Computer Science Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9781665400527

Javanese culture is a culture that the Javanese own, and the Javanese have started to exist with their culture since before the era of the Majapahit kingdom, and the Javanese are famous for their ability to sail the oceans. In addition, the many temples built, which are currently historical sites, prove that the Javanese are a large ethnic group that inhabits the island of Java. Other Javanese cultures include Javanese handicrafts such as kris, Javanese dance, Javanese batik art, Javanese wayang art, Javanese carving art, Javanese songs, and Javanese writings. All these cultures are part of the wealth of the unitary state of the Republic of Indonesia that must be preserved and passed down from one generation to another, and current technological developments tend to make the current generation less aware of preserving the culture of their ancestors. Therefore, we use technology to create an application for learning Javanese culture by applying game elements as entertainment elements. The business process of this application is modeled using use case diagrams and class diagrams to describe the database relationships. This mobile application is built using android studio and uses a MySQL database. © 2022 IEEE.

关键词： Android (operating system)

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：