检索结果-内蒙古大学图书馆

A Dynamic Multi-Scale Voxel Flow Network for Video Prediction

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Hu, Xiaotao Huang, Zhewei Huang, Ailin Xu, Jun Zhou, Shuchang College of Computer Science Nankai University China Megvii Technology China Wuhan University China School of Statistics and Data Science Nankai University China

The performance of video prediction has been greatly boosted by advanced deep neural networks. However, most of the current methods suffer from large model sizes and require extra inputs, e.g., semantic/depth maps, for promising performance. For efficiency consideration, in this paper, we propose a Dynamic Multi-scale Voxel Flow Network (DMVFN) to achieve better video prediction performance at lower computational costs with only RGB images, than previous methods. The core of our DMVFN is a differentiable routing module that can effectively perceive the motion scales of video frames. Once trained, our DMVFN selects adaptive sub-networks for different inputs at the inference stage. Experiments on several benchmarks demonstrate that our DMVFN is an order of magnitude faster than Deep Voxel Flow [32] and surpasses the state-of-the-art iterative-based OPT [59] on generated image quality. Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics

Facilitating Interdisciplinary Knowledge Transfer with Research Paper Recommender Systems

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Cunningham, Eoghan Smyth, Barry Greene, Derek School of Computer Science University College Dublin Dublin Ireland Insight Centre for Data Analytics Dublin Ireland

In the recommender systems literature, novelty and diversity have been identified as key properties of useful recommendations. However, these properties have received limited attention in the specific sub-field of research paper recommender systems. In this work, we argue for the importance of offering novel and diverse research paper recommendations to scientists in order to reduce siloed reading, break down filter bubbles, and promote interdisciplinary research. We propose a novel framework for evaluating the novelty and diversity of research paper recommendations that leverages methods from network analysis and natural language processing. Using this framework, we show that the choice of representational method within a larger research paper recommendation system can have a measurable impact on the nature of downstream recommendations, specifically on their novelty and diversity. We describe a paper embedding method that provides more novel and diverse recommendations without sacrificing precision, compared to other state-of-the-art baselines. © 2023, CC BY.

关键词： Knowledge representation

Deep Residual and Classified Neural Networks for Inverse Halftoning

学校读者我要写书评

暂无评论

Deep Residual and Classified Neural Networks for Inverse Hal...

Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA)

作者： Jing-Ming Guo S. Sankarasrinivasan Let Viet Hung Wei Liu Department of Electrical Engineering National Taiwan University of Science and Technology Taipei Taiwan School of Data and Computer Science Sun Yat-Sen University China

Inverse Halftoning is an ill-posed problem which restores a continuous-tone image from a halftone image. Many conventional inverse halftoning methods have tried to solve this problem, yet the recovered images still suffer several unwanted artifacts and fine details losses. In addition, recent deep neural network-based approaches have shown their advantages on restoration of the high-quality images with rich textures and detailed information. However, it is truly challenging for these deep learning methods to reconstruct a variety of different halftone patterns. For instance, the model trained with the halftone patterns of homogenous distribution cannot perform ideally for high structural information patterns. To solve this problem, an inverse halftoning based on deep residual neural network (DRNN) and variance classification is proposed. The proposed method utilizes benefits of progressive learning concept involving two main stages: First, the DRNN extracts numerous intrinsic features of an image, and significantly removes the halftone patterns. Subsequently, consecutive deep residual blocks are integrated to network restoring the fine details with good accuracy. Consequently, the proposed model comprises the integration of various DRNNs which are trained over various statistical ranges with respect to the statistics of halftone patches. Comprehensive experimental results demonstrate that the proposed deep learning-based technique significantly outperforms not only the conventional methods but also deep learning approaches.

关键词：

Learning dynamics of gradient descent optimization in deep neural networks

学校读者我要写书评

暂无评论

science China(Information sciences) 2021年第5期64卷 17-31页

作者： Wei WU Xiaoyuan JING Wencai DU Guoliang CHEN School of Computer Science Wuhan University Institute of Data Science City University of Macau College of Computer Science and Software Engineering Shenzhen University

Stochastic gradient descent(SGD)-based optimizers play a key role in most deep learning models,yet the learning dynamics of the complex model remain obscure. SGD is the basic tool to optimize model parameters, and is improved in many derived forms including SGD momentum and Nesterov accelerated gradient(NAG). However, the learning dynamics of optimizer parameters have seldom been studied. We propose to understand the model dynamics from the perspective of control theory. We use the status transfer function to approximate parameter dynamics for different optimizers as the first-or second-order control system, thus explaining how the parameters theoretically affect the stability and convergence time of deep learning models, and verify our findings by numerical experiments.

关键词： learning dynamics deep neural networks gradient descent control model transfer function

Domain Mastery Benchmark: An Ever-Updating Benchmark for Evaluating Holistic Domain Knowledge of Large Language Model A Preliminary Release

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Gu, Zhouhong Zhu, Xiaoxuan Ye, Haoning Zhang, Lin Xiong, Zhuozhi Li, Zihan He, Qianyu Jiang, Sihang Feng, Hongwei Xiao, Yanghua Shanghai Key Laboratory of Data Science School of Computer Science Fudan University China Fudan-Aishu Cognitive Intelligence Joint Research Center China

Domain knowledge refers to the in-depth understanding, expertise, and familiarity with a specific subject, industry, field, or area of special interest. The existing benchmarks are all lack of an overall design for domain knowledge evaluation. Holding the belief that the real ability of domain language understanding can only be fairly evaluated by an comprehensive and in-depth benchmark, we introduces the Domma, a Domain Mastery Benchmark. DomMa targets at testing Large Language Models (LLMs) on their domain knowledge understanding, it features extensive domain coverage, large data volume, and a continually updated data set based on Chinese 112 first-level subject classifications. DomMa consist of 100,000 questions in both Chinese and English sourced from graduate entrance examinations and undergraduate exams in Chinese college. We have also propose designs to make benchmark and evaluation process more suitable to LLMs. Our benchmark will be release in https://***/MikeGu721/DomainMasteryBenchmark. © 2023, CC BY.

关键词： Domain Knowledge

Cognitive Bias and Reassignment: Who Can Contribute High Quality LLM data 39

学校读者我要写书评

暂无评论

Cognitive Bias and Reassignment: Who Can Contribute High Qua...

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Gao, Yunfan Xiong, Yun Hu, Zhongyuan Zhang, Yiming Wang, Meng Wang, Haofen Shanghai Research Institute for Intelligent Autonomous Systems Tongji University China Shanghai Key Laboratory of Data Science School of Computer Science Fudan University China College of Design and Innovation Tongji University China

ISBN: (纸本)157735897X

The rapid development of Large Language Models has highlighted the urgent need for large-scale, high-quality, and diverse data. We have launched an LLM data co-creation platform aimed at bringing together a wide range of participants to contribute data. Within six months, the platform has attracted over 10,000 participants who contributed more than 150,000 data entries across more than 200 tasks. An observable user cohort was constructed around the question, "Who is the best data contributor?" along with sub-questions concerning user preferences, task competence, and more. Through a detailed analysis of data contributors, this paper reveals several data collection patterns related to human factors. It reveals that contributors who provide high-quality data often do not meet initial expectations, as their behavior exhibits typical characteristics of the Dunning-Kruger effect. This paper examined the cognitive bias between users’ self-assessment and actual abilities, where individuals tend to overestimate their capabilities in certain tasks, leading to a decreased willingness to continue contributing and a consequent waste of human resources. To address this issue, we propose a task reassignment method based on multi-task fine-tuning of small language models (SLMs) to better align user groups with appropriate task types. After the reallocation, we observed a significant increase in user engagement and platform benefits, along with improved overall platform efficiency. The versatility of this method makes it applicable to broader data collection scenarios. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词：

Learning Universal Multi-level Market Irrationality Factors to Improve Stock Return Forecasting 25

学校读者我要写书评

暂无评论

Learning Universal Multi-level Market Irrationality Factors ...

Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and data Mining V.1

作者： Chen Yang Jingyuan Wang Xiaohan Jiang Junjie Wu School of Computer Science and Engineering Beihang University Beijing China School of Computer Science and Engineering Beihang University Beijing China & School of Economics and Management Beihang University Beijing China MIIT Key Laboratory of Data Intelligence and Management Beihang University Beijing China & School of Economics and Management Beihang University Beijing China

ISBN: (纸本)9798400712456

Recent years have witnessed the perfect encounter of deep learning and quantitative trading has achieved great success in stock investment. Numerous deep learning-based models have been developed for forecasting stock returns, leveraging the powerful representation capabilities of neural networks to identify patterns and factors influencing stock prices. These models can effectively capture general patterns in the market, such as stock price trends, volume-price relationships, and time variations. However, the impact of special irrationality factors -- such as market sentiment, speculative behavior, market manipulation, and psychological biases -- has not been fully considered in existing deep stock forecasting models due to their relative abstraction as well as lack of explicit labels and data description. To fill this gap, we propose UMI, a Universal multi-level Market Irrationality factor model to enhance stock return forecasting. The UMI model learns factors that can reflect irrational behaviors in market from both individual stock and overall market levels. For the stock-level, UMI construct an estimated rational price for each stock, which is cointegrated with the stock's actual price. The discrepancy between the actual and the rational prices serves as a factor to indicate stock-level irrational events. Additionally, we define market-level irrational behaviors as anomalous synchronous fluctuations of stocks within a market. Using two self-supervised representation learning tasks, i.e., sub-market comparative learning and market synchronism prediction, the UMI model incorporates market-level irrationalities into a market representation vector, which is then used as the market-level irrationality factor. We also developed a forecasting model that captures both temporal and relational dependencies among stocks, accommodating the UMI factors. Extensive experiments on U.S. and Chinese stock markets with competitive baselines demonstrate our model's effectiveness a

关键词： deep learning

Towards Coarse-grained Visual Language Navigation Task Planning Enhanced by Event Knowledge Graph

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhao, Kaichen Liu, Haoyu Song, Yaoxian Li, Tiefeng Zhao, Haiquan Li, Zhixu Shanghai Key Laboratory of Data Science School of Computer Science Fudan University Shanghai China Research Center for Intelligent Robotics Zhejiang Lab Hangzhou China Zhejiang University Hangzhou China

Visual language navigation (VLN) is one of the important research in embodied AI. It aims to enable an agent to understand the surrounding environment and complete navigation tasks. VLN instructions could be categorized into coarse-grained and fine-grained commands. Fine-grained command describes a whole task with subtasks step-by-step. In contrast, coarse-grained command gives an abstract task description, which more suites human habits. Most existing work focuses on the former kind of instruction in VLN tasks, ignoring the latter abstract instructions belonging to daily life scenarios. To overcome the above challenge in abstract instruction, we attempt to consider coarse-grained instruction in VLN by event knowledge enhancement. Specifically, we first propose a prompt-based framework to extract an event knowledge graph (named VLN-EventKG) for VLN integrally over multiple mainstream benchmark datasets. Through small and large language model collaboration, we realize knowledge-enhanced navigation planning (named EventNav) for VLN tasks with coarse-grained instruction input. Additionally, we design a novel dynamic history backtracking module to correct potential error action planning in real time. Experimental results in various public benchmarks show our knowledge-enhanced method has superiority in coarse-grained-instruction VLN using our proposed VLN-EventKG with over 5% improvement in success rate. Our project is available at https://***/view/vln-eventkg © 2024, CC BY-NC-SA.

关键词： Knowledge graph

Surrogate-Assisted Neighborhood Search with only a Few Weight Vectors for Expensive Large-Scale Multiobjective Binary Optimization

学校读者我要写书评

暂无评论

IEEE Transactions on Evolutionary Computation 2024年

作者： Gu, Haoran Wang, Handing Mei, Yi Zhang, Mengjie Jin, Yaochu Xidian University School of Artificial Intelligence Xi'an710071 China Victoria University of Wellington Centre for Data Science and Artificial Intelligence School of Engineering and Computer Science Wellington6140 New Zealand Westlake University Trustworthy and General Artificial Intelligence Laboratory School of Engineering Hangzhou310030 China

Large-scale multiobjective binary optimization problems often occur in real-world applications, where the function evaluation can only be performed through computationally expensive simulations, which renders standard exact and heuristic methods ineffective. Aggregation-based surrogate-assisted multiobjective evolutionary algorithms have been developed and shown to be promising for solving such problems. They define a set of uniformly distributed weight vectors as search directions, on which all search resources are placed to perform the evolution. However, the Pareto fronts of large-scale multiobjective binary optimization problems are discrete and non-uniform. As a result, many weight vectors are useless, thus a lot of search resources are wasted. To address this challenge, we propose a surrogate-assisted neighborhood search (SANS) for expensive large-scale multiobjective binary optimization. SANS uses only a few weight vectors to save search resources while maintaining an adequate diversity. To further utilize the limited search resources, a Q-learning-based method is designed to dynamically allocate search resources to weight vectors. Furthermore, a surrogate-assisted variable neighborhood search is developed to speed up the search without getting trapped in a local optimum prematurely. To robustly and reliably predict the quality of the found solutions, global and local surrogate models are trained by different training samples and then work collaboratively. The experimental results have demonstrated the superiority of SANS over seven state-of-the-art algorithms on the multiobjective binary optimization problems with up to 1000 decision variables using only 500 real solution evaluations. © 1997-2012 IEEE.

关键词： Vectors