检索结果-内蒙古大学图书馆

Topic Modelling using Transfer learning: Issues and Challenges

学校读者我要写书评

暂无评论

Topic Modelling using Transfer Learning: Issues and Challeng...

Intelligent Control, Computing and Communications (IC3), International Conference on

作者： Rama Krishna K Kaipa Sandhya Praveen Gujjar J Raghavendra M Devadas Vani Hiremani Preethi Department of Artificial Intelligence and Machine Learning Impact college of Engineering and Applied Sciences Bengaluru India Department of Data Science Impact college of Engineering and applied sciences Bengaluru India Faculty of Management Studies JAIN (Deemed-to-be University) Bengaluru India Department of Information Technology Manipal Institute of Technology Bengaluru Manipal Academy of Higher Education (MAHE) Manipal India Symbiosis Institute of Technology Symbiosis International (Deemed) University Pune India

ISBN: (数字)9798331527495

ISBN: (纸本)9798331527501

A machine learning method, transfer learning, uses information from one job or area to enhance effectiveness in a related but distinct task or area. Transfer learning enables using pre-trained models that were previously learned using a sizable and varied dataset in the original domain instead of creating a model from the beginning for the intended activity. Usually, using a lesser dataset, the pre-trained model's representations and knowledge may be improved upon or applied to the intended goal or area. In natural language processing (NLP) and machine learning, topic modeling is an approach that finds and extracts the fundamental concepts or subjects from a group of documents or textual data. Despite prior experience with the aspects, this unsupervised learning method seeks to uncover the text's latent semantic structure. This paper's primary goal is to investigate the possibilities of transfer learning for text topic modeling. Additionally, to assess the correctness of the pre-trained model's performance. The paper also describes the issues and challenges associated with the text for performing the topic modeling using transfer learning. This paper delves into the methodologies and techniques of topic modeling using transfer learning. It explores how this novel approach transcends the constraints of conventional topic modeling, enabling the extraction of meaningful topics from data-rich source domains to improve topic modeling performance in data-scarce or specialized target domains. By integrating the principles of transfer learning into the topic modeling framework, we unlock the potential to enhance the interpretability, accuracy, and applicability of topic modeling across diverse domains and applications. This paper sheds light on the transformative power of transfer learning in topic modeling.

关键词： Performance evaluation Adaptation models Accuracy Computational modeling Transfer learning Semantics data models Natural language processing data mining Unsupervised learning

Optimal Kernel Choice for Score Function-based Causal Discovery

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wang, Wenjie Huang, Biwei Liu, Feng You, Xinge Liu, Tongliang Zhang, Kun Gong, Mingming School of Mathematics and Statistics The University of Melbourne Australia Department of Machine Learning Mohamed bin Zayed University of Artificial Intelligence United Arab Emirates University of California San Diego United States School of Computing and Information Systems The University of Melbourne Australia Huazhong University of Science and Technology China School of Computer Science Faculty of Engineering The University of Sydney Australia Department of Philosophy Carnegie Mellon University United States

Score-based methods have demonstrated their effectiveness in discovering causal relationships by scoring different causal structures based on their goodness of fit to the data. Recently, Huang et al. (2018) proposed a generalized score function that can handle general data distributions and causal relationships by modeling the relations in reproducing kernel Hilbert space (RKHS). The selection of an appropriate kernel within this score function is crucial for accurately characterizing causal relationships and ensuring precise causal discovery. However, the current method involves manual heuristic selection of kernel parameters, making the process tedious and less likely to ensure optimality. In this paper, we propose a kernel selection method within the generalized score function that automatically selects the optimal kernel that best fits the data. Specifically, we model the generative process of the variables involved in each step of the causal graph search procedure as a mixture of independent noise variables. Based on this model, we derive an automatic kernel selection method by maximizing the marginal likelihood of the variables involved in each search step. We conduct experiments on both synthetic data and real-world benchmarks, and the results demonstrate that our proposed method outperforms heuristic kernel selection methods. © 2024, CC BY.

关键词： Optimization

Disease and Medications Text Visualization Using Scattertext

学校读者我要写书评

暂无评论

Disease and Medications Text Visualization Using Scattertext

Intelligent Control, Computing and Communications (IC3), International Conference on

作者： Rama Krishna K Kaipa Sandhya Praveen Gujjar J Raghavendra M Devadas Vani Hiremani Sapna R Department of Artificial Intelligence and Machine Learning Impact college of Engineering and Applied Sciences Bengaluru India Department of Data Science Impact college of Engineering and applied sciences Bengaluru India Faculty of Management Studies JAIN (Deemed-to-be University) Bengaluru India Department of Information Technology Manipal Institute of Technology Bengaluru Manipal Academy of Higher Education (MAHE) Manipal India Symbiosis Institute of Technology Symbiosis International (Deemed) University Pune India

ISBN: (数字)9798331527495

ISBN: (纸本)9798331527501

Before text data can be analysed and visualised, it must be thoroughly cleaned due to its messy nature. data visualizations use the data to tell an engaging and simple-to-read story. That is what the Scattertext tool does. Scatter text allows which terms stand out in your corpora and how their frequency varies across categories. The dataset used in this paper includes reviews of drugs. This dataset offers a useful window into how people discuss their thoughts and experiences using various drugs. The drug names, user ratings, and the actual written reviews are among the details included in the dataset. This dataset consists of drug reviews collected from various sources, hence expressing a wide range of opinions and feelings. Each row in this dataset represents the experience a user had with a given medication and hence allows us to explore those emotions that are associated with any particular set of drugs. Understanding opinion expressed in the review about drugs might go a long way in helping decision-making processes in both healthcare and pharmaceuticals. The scatter plot distinguishes between labelled points and unlabelled ones by using a custom algorithm for Scattertext. By hovering their cursor over the labelled areas done client-side, users can interactively see which terms correspond to which points.

关键词： Drugs Reviews Decision making data visualization Medical services Market research Feature extraction Diseases Intelligent control

Analyzing the Effects of Different Urban and Rural Characteristics on Surface Urban Heat Island Intensity Using the Spatial Autoregressive Model

学校读者我要写书评

暂无评论

SSRN

SSRN 2023年

作者： Wongsi, Noppachai Wongsai, Sangdao Wanishsakpong, Wandee Suwanprasit, Chanida Department of Mathematics and Statistics Faculty of Science and Technology Thammasat University Pathumthani12121 Thailand Thammasat University Research Unit in Data Learning Thammasat University Pathumthani12121 Thailand Department of Statistics Faculty of Science Kasetsart University Bangkok10900 Thailand Department of Geography Faculty of Social Sciences Chiang Mai University Chiang Mai50200 Thailand College of Arts Media and Technology Chiang Mai University Chiang Mai50200 Thailand

This study introduces an approach for quantifying the SUHII using the generalized mixed spatial autoregressive (SAR) model, in the Eastern Economic Corridor of Thailand. A comparative analysis was performed using the mean differences and simple linear regression (SLR), SLR with kernel density estimation (KDE), and spatial autoregressive regression (SAR) models. A year of MODIS LST was the dependent variable. The independent variables were the ISA percentage (%ISA), derived from the Sentinel-2 image using the RF classifier, and the geographical locations. Our findings indicated that using the mean differences on pixels without water surface and altitude led to the underestimated SUHII. By fitting the SLR with KDE on the smoothed %ISA, the LST was overfitted. The SLR applied to the original dataset mitigated the overestimation issue, but could not address the spatial dependency. The SAR model revealed a significant reduction in the spatial dependency and enhanced the model accuracy without overfitting. The 95% confidence intervals for LST from the SAR were found to be twice as narrow as those from other models. To evaluate the spatial assumption, we suggested making a map of how the spatial residuals are spread out, showing if there are clusters or dispersed. © 2023, The Authors. All rights reserved.

关键词： Atmospheric temperature

GSCLIP: A Framework for Explaining Distribution Shifts in Natural Language

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhu, Zhiying Liang, Weixin Zou, James Department of Machine Learning Carnegie Mellon University PA United States Department of Computer Science Stanford University CA United States Department of Biomedical Data Science Stanford University CA United States Chan Zuckerberg Biohub San FranciscoCA United States

Helping end users comprehend the abstract distribution shifts can greatly facilitate AI deployment. Motivated by this, we propose a novel task, dataset explanation. Given two image data sets, dataset explanation aims to automatically point out their dataset-level distribution shifts with natural language. Current techniques for monitoring distribution shifts provide inadequate information to understand datasets with the goal of improving data quality. Therefore, we introduce GSCLIP, a training-free framework to solve the dataset explanation task. In GSCLIP, we propose the selector as the first quantitative evaluation method to identify explanations that are proper to summarize dataset shifts. Furthermore, we leverage this selector to demonstrate the superiority of a generator based on language model generation. Systematic evaluation on natural data shift verifies that GSCLIP, a combined system of a hybrid generator group and an efficient selector is not only easy-to-use but also powerful for dataset explanation at scale. Copyright © 2022, The Authors. All rights reserved.

关键词： machine learning

Probabilistic task modelling for meta-learning

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Nguyen, Cuong Do, Thanh-Toan Carneiro, Gustavo Australian Institute for Machine Learning University of Adelaide Australia Department of Data Science and AI Monash University Australia

We propose probabilistic task modelling - a generative probabilistic model for collections of tasks used in meta-learning. The proposed model combines variational auto-encoding and latent Dirichlet allocation to model each task as a mixture of Gaussian distribution in an embedding space. Such modelling provides an explicit representation of a task through its task-theme mixture. We present an efficient approximation inference technique based on variational inference method for empirical Bayes parameter estimation. We perform empirical evaluations to validate the task uncertainty and task distance produced by the proposed method through correlation diagrams of the prediction accuracy on testing tasks. We also carry out experiments of task selection in meta-learning to demonstrate how the task relatedness inferred from the proposed model help to facilitate meta-learning algorithms. © 2021, CC BY.

关键词： Mixtures

Be Intentional About Fairness!: Fairness, Size, and Multiplicity in the Rashomon Set

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Dai, Gordon Ravishankar, Pavan Yuan, Rachel Neill, Daniel B. Black, Emily Machine Learning for Good Laboratory Department of Computer Science NYU Courant Institute of Mathematical Sciences New York University United States Robert F. Wagner Graduate School of Public Service New York University United States Center for Urban Science and Progress Tandon School of Engineering New York University United States Center for Data Science New York University United States Department of Computer Science and Engineering Tandon School of Engineering New York University United States

When selecting a model from a set of equally performant models, how much unfairness can you really reduce? Is it important to be intentional about fairness when choosing among this set, or is arbitrarily choosing among the set of "good" models good enough? Recent work has highlighted that the phenomenon of model multiplicity—where multiple models with nearly identical predictive accuracy exist for the same task—has both positive and negative implications for fairness, from strengthening the enforcement of civil rights law in AI systems to showcasing arbitrariness in AI decision-making. Despite the enormous implications of model multiplicity, there is little work that explores the properties of sets of equally accurate models, or Rashomon sets, in general. In this paper, we present five main theoretical and methodological contributions which help us to understand the relatively unexplored properties of the Rashomon set, in particular with regards to fairness. Our contributions include methods for efficiently sampling models from this set and techniques for identifying the fairest models according to key fairness metrics such as statistical parity. We also derive the probability that an individual’s prediction will be flipped within the Rashomon set, as well as expressions for the set’s size and the distribution of error tolerance used across models. These results lead to policy-relevant takeaways, such as the importance of intentionally looking for fair models within the Rashomon set, and understanding which individuals or groups may be more susceptible to arbitrary decisions. © 2025, CC BY.

关键词：

From 5G to 6G: The Role of AI, machine learning, and Deep learning in Wireless Systems

学校读者我要写书评

暂无评论

From 5G to 6G: The Role of AI, Machine Learning, and Deep Le...

Sentiment Analysis and Deep learning (ICSADL), International Conference on

作者： Pradnyawant M. Gote Praveen Kumar Prateek Verma Prajyot Yesankar Adesh Pawar Saniya Saratkar Department of Computer Science & Design Faculty of Engineering and Technology Datta Meghe Institute of Higher Education and Research (DU) Wardha Maharashtra India Department of Computer Science & Medical Engineering Faculty of Engineering and Technology Datta Meghe Institute of Higher Education and Research (DU) Wardha Maharashtra India Department of Artificial Intelligence & Machine Learning Faculty of Engineering and Technology Datta Meghe Institute of Higher Education and Research (DU) Wardha Maharashtra India Department of Artificial Intelligence & Data Science Faculty of Engineering and Technology Datta Meghe Institute of Higher Education and Research (DU) Wardha Maharashtra India

ISBN: (数字)9798331523923

ISBN: (纸本)9798331523930

The swift progression of wireless communication technologies-specifically from 5G to 6G is an approach that could be the most significant revolutionary leap towards changing connectivity and data transmission forever. Additionally, algorithms of machine learning (ML) and deep learning (DL) in 5G and 6G can offer promising advancements toward predictive maintenance, dynamic spectrum management, and enhanced security. The review of the potentially transformative nature of these technologies underlines the need for future directions of research efforts called for in circumventing challenges towards making full capability exploitation of 6G possible.

关键词： 6G mobile communication Wireless communication Deep learning Tactile Internet Technological innovation Sentiment analysis 5G mobile communication Reviews Heuristic algorithms Security

Adaptive atomic basis sets

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Khan, Danish Ach, Maximillian L. von Lilienfeld, O. Anatole Chemical Physics Theory Group Department of Chemistry University of Toronto St. George Campus TorontoON Canada Vector Institute for Artificial Intelligence TorontoONM5S 1M1 Canada Department of Physics University of Toronto St. George Campus TorontoON Canada Munich Germany Acceleration Consortium University of Toronto TorontoON Canada Department of Materials Science and Engineering University of Toronto St. George Campus TorontoON Canada Machine Learning Group Technische Universität Berlin Institute for the Foundations of Learning and Data Berlin10587 Germany Berlin Institute for the Foundations of Learning and Data Berlin10587 Germany

Atomic basis sets are widely employed within quantum mechanics based simulations of matter. We introduce a machine learning model that adapts the basis set to the local chemical environment of each atom, prior to the start of self consistent field (SCF) calculations. In particular, as a proof of principle and because of their historic popularity, we have studied the Gaussian type orbitals from the Pople basis set, i.e. the STO-3G, 3-21G, 6-31G and 6-31G*. We adapt the basis by scaling the variance of the radial Gaussian functions leading to contraction or expansion of the atomic orbitals. A data set of optimal scaling factors for C, H, O, N and F were obtained by variational minimization of the Hartree-Fock (HF) energy of the smallest 2500 organic molecules from the QM9 database. Kernel ridge regression based machine learning (ML) prediction errors of the change in scaling decay rapidly with training set size, typically reaching less than 1% for training set size 2000. Overall, we find systematically lower variance, and consequently the larger training efficiencies, when going from hydrogen to carbon to nitrogen to oxygen. Using the scaled basis functions obtained from the ML model, we conducted HF calculations for the subsequent 30’000 molecules in QM9. In comparison to the corresponding default Pople basis set results we observed improved energetics in up to 99% of all cases. With respect to the larger basis set 6-311G(2df,2pd), atomization energy errors are lowered on average by ∼31, 107, 11, and 11 kcal/mol for STO-3G, 3-21G, 6-31G and 6-31G*, respectively — with negligible computational overhead. We illustrate the high transferability of adaptive basis sets for larger out-of-domain molecules relevant to addiction, diabetes, pain, aging. Generally, deviation from cc-pvQZ based HF results for counter-poise corrected atomization energies, HOMO-LUMO gaps, and dipole moments also decreases — in line with the variational principle. © 2024, CC BY.

关键词： Quantum theory