Skip to content
Free planNo credit card requiredStart writing
AI for Research home
AI for Research home
Dashboard
Sign inGet started

Search 250 million research papers

Read the abstract, see who wrote it and who cites it, and get the citation in APA, MLA, Chicago or Harvard.

About 6,254 papers for “External Data Representation”

External Data-Enhanced Meta-Representation for Adaptive Probabilistic Load Forecasting

Haoran Li, Muhao Guo, Marija Ilić et al. · ArXiv.org · 2025

Accurate residential load forecasting is critical for power system reliability with rising renewable integration and demand-side flexibility. However, most statistical and machine learning models treat external factors, such as weather, calendar effects, and pricing, as extra input, ignoring their heterogeneity, and thus limiting the extraction of useful external information. We propose a paradigm shift: external data should serve as meta-knowledge to dynamically adapt the forecasting model itself. Based on this idea, we design a meta-representation framework using hypernetworks that modulate selected parameters of a base Deep Learning (DL) model in response to external conditions. This provides both expressivity and adaptability. We further integrate a Mixture-of-Experts (MoE) mechanism to enhance efficiency through selective expert activation, while improving robustness by filtering redundant external inputs. The resulting model, dubbed as a Meta Mixture of Experts for External data (M2oE2), achieves substantial improvements in accuracy and robustness with limited additional overhead, outperforming existing state-of-the-art methods in diverse load datasets. The dataset and source code are publicly available at https://github.com/haorandd/M2oE2\_load\_forecast.git.

Cited by 0Open access

Adversarial Speaker Disentanglement Using Unannotated External Data for Self-supervised Representation-based Voice Conversion

Xintao Zhao, Shuai Wang, Chao Yang et al. · 2023

Nowadays, recognition-synthesis-based methods have been quite popular with voice conversion (VC). By introducing linguistics features with good disentangling characters extracted from an automatic speech recognition (ASR) model, the VC performance achieved considerable breakthroughs. Recently, self-supervised learning (SSL) methods trained with a large-scale unannotated speech corpus have been applied to downstream tasks focusing on the content information, which is suitable for VC tasks. However, a huge amount of speaker information in SSL representations degrades timbre similarity and the quality of converted speech significantly. To address this problem, we proposed a high-similarity any-to-one voice conversion method with the input of SSL representations. We incorporated adversarial training mechanisms in the synthesis module using external unannotated corpora. Two auxiliary discriminators were trained to distinguish whether a sequence of mel-spectrograms has been converted by the acoustic model and whether a sequence of content embeddings contains speaker information from external corpora. Experimental results show that our proposed method achieves comparable similarity and higher naturalness than the supervised method, which needs a huge amount of annotated corpora for training and is applicable to improve similarity for VC methods with other SSL representations as input.

Cited by 7

Adversarial Speaker Disentanglement Using Unannotated External Data for Self-supervised Representation Based Voice Conversion

Xintao Zhao, Shuai Wang, Chao Yang et al. · arXiv (Cornell University) · 2023

Nowadays, recognition-synthesis-based methods have been quite popular with voice conversion (VC). By introducing linguistics features with good disentangling characters extracted from an automatic speech recognition (ASR) model, the VC performance achieved considerable breakthroughs. Recently, self-supervised learning (SSL) methods trained with a large-scale unannotated speech corpus have been applied to downstream tasks focusing on the content information, which is suitable for VC tasks. However, a huge amount of speaker information in SSL representations degrades timbre similarity and the quality of converted speech significantly. To address this problem, we proposed a high-similarity any-to-one voice conversion method with the input of SSL representations. We incorporated adversarial training mechanisms in the synthesis module using external unannotated corpora. Two auxiliary discriminators were trained to distinguish whether a sequence of mel-spectrograms has been converted by the acoustic model and whether a sequence of content embeddings contains speaker information from external corpora. Experimental results show that our proposed method achieves comparable similarity and higher naturalness than the supervised method, which needs a huge amount of annotated corpora for training and is applicable to improve similarity for VC methods with other SSL representations as input.

Cited by 0Open access

Students’ mathematical representation skills in solving numeracy problems with profession context

Dimas Danar Septiadi, Ikha Yuliati, Rany Widyastuti · Desimal Jurnal Matematika · 2022

This research aims to describe students’ mathematical representation skills in solving numeracy problem with profession context. This is qualitative descriptive research with 12 students for its research subjects. There are four steps in analyzing the data, those are data collection, data condensation, data display, and conclusions. To check the validity, researchers used techniques triangulation. This research shows that, among those 12 subjects, there are four types of model in solving numeracy problem with profession context. Those are first, students use visual and external representation; second, they perform visual, verbal, math model, and internal representation; third, they use verbal, visual, and external representation; and fourth of all, they use verbal and external representation. To make this result fruitfull, other researchers can try to design mathematics learning activities to train students’ mathematical representation skill such that when students solve numeracy problems, they will not face any difficulties.

Cited by 1Open access

State Estimation Based External Network Solution for on-line Security Analysis

Kenneth Geisler, Anjan Bose · IEEE Power Engineering Review · 1983

A major difficulty in on-line security analysis of power systems is the representation of the external but interconnected network from which very little or no real time data are available. One method used for the determination of the external system state is to extrapolate or assume real time data for a large external network representation, which is then solved by a power flow in such a way that the mismatches at the boundary between this external system solution and the state estimated internal network is minimized. This paper presents a method that extends the state estimation techniques, instead of using the power flow for the external network solution. The results indicate improved solution flexibility, sensitivity, error detection and location abilities, and significant reduction in boundary errors.

Cited by 3

Chinese Named Entity Recognition for Automobile Fault Texts Based on External Context Retrieving and Adversarial Training

Shuhai Wang, Linfu Sun · Entropy · 2025

Identifying key concepts in automobile fault texts is crucial for understanding fault causes and enabling diagnosis. However, effective mining tools are lacking, leaving much latent information unexplored. To solve the problem, this paper proposes Chinese named entity recognition for automobile fault texts based on external context retrieval and adversarial training. First, we retrieve external contexts by using a search engine. Then, the input sentence and its external contexts are respectively fed into Lexicon Enhanced BERT to improve the text embedding representation. Furthermore, the input sentence and its external contexts embedding representation are fused through the attention mechanism. Then, adversarial samples are generated by adding perturbations to the fusion vector representation. Finally, the fusion vector representation and adversarial samples are input into the BiLSTM-CRF layer as training data for entity labeling. Our model is evaluated on the automotive fault datasets, Weibo and Resume datasets, and achieves state-of-the-art results.

Cited by 2Open access

Application of Topic Knowledge in Graduate EFL Academic Writing in China

Jin Bi, Xiaoqing Qin · Chinese Journal of Applied Linguistics · 2012

The present study examines the relationship of knowledge resources of the topic and their representation to the self-evaluation of content in academic writing. Factor analysis of the questionnaire data gathered from 403 non-English-major graduate students yields four types of knowledge resources and eight types of knowledge representation strategies. Data analysis shows that personal experience is the most preferred knowledge resource, and that the most frequently used strategies of knowledge representation range from subjective perspective, contrastive evidence, objective statement, paraphrasing, and comprehensive processing to evidence evaluation. Further correlation analysis indicates that content evaluation is significantly correlated to three external knowledge resources (authoritative evidence, multi-source evidence and professional knowledge) and five external knowledge representation strategies (objective statement, paraphrasing, comprehensive processing, evidence evaluation, and objective perspective).

Cited by 1

Representing Programs as Hypertext

Kurt Nørmark, Kasper Østerbye · 1994

This paper discusses the representation of programs as hypertext. Representing programs as hypertext confronts us with the issue of which program structures are to be represented by the external node and link structure, and which program structures are to be represented internal to the nodes. The concept of anchors plays an important role in the interface between the external and the internal structure. We will examine five different models for the interplay between the internal and external structures. 1 Introduction This paper is obtained by ftp from ftp.iesd.auc.dk:/pup/reports/papers/normark94a. It is an exact copy of the paper published as [7] The program representation (sometimes called the internal program representation for clarity) is the data structure which represents a program. The program representation is seen as a contrast to the external presentation of the program on the screen, which sometimes is called a view. Text and abstract syntax trees (ASTs) are common...

Cited by 0

ActionFlowNet: Learning Motion Representation for Action Recognition

Joe Yue-Hei Ng, Jonghyun Choi, Jan Neumann et al. · 2018

We present a data-efficient representation learning approach to learn video representation with small amount of labeled data. We propose a multitask learning model ActionFlowNet to train a single stream network directly from raw pixels to jointly estimate optical flow while recognizing actions with convolutional neural networks, capturing both appearance and motion in a single model. Our model effectively learns video representation from motion information on unlabeled videos. Our model significantly improves action recognition accuracy by a large margin (23.6%) compared to state-of-the-art CNN-based unsupervised representation learning methods trained without external large scale data and additional optical flow input. Without pretraining on large external labeled datasets, our model, by well exploiting the motion information, achieves competitive recognition accuracy to the models trained with large labeled datasets such as ImageNet and Sport-1M.

Cited by 106

Identifying gaps for research prioritisation: Global burden of external causes of injury as reflected in the Cochrane Database of Systematic Reviews

Karimkhani, Chante, Trikha, Ritika, Aksut, Baran et al. · Figshare · 2016

Importance Burden of disease should impact research prioritisation. Objective: To analyse the Cochrane Database of Systematic Reviews (CDSR) and determine whether systematic reviews and protocols accurately represent disease burden, as measured by disability-adjusted life years (DALYs) from the Global Burden of Disease (GBD) 2010 Study. Methods: Two investigators collected GBD disability metrics for 12 external causes of injury in the GBD 2010 Study. These external causes were then assessed for systematic review and protocol representation in CDSR. Data was collected during the month of April 2015. There were no study participants aside from the researchers. Percentage of total 2010 DALYs, 2010 DALY rank, and median DALY percent change from 1990 to 2010 of the 12 external causes of injury were compared with CDSR representation of systematic reviews and protocols. Data were analysed for correlation using Spearman rank correlation. Results: Eleven of the 12 causes were represented by at least one systematic review or protocol in CDSR; the category collective violence and legal intervention had no representation in CDSR. Correlation testing revealed a strong positive correlation that was statistically significant. Representation of road injury; interpersonal violence; fire, heat, and hot substances; mechanical forces; poisonings, adverse effect of medical treatment, and animal contact was well aligned with respect to DALY. Representation of falls was greater compared to DALY, while self-harm, exposure to forces of nature, and other transport injury representation was lower compared to DALY. Conclusions and relevance CDSR representation of external causes of injury strongly correlates with disease burden. The number of systematic reviews and protocols was well aligned for seven out of 12 causes of injury. These results provide high-quality and transparent data that may guide future prioritisation decisions.

Cited by 0Open access

Identifying gaps for research prioritisation: Global burden of external causes of injury as reflected in the Cochrane Database of Systematic Reviews

Karimkhani, Chante, Trikha, Ritika, Aksut, Baran et al. · Swinburne Research Bank (Swinburne University of Technology) · 2016

Importance Burden of disease should impact research prioritisation. Objective: To analyse the Cochrane Database of Systematic Reviews (CDSR) and determine whether systematic reviews and protocols accurately represent disease burden, as measured by disability-adjusted life years (DALYs) from the Global Burden of Disease (GBD) 2010 Study. Methods: Two investigators collected GBD disability metrics for 12 external causes of injury in the GBD 2010 Study. These external causes were then assessed for systematic review and protocol representation in CDSR. Data was collected during the month of April 2015. There were no study participants aside from the researchers. Percentage of total 2010 DALYs, 2010 DALY rank, and median DALY percent change from 1990 to 2010 of the 12 external causes of injury were compared with CDSR representation of systematic reviews and protocols. Data were analysed for correlation using Spearman rank correlation. Results: Eleven of the 12 causes were represented by at least one systematic review or protocol in CDSR; the category collective violence and legal intervention had no representation in CDSR. Correlation testing revealed a strong positive correlation that was statistically significant. Representation of road injury; interpersonal violence; fire, heat, and hot substances; mechanical forces; poisonings, adverse effect of medical treatment, and animal contact was well aligned with respect to DALY. Representation of falls was greater compared to DALY, while self-harm, exposure to forces of nature, and other transport injury representation was lower compared to DALY. Conclusions and relevance CDSR representation of external causes of injury strongly correlates with disease burden. The number of systematic reviews and protocols was well aligned for seven out of 12 causes of injury. These results provide high-quality and transparent data that may guide future prioritisation decisions.

Cited by 0Open access

PIC-INR: Scan-specific unsupervised IVIM parameter mapping using physics-informed convolutional implicit neural representation

Jiechao Wang, Lu Wang, Congbo Cai et al. · Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025

Motivation: To eliminate the reliance on abundant high-quality reference maps and to address the inconsistent distribution induced from external training data in learning-based IVIM parameter mapping. Goal(s): To develop a scan-specific unsupervised learning method that accurately maps IVIM parameters from noisy diffusion-weighted images without depending on external training data. Approach: The proposed method employs a physics-informed mechanism and convolutional implicit neural representation to generate IVIM parameter maps solely from diffusion-weighted images themselves. Results: Numerical phantom and in vivo human brain results demonstrate that our proposed method generates accurate IVIM parameter maps with improved quality. Impact: Our scan-specific unsupervised method, employing physics-informed convolutional implicit neural representation, has successfully achieved accurate IVIM parameter mapping without relying on reference maps and external data during network training.

Cited by 0

Disentangled and Robust Representation Learning for Bragging Classification in Social Media

Xiang Li, Yucheng Zhou · arXiv (Cornell University) · 2022

Researching bragging behavior on social media arouses interest of computational (socio) linguists. However, existing bragging classification datasets suffer from a serious data imbalance issue. Because labeling a data-balance dataset is expensive, most methods introduce external knowledge to improve model learning. Nevertheless, such methods inevitably introduce noise and non-relevance information from external knowledge. To overcome the drawback, we propose a novel bragging classification method with disentangle-based representation augmentation and domain-aware adversarial strategy. Specifically, model learns to disentangle and reconstruct representation and generate augmented features via disentangle-based representation augmentation. Moreover, domain-aware adversarial strategy aims to constrain domain of augmented features to improve their robustness. Experimental results demonstrate that our method achieves state-of-the-art performance compared to other methods.

Cited by 0Open access

The impact of creative accounting on the faithful representation of financial Reporting: A case of commercial banks in Zimbabwe

Onesmo Guti, Kumbirai W Mukonori · International Journal of Research in Management · 2022

The study aimed to explore the impact of creative accounting on the faithful representation of financial reporting in the case of commercial banks in Zimbabwe. The specific objectives of the study were to explore the effect of government regulation in combating creative accounting on the faithful representation of financial reports and ascertain the effect of external audit on creative accounting as a tool to improve the faithful representation of financial reports. The research methodology used was quantitative through a systematic investigation, and a pragmatic philosophy was adopted. The population of the study consisted of commercial banks in Zimbabwe. The sample size was determined with the help of Krejcie and Morgan (1970). The data was collected using questionnaires distributed online. The study processed data using SPSS and analyzed the data using descriptive statistics, regression, and Pearson’s correlation coefficient. The study concluded that, government regulations, if implemented properly, have the potential to improve the state of faithful representation of financial reports; if the government of the country is corrupt, it negatively affects faithful representation of financial statements, and there is a positive relationship between the quality of external audits and faithful representation of financial reports.

Cited by 0Open access

Adaptive selection: External knowledge and internal context representation for few-shot intention recognition

Jingfan Tang, Pengfei Li, Xuefeng Zhang et al. · Alexandria Engineering Journal · 2025

Intention recognition models usually need to be trained on a large number of data, and when new intentions appear in the discriminant field, only a small amount of data is available. The method based on few-shot learning can well deal with this problem. In existing methods, the intent representation is often obtained by summing or averaging the sample representations. This will make the distance between the intents too close and cause the classification to fail. In this paper, we propose a few-shot intent recognition method based on adaptive selection of external knowledge and internal context representation. This method combines external knowledge representation extracted from a knowledge graph with internal context representation within sentences. It achieves hierarchical modeling of sentences through a dynamic routing algorithm and adaptively selects entity semantic representation information to enhance the semantic representation of entities in sentences. Experiments conducted on the Banking 77 dataset demonstrate that this method outperforms existing methods in both cross-domain and same-domain scenarios, particularly achieving an accuracy rate of 80.73 % under the 3-way 30-shot setting.

Cited by 0Open access