Felipe Viegas

Felipe Viegas

Auto-generated profile, based on publicly available information

Felipe Viegas

UFMG

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
13 July 2023 in ACM Computing Surveys
Verified
Journal Article
Open access
WCWashington Cunha
Washington Cunha
FVFelipe Viegas
Felipe Viegas
CFCelso França
Celso França
TRThierson Rosa
Thierson Rosa
LRLeonardo Rocha
Leonardo Rocha
MGMarcos André Gonçalves
Marcos André Gonçalves
Progress in natural language processing has been dictated by the rule of more : more data, more computing power, more complexity, best exemplified by deep learning Transformers. However, training (or fine-tuning) large dense models for specific applications usually requires significant amounts of computing resources. One way to ameliorate this problem is through data engineering instead of the algorithmic or hardware perspectives. Our focus here is an under-investigated data engineering technique, with enormous potential in the current scenario – Instance Selection (IS) (a.k.a. Selective Sampling, Prototype Selection). The IS goal is to reduce the training set size by removing noisy or redundant instances while maintaining or improving the effectiveness (accuracy) of the trained models and reducing the training process cost. We survey classical and recent state-of-the-art IS techniques and provide a scientifically sound comparison of IS methods applied to an essential natural language processing task—Automatic Text Classification (ATC). IS methods have been normally applied to small tabular datasets and have not been systematically compared in ATC. We consider several neural and non-neural state-of-the-art ATC solutions and many datasets. We answer several research questions based on tradeoffs induced by a tripod (training set reduction, effectiveness, and efficiency). Our answers reveal an enormous unfulfilled potential for IS solutions. Specially, we show that in 12 out of 19 datasets, specific IS methods—namely, Condensed Nearest Neighbor, Local Set-based Smoother, and Local Set Border Selector—can reduce the size of the training set without effectiveness losses. Furthermore, in the case of fine-tuning the Transformer methods, the IS methods reduce the amount of data needed, without losing effectiveness and with considerable training-time gains.

1 January 2018 in Neurocomputing
Verified
Journal Article
Closed access
FVFelipe Viegas
Felipe Viegas
LRLeonardo Rocha
Leonardo Rocha
MGMarcos Gonçalves
Marcos Gonçalves
FMFernando Mourão
Fernando Mourão
GSGiovanni Sá
Giovanni Sá
ISIsac Sandin
Isac Sandin

1 January 2018
Verified
Conference Paper
Open access
WLWashington Luiz
Washington Luiz
FVFelipe Viegas
Felipe Viegas
RARafael Alencar
Rafael Alencar
FMFernando Mourão
Fernando Mourão
TSThiago Salles
Thiago Salles
LRLeonardo Rocha
Leonardo Rocha

1 July 2020 in Information Processing & Management
Verified
Journal Article
Closed access
WCWashington Cunha
Washington Cunha
SCSérgio Canuto
Sérgio Canuto
FVFelipe Viegas
Felipe Viegas
TSThiago Salles
Thiago Salles
CGChristian Gomes
Christian Gomes
LRLeonardo Rocha
Leonardo Rocha

30 January 2019
Verified
Conference Paper
Open access
FVFelipe Viegas
Felipe Viegas
SCSérgio Canuto
Sérgio Canuto
CGChristian Gomes
Christian Gomes
WLWashington Luiz
Washington Luiz
TRThierson Rosa
Thierson Rosa
MGMarcos André Gonçalves
Marcos André Gonçalves

1 January 2020
Verified
Conference Paper
Open access
PDF file
FVFelipe Viegas
Felipe Viegas
WCWashington Cunha
Washington Cunha
CGChristian Gomes
Christian Gomes
APAntônio Pereira
Antônio Pereira
LRLeonardo Rocha
Leonardo Rocha
MGMarcos Goncalves
Marcos Goncalves
Felipe Viegas, Washington Cunha, Christian Gomes, Antônio Pereira, Leonardo Rocha, Marcos Goncalves. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020.