Iury Cleveston

Iury Cleveston

Auto-generated profile, based on publicly available information

Iury Cleveston

Universidade Estadual de Campinas

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
1 January 2025 in Texto Livre
Verified
Journal Article
Open access
PDF file
LMLavínia de Carvalho Moraes
Lavínia de Carvalho Moraes
ISIrene Cristina Silvério
Irene Cristina Silvério
RMRafael Alexandre Sousa Marques
Rafael Alexandre Sousa Marques
BABianca de Castro Anaia
Bianca de Castro Anaia
DPDandara Freitas de Paula
Dandara Freitas de Paula
RFRaquel Meister Ko Freitag
Raquel Meister Ko Freitag
Linguistic ambiguity continues to represent a significant challenge for natural language processing (NLP) systems, notwithstanding the advancements in architectures such as Transformers and BERT. Inspired by the recent success of instructional models like ChatGPT and Gemini (named Bard in 2023), this study aims to analyze and discuss linguistic ambiguity within these models, focusing on three types prevalent in Brazilian Portuguese: semantic, syntactic, and lexical ambiguity. We created a corpus comprising 120 sentences, both ambiguous and unambiguous, for classification, explanation, and disambiguation. The models’ capability to generate ambiguous sentences was also explored by soliciting sets of sentences for each type of ambiguity. The results underwent qualitative analysis, drawing on recognized linguistic references, and quantitative assessment based on the accuracy of the responses obtained. It was evidenced that even the most sophisticated models, such as ChatGPT and Gemini, exhibit errors and deficiencies in their responses, often providing inconsistent explanations. Furthermore, the accuracy peaked at 49.58%,indicating the need for descriptive studies for supervised learning.

19 December 2025 in Cadernos de Linguística
Verified
Journal Article
Open access
PDF file
RFRaquel Freitag
Raquel Freitag
MVMarcia dos Santos Machado Vieira
Marcia dos Santos Machado Vieira
JBJuliana Bertucci Bertucci Barbosa
Juliana Bertucci Bertucci Barbosa
MJMiguel Oliveira, Jr.
Miguel Oliveira, Jr.
CACleber Ataíde
Cleber Ataíde
VDValter de Carvalho Dias
Valter de Carvalho Dias
Generative artificial intelligence is based on large-scale language models (LLMs), which are trained with data most often collected without consent or in breach of copyright. LLMs are trained with billions of words and millions of parameters, but we don't know exactly which texts are selected in the training or which parameters are controlled. While unsupervised learning requires a large volume of data, demanding more and more computational costs and generating energy impacts, supervised learning with structured and tagged data can optimize this process; more than that: supervised learning with structured and tagged data resulting from language documentation projects can contribute directly to the National Artificial Intelligence Plan: “Develop advanced language models in Portuguese, with national data that encompasses our cultural, social and linguistic diversity, to strengthen sovereignty in AI.” In Brazil, in addition to Portuguese and its varieties, there are more than 250 other languages (indigenous, immigration, sign language), which are neglected in digital inclusion due to a lack of structured data. The consortium of laboratories and research groups in this INCT aims to prepare linguistic data for the training of LLMs, considering Brazil's linguistic diversity, with the development of a joint protocol for collecting linguistic data in the field, to be replicated in the groups and laboratories longitudinally, as well as transcription procedures, as well as procedures for transcribing, aligning and labeling linguistic data to create a data set that represents Brazilian linguistic diversity, and conducting studies on linguistic processing of diversity to fine-tune LLMs, helping to reduce asymmetries and prejudice resulting from training LLMs with translations from English. A inteligência artificial generativa é baseada em modelos de língua em larga escala (LLMs), que são treinados com dados na maioria das vezes coletados sem consentimento ou infringindo direitos autorais. LLMs são treinados com bilhões de palavras e milhões de parâmetros, mas não sabemos exatamente quais textos são selecionados no treinamento ou quais parâmetros são controlados. Enquanto o aprendizado não supervisionado requer um grande volume de dados, demandando cada vez mais custos computacionais e gerando impactos energéticos, o aprendizado supervisionado, com dados estruturados e etiquetados pode otimizar esse processo; mais do que isso: o aprendizado supervisionado com dados estruturados e etiquetados resultantes de projetos de documentação linguística podem contribuir diretamente para o Plano Nacional de Inteligência Artificial: “Desenvolver modelos avançados de linguagem em português, com dados nacionais que abarcam nossa diversidade cultural, social e linguística, para fortalecer a soberania em IA.” No Brasil, além do português e suas variedades, há mais de 250 outras línguas (indígenas, de imigração, sinalizadas), negligenciadas na inclusão digital por falta de dados estruturados. O consórcio de laboratórios e grupos de pesquisa neste INCT visa a preparação de dados linguísticos para o treinamento de LLMs, considerando a diversidade linguística brasileira, com o desenvolvimento de um protocolo conjunto de coleta de dados linguísticos em campo, a ser replicado nos grupos e laboratórios longitudinalmente, assim como procedimentos de transcrição, alinhamento e etiquetagem de dados linguísticos para a constituição de um conjunto de dados que represente a diversidade linguística brasileira, e a realização de estudos sobre processamento linguístico da diversidade para o ajuste fino dos LLMs, contribuindo para a redução de assimetrias e preconceito resultantes do treino de LLMs com traduções do inglês.

21 November 2023
Verified
Thesis
Open access
ICIury Cleveston
Iury Cleveston

25 April 2024 in arXiv
Verified
Preprint
Open access
PDF file
LMLavínia de Carvalho Moraes
Lavínia de Carvalho Moraes
ISIrene Cristina Silvério
Irene Cristina Silvério
RMRafael Alexandre Sousa Marques
Rafael Alexandre Sousa Marques
BABianca de Castro Anaia
Bianca de Castro Anaia
DPDandara Freitas de Paula
Dandara Freitas de Paula
RFRaquel Meister Ko Freitag
Raquel Meister Ko Freitag
Linguistic ambiguity continues to represent a significant challenge for natural language processing (NLP) systems, notwithstanding the advancements in architectures such as Transformers and BERT. Inspired by the recent success of instructional models like ChatGPT and Gemini (In 2023, the artificial intelligence was called Bard.), this study aims to analyze and discuss linguistic ambiguity within these models, focusing on three types prevalent in Brazilian Portuguese: semantic, syntactic, and lexical ambiguity. We create a corpus comprising 120 sentences, both ambiguous and unambiguous, for classification, explanation, and disambiguation. The models capability to generate ambiguous sentences was also explored by soliciting sets of sentences for each type of ambiguity. The results underwent qualitative analysis, drawing on recognized linguistic references, and quantitative assessment based on the accuracy of the responses obtained. It was evidenced that even the most sophisticated models, such as ChatGPT and Gemini, exhibit errors and deficiencies in their responses, with explanations often providing inconsistent. Furthermore, the accuracy peaked at 49.58 percent, indicating the need for descriptive studies for supervised learning.

27 September 2021
Verified
Book Chapter
Closed access
PDF file
ACAlana de Santana Correia
Alana de Santana Correia
ICIury Cleveston
Iury Cleveston
VSViviane Bonadia dos Santos
Viviane Bonadia dos Santos
Sandra AvilaSandra Avila
Sandra Avila
ECEsther Colombini
Esther Colombini

21 May 2025 in SciELO Preprints
Verified
Preprint
Open access
RFRaquel Meister Ko Freitag
Raquel Meister Ko Freitag
MVMarcia dos Santos Machado Vieira
Marcia dos Santos Machado Vieira
JBJuliana Bertucci Bertucci Barbosa
Juliana Bertucci Bertucci Barbosa
MJMiguel Oliveira, Jr.
Miguel Oliveira, Jr.
CACleber Ataíde
Cleber Ataíde
VDValter de Carvalho Dias
Valter de Carvalho Dias
Generative artificial intelligence is based on large-scale language models (LLMs), which are trained with data most often collected without consent or in breach of copyright. LLMs are trained with billions of words and millions of parameters, but we don't know exactly which texts are selected in the training or which parameters are controlled. While unsupervised learning requires a large volume of data, demanding more and more computational costs and generating energy impacts, supervised learning with structured and tagged data can optimize this process; more than that: supervised learning with structured and tagged data resulting from language documentation projects can contribute directly to the National Artificial Intelligence Plan: “Develop advanced language models in Portuguese, with national data that encompasses our cultural, social and linguistic diversity, to strengthen sovereignty in AI.” In Brazil, in addition to Portuguese and its varieties, there are more than 250 other languages (indigenous, immigration, sign language), which are neglected in digital inclusion due to a lack of structured data. The consortium of laboratories and research groups in this INCT aims to prepare linguistic data for the training of LLMs, considering Brazil's linguistic diversity, with the development of a joint protocol for collecting linguistic data in the field, to be replicated in the groups and laboratories longitudinally, as well as transcription procedures, as well as procedures for transcribing, aligning and labeling linguistic data to create a data set that represents Brazilian linguistic diversity, and conducting studies on linguistic processing of diversity to fine-tune LLMs, helping to reduce asymmetries and prejudice resulting from training LLMs with translations from English.