Machine learning researcher at Unesp working on topic modeling, graph neural networks, and context-aware retrieval
3
papers here
0
reviewed by Evolve
3
first-author
0
corresponding
—
works · OpenAlex
1
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
Marked as insufficient. The record shows only 5 works in the last three years and an i10-index of 0, too thin to score output; this axis would measure total publication volume and its consistency over time.
Impact—
Marked as insufficient. With an h-index of 1 and i10-index of 0, there are too few citations to score impact; this axis would measure citation counts, per-paper influence, and uptake of the work by other researchers.
Focus—
Marked as insufficient. Topic signals exist — Topic Modeling (3) and Advanced Graph Neural Networks (2), plus three single-paper topics — but the record is too short to score; this axis would measure how consistently output concentrates on a few themes.
Collaboration36/100
Scores 36 on Evolve's 0-100 collaboration axis. The OpenAlex record shows 5 distinct coauthors in the recent three-year window, and the catalog lists five frequent collaborators — Daniel Carlos Guimarães Pedronette, Gustavo Rosseto Letício, Vinicius Atsushi Sato Kawai, Lucas Pascotti Valem, and Andre Freitas. All three catalog papers list Almeida as first author with no corresponding-author role, consistent with a junior researcher embedded in a small, stable team rather than a broad network.
Activity52/100
Scores 52 on the activity axis. The raw record counts 5 works in the last three years, with per-year output rising from 2 works in 2025 to 3 in 2026, and all three Evolve-catalog papers appeared as arXiv preprints in 2026. This is a modest but accelerating pace typical of an early-stage researcher; the low h-index of 1 confirms the profile is still being established rather than dormant.
Reproducibility—
Marked as insufficient. None of the three catalog papers carries a reproducibility assessment (reproReadiness is null for all), so this axis cannot be scored; it would measure code, data, and setup availability that lets others re-run the reported experiments.
Research summary
Thiago César Castilho Almeida is an early-career researcher at São Paulo State University (Unesp), Brazil, working at the intersection of machine learning, information retrieval, and natural language processing. This profile is name-resolved from public OpenAlex bibliographic data (ORCID 0000-0002-2167-0463). The record is recent: an h-index of 1, an i10-index of 0, 5 works in the last three years, and per-year counts rising from 2 works in 2025 to 3 in 2026. Three first-authored arXiv preprints from 2026 anchor the Evolve catalog entry: a manifold-aware topic modeling approach via rank-based prototypes, context-aware interpretable representations for retrieval and graph convolutional network classification, and graph- and rank-based contextual embeddings for textual and multimedia data. The topic record mirrors this agenda — Topic Modeling (3 works) and Advanced Graph Neural Networks (2) lead, with single works in face and expression recognition, NLP techniques, and information retrieval and search behavior. Collaboration is concentrated: 5 distinct coauthors in the recent window, including Daniel Carlos Guimarães Pedronette, Lucas Pascotti Valem, Andre Freitas, Gustavo Rosseto Letício, and Vinicius Atsushi Sato Kawai. Across the catalog papers Almeida appears as first author without corresponding-author status, consistent with a researcher establishing independence within an established group. Impact and reproducibility axes are not yet scored; the profile should be read as an early-stage snapshot rather than a mature record.
Collaboration pattern
Collaboration is compact and stable: 5 distinct coauthors appear in the recent three-year window (collaboration score 36), and the profile lists five frequent collaborators — Daniel Carlos Guimarães Pedronette, Gustavo Rosseto Letício, Vinicius Atsushi Sato Kawai, Lucas Pascotti Valem, and Andre Freitas. All three catalog papers are first-authored by Almeida with no corresponding role, suggesting work within a mentor-led Unesp group rather than an independent or international network so far.
Research keywords
Highlights
A Manifold-Aware Topic Modeling Approach via Rank-Based Prototypes
First-authored 2026 arXiv preprint and the catalog entry behind the profile's most frequent topic, Topic Modeling (3 works); the title indicates topic modeling that uses rank-based prototypes with manifold awareness. No Evolve verdict or reproducibility assessment is available yet.
Context-Aware Interpretable Representations for Retrieval and Graph Convolutional Network Classification
First-authored 2026 arXiv preprint connecting retrieval-side context to interpretable representations and GCN-based classification; it matches the profile's Advanced Graph Neural Networks topic (2 works) and Information Retrieval and Search Behavior topic. First author, non-corresponding; arXiv source.
Effective Graph and Rank-based Contextual Embeddings for Textual and Multimedia Data
First-authored 2026 arXiv preprint on graph- and rank-based contextual embeddings spanning text and multimedia; consistent with the profile's topic mix of graph neural networks, NLP techniques, and multimedia-oriented retrieval. No peer-reviewed venue is recorded yet.
Timeline
- 2025OpenAlex counts-by-year data records 2 works in 2025 — the earliest activity captured for this profile.
- 2026First-authored arXiv preprint 'A Manifold-Aware Topic Modeling Approach via Rank-Based Prototypes' appears; Topic Modeling is the profile's most frequent topic (3 works).
- 2026First-authored arXiv preprint 'Context-Aware Interpretable Representations for Retrieval and Graph Convolutional Network Classification' appears, matching the Advanced Graph Neural Networks topic (2 works).
- 2026First-authored arXiv preprint 'Effective Graph and Rank-based Contextual Embeddings for Textual and Multimedia Data' appears, bringing the recent three-year total to 5 works.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric metrics (h-index, i10-index, counts by year, topics) come from OpenAlex only; other databases may differ.
- All three catalog papers are 2026 arXiv preprints; no peer-reviewed venue, Evolve verdict, field label, or reproducibility assessment is available yet.
- Output, impact, focus, and reproducibility dimensions are unscored because the underlying data is insufficient at this stage.
- The profile is name-resolved from public bibliographic data (OpenAlex; ORCID 0000-0002-2167-0463); no biographical information was independently verified.