Daniel Carlos Guimarães Pedronette

Research profile

Research summary

Daniel Carlos Guimarães Pedronette is a computer scientist at Universidade Estadual Paulista (Unesp), Brazil, whose OpenAlex-linked profile centers on image and multimedia retrieval, rank-based contextualization, and graph-structured learning. The bibliometric snapshot records an h-index of 20, an i10-index of 47, and 46 works over the last three years, with publications logged every year from 2008 (2 works) to 2026 (14 works) and a peak of 19 works in 2025. Topic aggregation places Image Retrieval and Classification Techniques (92) and Advanced Image and Video Retrieval Techniques (88) at the core, followed by Information Retrieval and Search Behavior (61), Domain Adaptation and Few-Shot Learning (22), and Advanced Neural Network Applications (20). Three 2026 arXiv preprints in the Evolve catalog — context-aware interpretable representations for retrieval and graph convolutional network classification, graph- and rank-based contextual embeddings for textual and multimedia data, and manifold-aware topic modeling via rank-based prototypes — carry him as last author and show the retrieval toolkit extending toward interpretability, manifolds, and topic models. Collaboration scores 36, with 5 distinct recent co-authors including Lucas Pascotti Valem and Andre Freitas. Four radar axes (output, impact, focus, reproducibility) remain unscored pending additional data.

Research keywords

Highlights

• Context-Aware Interpretable Representations for Retrieval and Graph Convolutional Network Classification

  A 2026 arXiv preprint with Pedronette in the last-author position, pairing context-aware interpretable representations with graph convolutional network classification — a continuation of the contextual rank-analysis line that dominates his topic record, led by Image Retrieval and Classification Techniques (92 works).

• Effective Graph and Rank-based Contextual Embeddings for Textual and Multimedia Data

  A 2026 arXiv preprint, last-authored by Pedronette, extending graph- and rank-based contextual embeddings across textual and multimedia data. It sits squarely in his two largest OpenAlex topic clusters: Advanced Image and Video Retrieval Techniques (88) and Information Retrieval and Search Behavior (61).

• A Manifold-Aware Topic Modeling Approach via Rank-Based Prototypes

  A 2026 arXiv preprint with Pedronette as last author, carrying rank-based prototypes into manifold-aware topic modeling. It signals the researcher's retrieval-originated rank analysis being transferred into text-oriented representation learning, adjacent to his Domain Adaptation and Few-Shot Learning cluster (22 works).

Timeline

1. 2008First recorded works appear in the OpenAlex snapshot (2 publications), opening an unbroken annual output run that continues through 2026.
2. 2012Output reaches 10 works, an early high-water mark in the recorded record.
3. 202112 works recorded, starting a sustained run of 11 or more works per year through 2026.
4. 2025Peak annual output of 19 works, the highest single-year count in the 2008-2026 record.
5. 2026Three arXiv preprints enter the Evolve catalog — context-aware interpretable representations, graph- and rank-based contextual embeddings, and manifold-aware topic modeling — all listing Pedronette as last author.

Papers in the Evolve catalog

Frequent co-authors