Luis F. Garcia

—

Unclaimed · AI-structured profileOpenAlex

KGQA researcher: traceable multi-hop graph navigation and path-fidelity evaluation beyond answer accuracy

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-15) 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—

Not yet scored: output data is too sparse to evaluate — the OpenAlex record holds only two works (both 2026) with h-index and i10-index at 0. This axis would measure total publication volume, venue spread, and productivity over time.

Impact—

Not yet scored: citation-impact data is unavailable — h-index and i10-index are both 0, with no citation counts recorded beyond the two 2026 works. This axis would measure how often the work is cited and its uptake by later research.

Focus0/100

Evolve scores focus at 0, the lowest reading on this radar. The raw field distribution is evenly split — 0.5 for knowledge-graph NLP and 0.5 for KGQA/LLM reasoning — and OpenAlex assigns three topics (Topic Modeling, Semantic Web and Ontologies, Advanced Graph Neural Networks) at two works each. With only two papers indexed, concentration is genuinely undetermined rather than scattered.

Collaboration32/100

Collaboration scores 32, a modest level consistent with the raw counts: 4 distinct coauthors (Eduin E. Hernandez, Nurassyl Askar, Sergio A. Diaz, Stefano Rini) across two 2026 papers, on which Garcia sits in a middle author position without corresponding credit. The pattern suggests a contributing member of a small, stable team rather than an independent group lead so far.

Activity32/100

Activity scores 32, matching a thin recent record: OpenAlex counts 2 works in the last three years, both appearing in 2026 (counts by year list 2026 with 2 works and nothing earlier). The two arXiv preprints were reviewed by Evolve in the same cycle, so current output exists, but a multi-year cadence cannot yet be assessed.

Reproducibility78/100

Reproducibility is the strongest axis at 78, computed as a mean of 78 over 2 reviewed papers. Evolve rates 'Theseus in the Graph' high for reproducibility readiness (gold-path supervised benchmark with fidelity metrics), while 'The Path Matters' is medium (sound design, but small benchmarks and no significance tests). The above-average score reflects unusually explicit evaluation artifacts for an early record.

Research summary

Luis F. Garcia is an early-career researcher working at the intersection of natural language processing and knowledge graphs, with a profile name-resolved from public OpenAlex bibliographic data (no verified ORCID or affiliation is on record). The indexed record is very recent: OpenAlex lists exactly two works in 2026, both arXiv preprints reviewed in the Evolve catalog, in which Garcia appears in a middle author position without corresponding-author credit. Both papers converge on one problem: making multi-hop reasoning over knowledge graphs traceable rather than answer-only. 'Theseus in the Graph' (2026) frames KGQA as question-conditioned graph navigation supervised with gold paths and evaluated with fidelity metrics, and carries a high reproducibility-readiness rating, though Evolve notes that the closest prior baselines were excluded. 'The Path Matters' (2026) extends this to evaluation methodology, showing that navigation fidelity and answer accuracy can rank small language models differently, and introduces path-based measures such as Path Edit Distance; its reproducibility readiness is rated medium, with small benchmarks and no significance tests flagged. OpenAlex topic assignments (Topic Modeling, Semantic Web and Ontologies, Advanced Graph Neural Networks, two works each) and a 50/50 field split between knowledge graphs and KGQA/LLM reasoning describe a coherent but narrow early record. Citation metrics (h-index 0, i10-index 0) plus modest collaboration (4 distinct coauthors) and activity signals (score 32) mark this as a beginning profile.

Collaboration pattern

Collaboration is modest (radar score 32): Luis F. Garcia's two 2026 arXiv papers share one small, stable team, with 4 distinct coauthors recorded in OpenAlex — Eduin E. Hernandez, Nurassyl Askar, Sergio A. Diaz, and Stefano Rini. On both papers Garcia sits in a middle author position without corresponding-author credit, consistent with a contributing role inside a compact group rather than leading an independent lab. No cross-institutional or solo-author patterns can be inferred from this thin record.

Research keywords

knowledge graph question answeringmulti-hop reasoninggraph navigationpath fidelityPath Edit Distancetraceabilitysmall language modelsreinforcement learning agentssemantic web and ontologiestopic modelinggraph neural networksreproducibility

Highlights

THESEUS: KGQA as question-conditioned graph navigation with gold paths
Evolve's review credits this 2026 paper with framing knowledge graph question answering as traceable, question-conditioned graph navigation, supervised with gold paths and scored with fidelity metrics; the benchmark is judged solid, though the closest prior baselines were excluded from comparison. Reproducibility readiness: high.
Path fidelity can rerank small language models in KGQA
In the companion 2026 study, Evolve notes a controlled finding: small language models can rank differently when navigation-path fidelity is measured alongside answer accuracy, using path-level measures such as Path Edit Distance. The design is judged sound, but benchmarks are small and no significance tests are reported; reproducibility readiness: medium.
A shared fidelity-first evaluation methodology across both papers
Both Evolve-reviewed 2026 papers converge on evaluating multi-hop reasoning beyond answer accuracy — gold paths, path-fidelity metrics, traceability — which also underpins the profile's above-average reproducibility mean of 78 across the two works. The same small team (4 distinct coauthors) appears on both papers.

Timeline

2026Posted 'Theseus in the Graph: Towards Traceable Multi-Hop Graph Navigation' to arXiv, introducing question-conditioned graph navigation with gold paths and fidelity metrics (Evolve: solid benchmark, though closest prior baselines excluded).
2026Posted 'The Path Matters: Evaluating Small Language Models Beyond Answer Accuracy in KGQA' to arXiv, showing that navigation fidelity and answer accuracy can rank small language models differently and introducing Path Edit Distance.
2026OpenAlex year counts for this profile register both works under 2026 (recent-3-year works: 2), making 2026 the entire indexed publication history so far.

Papers in the Evolve catalog

Frequent co-authors

Eduin E. Hernandez×2 Nurassyl Askar×2 Sergio A. Diaz×2 Stefano Rini×2

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.

Bibliometrics come from OpenAlex only (identity: openalex); h-index and i10-index are 0, so the output and impact radar axes are unscored rather than confirmed zero-value achievements.
'Luis F. Garcia' is a fairly common name; this profile is name-resolved from public bibliographic data, with no verified ORCID, institution, or country on record.
The catalog holds only two Evolve-reviewed 2026 arXiv preprints; peer-review status, venue, and any earlier publications are outside this snapshot.
Scores rest on a tiny sample: reproducibility (mean 78) is averaged over 2 papers, and the focus score of 0 reflects an even 50/50 field split on a very early record.