Sören Auer

Leibniz University Hannover · Germany

Unclaimed · AI-structured profileORCID OpenAlex

Sören Auer — Semantic Web and knowledge engineering researcher at Leibniz University Hannover (OpenAlex h-index 52)

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

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—

The output axis — annual publication volume and productivity profile — is not scored here: this OpenAlex snapshot supplies totals by year and topic but the derived output dimension has not been computed, so no estimate is offered.

Impact—

The impact axis measures citation-based influence such as field-normalized citation rates. Although the snapshot lists an OpenAlex h-index of 52 and i10-index of 234, per-work citation data needed to score impact is absent, so the axis stays unscored.

Focus—

The focus axis would quantify topical concentration versus dispersion across the corpus. It is unscored here: topic labels exist in the snapshot, but the data required to compute this dimension has not yet been assembled, so no value is estimated.

Collaboration32/100

Scoring 32/100 for collaboration, this snapshot records only 4 distinct co-authors against 147 works over the last three years — an unusually tight footprint pointing to a small recurring core: Hamed Babaei Giglou, Jennifer D'Souza, Peio Popov and Mahsa Sanaei. On the three reviewed 2026 papers Auer sits last or middle author, consistent with a senior supervisory role rather than hands-on first-author work.

Activity100/100

Activity scores a full 100/100. The countsByYear series shows sustained output since 2004 and a steep recent climb: 30 works in 2022, 27 in 2023, 37 in 2024, 49 in 2025 and 61 in 2026 — the highest on record. Recent-3-year productivity stands at 147 works, confirming an exceptionally active, accelerating publishing pace.

Reproducibility—

The reproducibility axis would measure artifact, code and data availability across the corpus. It is unscored: only the three Evolve-reviewed 2026 papers carry readiness labels — each rated medium — far too narrow a base for a profile-level estimate.

Research summary

Sören Auer, based at Leibniz University Hannover (DE), is a Semantic Web and knowledge-engineering researcher whose OpenAlex snapshot shows an h-index of 52, an i10-index of 234, and 147 works over the last three years. The countsByYear series documents two decades of continuous publishing since 2004 and a steep recent acceleration — 37 works in 2024, 49 in 2025 and 61 in 2026, the highest annual count on record — yielding a perfect 100/100 activity score. Topic mass concentrates on Semantic Web and Ontologies (496), followed by Library Science and Information Systems (195), Topic Modeling (165), Biomedical Text Mining and Ontologies (70) and Research Data Management Practices (68). The three Evolve-reviewed 2026 arXiv papers show a researcher steering his agenda into the LLM era: benchmarking whether general sentence embeddings capture ontological reasoning (last author), fusing string-, KGE- and LLM-based ontology aligners by voting, and running a controlled 13-model study of when LLM scale helps ontology learning. Across them, Evolve's verdicts reward careful, honest benchmarking while flagging limits — synthetic labels without human validation, a single fusion strategy evaluated, tiny test sets without significance tests. A 32/100 collaboration score reflects a snapshot recording just 4 distinct co-authors against those 147 works, centered on Hamed Babaei Giglou, Jennifer D'Souza, Peio Popov and Mahsa Sanaei. Output, impact, focus and reproducibility dimensions remain unscored for lack of underlying data.

Collaboration pattern

The snapshot's collaboration footprint is narrow and stable: 4 distinct co-authors recorded across 147 recent-3-year works, with Hamed Babaei Giglou, Jennifer D'Souza, Peio Popov and Mahsa Sanaei recurring on the reviewed 2026 papers. Auer's last- and middle-author placements fit a senior supervisor pattern — defining problems and benchmarks while collaborators run experiments. The low distinct-coauthor count likely reflects snapshot coverage limits rather than the full network (collaboration score: 32/100).

Research keywords

Semantic WebOntologiesKnowledge GraphsOntology LearningOntology AlignmentLarge Language ModelsSentence EmbeddingsKnowledge Graph EmbeddingsRetrieval-Augmented GenerationBenchmarkingResearch Data ManagementSemantic Textual Similarity

Highlights

AVA benchmark: general NLP embeddings fail ontological reasoning
Evolve's review of this arXiv study (Auer as last author) reports that 25+ sentence embeddings cannot discriminate at ontology level, that fine-tuning nearly solves the task yet transfers poorly across domains, and that synthetic labels lack human validation — reproducibility rated medium.
OntoAligner-Ensemble fuses string, KGE and LLM aligners by voting
Reviewed on arXiv (2026), this work combines heterogeneous ontology alignment techniques — string-based, knowledge-graph-embedding and LLM-based — through voting-based fusion. Evolve credits solid engineering and honest results, but notes only one fusion strategy is actually evaluated; reproducibility medium.
Controlled study: LLM scale lifts ontology learning precision, non-monotonically
This 13-model controlled study asks when bigger LLMs help ontology learning. Evolve's verdict: scale mainly raises precision and does so non-monotonically; the design is sound but tiny test sets and the absence of significance tests limit the claims. Reproducibility medium.

Timeline

2026Last-authored an arXiv benchmark study (AVA) showing 25+ general NLP embeddings fail ontology-level discrimination, while fine-tuning nearly closes the gap yet transfers poorly.
2026Co-authored OntoAligner-Ensemble, a voting-based fusion of string-based, knowledge-graph-embedding and LLM ontology aligners — reviewed by Evolve as solid engineering with only one fusion strategy evaluated.
2026Middle-authored a controlled 13-model study of LLM scale in ontology learning, finding that scale mainly raises precision and does so non-monotonically.

Papers in the Evolve catalog

Frequent co-authors

Hamed Babaei Giglou×3 Jennifer D'Souza×3 Peio Popov×1 Mahsa Sanaei×1

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 (h-index 52, i10-index 234, countsByYear, topics, distinct co-authors) come from an OpenAlex-derived snapshot; per-work citation data is absent, so the output, impact, focus and reproducibility axes remain unscored.
Qualitative verdicts rest on only three Evolve-reviewed 2026 arXiv papers — a very recent, narrow sample of this researcher's much larger output.
Only 4 distinct co-authors are recorded against 147 recent-3-year works, which almost certainly under-counts the true collaboration network.
Identity is name-resolved from public bibliographic data (OpenAlex); affiliation and country are as provided, and no biographical details were independently verified.