Korbinian Riedhammer

Georg Simon Ohm University of Applied Sciences Nuremberg

Unclaimed · AI-structured profileORCID OpenAlex

Speech-technology researcher focused on multimodal ASR corpora and reusing latent speech representations in transcripts

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—

Output is not scored: this snapshot lacks verified publication-volume, venue, and productivity metrics. Once a fuller bibliographic record is available, this axis would measure total research output — publication count, venues, and annual productivity.

Impact—

Impact is not scored: citation counts, h-index, and influence measures are absent from this snapshot. Once available, this axis would measure community uptake — citation volume and the adoption of his corpora and methods.

Focus0/100

Focus scores 0 in this snapshot. The raw field distribution splits evenly — 0.5 for 'speech processing / automatic speech recognition' and 0.5 for 'speech processing and natural language processing' — and with only two reviewed papers, the measure treats this 50/50 split as no concentration. Substantively, both labels describe adjacent speech-technology areas, so the real thematic spread is narrower than the score suggests.

Collaboration32/100

Collaboration scores 32. The reviewed record lists 4 distinct co-authors — Thomas Ranzenberger, Steffen Freisinger, Tobias Bocklet, and Philipp Seeberger — forming a small, stable working group. Korbinian Riedhammer appears in the last (senior, non-corresponding) author position on both 2026 papers, a pattern consistent with supervising or directing the work rather than first-hand execution.

Activity—

Activity is not scored: the snapshot has no per-year publication record to gauge research tempo. This axis would measure output cadence — papers per year, recency of contributions, and continuity of publishing over time.

Reproducibility55/100

Reproducibility scores 55, the mean readiness (reproMean 55) across the 2 reviewed papers, each rated medium. Evolve's verdicts flag concrete gaps: the topic-localization paper reports consistent gains without significance tests, and the corpus paper's strong audio baselines are not yet backed by tested fine-tuning. Together they suggest solid engineering with verification steps still open.

Research summary

Korbinian Riedhammer is a speech-technology researcher at Georg Simon Ohm University of Applied Sciences Nuremberg (ORCID 0000-0003-3582-2154). This full-tier profile is name-resolved from public bibliographic data and rests on a deliberately narrow evidence base: two 2026 arXiv papers reviewed in the Evolve catalog, both listing him in the last (senior, non-corresponding) author position. The first, 'The Spoken Wikipedia Presentation Corpus,' introduces a 553-hour resource pairing Spoken Wikipedia audio with LLM-generated presentation slides for multimodal ASR and contextual-biasing research; Evolve notes strong audio baselines but zero-shot slide context failing, with fine-tuning untested. The second, 'Reusing Latent Speech Representations for Query-Conditioned Topic Localization in Transcripts,' reuses Whisper ASR encoder states for query-conditioned span localization, reporting consistent gains over text-only baselines that are domain-dependent and untested for significance. The field distribution splits evenly between 'speech processing / automatic speech recognition' and 'speech processing and natural language processing,' yielding a focus score of 0 in this snapshot. Collaboration scores 32, reflecting 4 distinct co-authors (Thomas Ranzenberger, Steffen Freisinger, Tobias Bocklet, Philipp Seeberger) in a compact, stable group. Reproducibility scores 55 across the 2 reviewed papers, each rated medium readiness. Output, impact, and activity are not scored because the underlying metrics are missing from this snapshot; career history, awards, and identity verification fall outside its scope.

Collaboration pattern

Within the reviewed record, collaboration is compact and recurrent: 4 distinct co-authors (Thomas Ranzenberger, Steffen Freisinger, Tobias Bocklet, Philipp Seeberger) appear across the two 2026 papers, and Korbinian Riedhammer sits in the last author position on both. The pattern suggests a stable, small research group at Georg Simon Ohm University of Applied Sciences Nuremberg, with senior-author supervision rather than broad external partnerships; the collaboration score is correspondingly moderate at 32.

Research keywords

automatic speech recognitionmultimodal ASRcontextual biasingquery-conditioned span localizationASR encoder representationsWhispertopic segmentationspoken document retrievalspoken language modelscorpus construction

Highlights

Reusing ASR encoder states for query-conditioned topic localization
In a 2026 arXiv paper, the group reuses latent speech representations (Whisper encoder states) for query-conditioned span localization in transcripts, with multimodal fusion delivering consistent gains over text-only baselines; the review notes the gains are domain-dependent and untested for significance.
A 553-hour multimodal ASR corpus built from Spoken Wikipedia
The Spoken Wikipedia Presentation Corpus (2026) pairs Spoken Wikipedia audio with LLM-generated presentation slides (553 h in total) to support multimodal ASR and contextual-biasing research; audio baselines are strong while zero-shot slide context fails, and fine-tuning remains untested.
Consistent senior-author role across the reviewed 2026 record
Both catalogued 2026 papers list Korbinian Riedhammer in the last author position (non-corresponding), working with the same recurring co-authors — a pattern consistent with directing speech-technology projects rather than first-author execution.

Timeline

2026Released 'The Spoken Wikipedia Presentation Corpus' on arXiv as last author — a 553-hour resource pairing Spoken Wikipedia audio with LLM-generated slides for multimodal ASR research.
2026Published 'Reusing Latent Speech Representations for Query-Conditioned Topic Localization in Transcripts' on arXiv, again as last author, reusing Whisper ASR encoder states for query-based span localization.
2026Both 2026 papers entered the Evolve review catalog with medium reproducibility readiness, covering two adjacent speech-processing field labels.

Papers in the Evolve catalog

Frequent co-authors

Thomas Ranzenberger×2 Steffen Freisinger×2 Tobias Bocklet×2 Philipp Seeberger×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.

Radar scores and co-author counts are computed from OpenAlex-derived bibliometrics only; no citation, h-index, or venue metrics are included.
The reviewed scope covers 2 arXiv papers from 2026; earlier publications and career history are outside this snapshot.
The reproducibility score (55) averages exactly 2 papers, so it is sensitive to small-sample effects.
Country information is absent from the source data; only the institution name is provided.