Early-career researcher in responsible speech AI: queer representation audits and participatory dataset curation
2
papers here
2
reviewed by Evolve
2
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-08) 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: the catalog contains only two 2026 arXiv preprints, too small a base to compute an output measure. This axis would track publication volume, venue placement, and progression from preprint to peer-reviewed work.
Impact—
Not yet scored: no citation counts, h-index, or uptake metrics are available for this record. This axis would measure how widely the dataset audit's 0–1.4% finding and the participatory curation framework are cited and adopted.
Focus0/100
Scores 0, and the raw field distribution explains why: the catalogued work splits evenly (0.5/0.5) between two nearly identical labels — "speech technology / responsible AI" and its fairness-accountability-transparency variant — which reads as a labeling artifact rather than genuine scatter. With only two first-authored 2026 preprints, the evidence is too thin to establish a stable topical focus, even though both papers target queer speech-data representation.
Collaboration28/100
Scores 28 on a 0–100 scale. The raw data shows just 3 distinct co-authors across the catalogued record — Anaelia Ovalle, Adina Williams, and Levent Sagun — and both papers are first-authored without corresponding-author roles. This points to a small, concentrated collaboration circle rather than an extended network, a pattern typical of an early-stage researcher whose co-authorship footprint has not yet diversified across groups or venues.
Activity—
Not yet scored: with only two preprints, both dated 2026, there is no multi-year record from which to compute publication cadence. This axis would measure annual output rhythm, sustained engagement, and consistency of research activity over time.
Reproducibility20/100
Scores 20, the mean reproducibility readiness (reproMean 20) across 2 assessed papers. Evolve rates both first-authored preprints "low": the participatory-curation position paper proposes a four-phase framework with no pilot or supporting evidence, while the dataset audit is descriptive, with thin statistics behind its central finding of 0–1.4% queer speakers across 8 datasets. The score reflects how early-stage and conceptual this agenda currently is.
Research summary
Brooklyn Sheppard is a researcher at the University of Calgary working at the intersection of speech technology and responsible AI, focusing on how queer and LGBTQIA+ communities are represented in the datasets that underpin speech systems. The catalogued record is early-stage but coherent: two first-authored 2026 arXiv preprints. The first, "Queer inclusion in speech datasets," audits 8 speech datasets and finds queer speakers make up only 0–1.4% of speakers, then proposes a four-way taxonomy of the practical tensions — around gender annotation and participatory collection — that produce this under-representation. The second, "Towards participatory speech dataset curation," pairs a queer case study with a conceptual four-phase framework for participatory curation, grounded in co-design and data governance. Evolve's review characterizes the audit as descriptive but solid and timely, though statistically thin, and the framework as well-motivated but unvalidated, with no pilot evidence; both carry low reproducibility readiness. The bibliometric footprint is correspondingly small: 3 distinct co-authors (Anaelia Ovalle, Adina Williams, Levent Sagun), a collaboration score of 28, and no citation or impact metrics yet. Overall, Sheppard reads as an emerging voice in participatory, fairness-oriented speech dataset research whose agenda — auditing representation and redesigning curation — is clearly defined, but whose empirical validation, publication venues, and citation record remain to be established.
Collaboration pattern
Collaboration is narrow and consistent: only 3 distinct co-authors appear in the record (collaboration score 28), with Anaelia Ovalle, Adina Williams, and Levent Sagun as the recurring partners on both 2026 first-authored preprints. No corresponding-author role appears yet, and there is no evidence of network breadth across institutions. The pattern suggests a tight working group around queer speech-data research; the circle would be expected to widen as the agenda moves from position papers to empirical studies.
Research keywords
Highlights
Audits queer representation across 8 speech datasets
Proposes a four-phase participatory framework for queer speech data curation
Timeline
- 2026Posts first-authored arXiv audit "Queer inclusion in speech datasets," reporting 0–1.4% queer speakers across 8 datasets and proposing a 4-way tension taxonomy.
- 2026Posts first-authored arXiv position paper "Towards participatory speech dataset curation," introducing a four-phase framework grounded in a queer case study.
- 2026The two 2026 preprints, co-authored with Anaelia Ovalle, Adina Williams, and Levent Sagun, together establish a research line linking speech technology with responsible AI and participatory data governance.
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.
- Bibliometrics and radar scores are derived from OpenAlex-based data only; no citation, impact, or venue-quality metrics are available.
- The catalog holds only two 2026 arXiv preprints, both rated low reproducibility readiness, so every score rests on a very small sample.
- No ORCID or country is linked; "Brooklyn Sheppard" is name-resolved from public bibliographic data and should be verified before citing this profile.
- Missing output, impact, and activity scores reflect absent data, not poor performance.