Kiki van der Heijden
Donders Institute for Brain, Cognition and Behaviour · Netherlands
Spatial-audio ML researcher: equivariant SELD networks and self-supervised soundfield embeddings (Donders Institute)
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-09) 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—
No output score yet: the OpenAlex record shows h-index 0 and i10-index 0 with no usable publication aggregates, so cumulative output cannot be assessed. This axis would measure total and per-year publication volume across venues and preprint servers.
Impact—
No impact score yet: with h-index 0, i10-index 0 and no citation aggregates in the OpenAlex record, citation influence cannot be measured. This axis would track citations, citing works and their distribution over time once the record accumulates them.
Focus0/100
Scores 0/100 — the lowest concentration in the snapshot. The 4 indexed works spread across four OpenAlex topics: Speech Recognition and Synthesis (4), Speech and Audio Processing (4), Music and Audio Processing (3), and Embedded Systems Design Techniques (1), while the Evolve field split is even (0.5 spatial-audio/audio signal processing, 0.5 SELD). With so few works, any topical split reads as dispersion.
Collaboration22/100
Scores 22/100. The OpenAlex record lists just 2 distinct co-authors, with Goksenin Yuksel and Marcel van Gerven recurring across the Evolve-reviewed work; both 2026 catalog papers sit in this small circle, with van der Heijden in the last author position. The low score reflects a compact, stable team rather than an extended network, and OpenAlex identity resolution may also undercount links.
Activity47/100
Scores 47/100. The record counts 4 works in the last three years (recent3yWorks 4), and countsByYear places all of them in 2026 — including two arXiv preprints reviewed by Evolve that year. The pattern is a concentrated recent burst rather than a sustained multi-year cadence, consistent with a researcher whose visible publishing record is just getting under way.
Reproducibility100/100
Scores 100/100, computed from a reproducibility mean of 100 across 2 reviewed papers (reproMean 100, reproPapers 2). Both Evolve-reviewed papers carry a high reproducibility rating: EquiSELD reports an explicit O(3)-versus-SO(3) ablation plus comparisons on TAU2021 and STARSS23, and Bearings quantifies ablated SELD gains (F20° error from below 4 to 50.4 on TAU). Strong, but the evidence base is only two papers.
Research summary
Kiki van der Heijden is a researcher at the Donders Institute for Brain, Cognition and Behaviour (Netherlands) working on machine learning for spatial audio. This profile is name-resolved from public OpenAlex bibliographic data. The record is early-stage: h-index and i10-index are both 0, and OpenAlex counts 4 works in the last three years, all placed in 2026. Two of these are arXiv preprints reviewed in the Evolve catalog, on both of which van der Heijden is listed in the last author position (not corresponding author). EquiSELD (2026) trains exactly O(3)-equivariant attention networks for sound event localization and detection, winning on TAU2021, though the O(3)-versus-SO(3) ablation shows no clear gain and STARSS23 wins fall within noise. Bearings (2026) introduces self-supervised soundfield embeddings from first-order Ambisonics, claimed as the first pluggable into frozen encoders, with solid ablated SELD gains (F20° error from below 4 to 50.4 on TAU) and a moderate lead over hand-crafted statistics. OpenAlex topics cluster around speech recognition and synthesis, speech and audio processing, and music and audio processing. Collaboration is compact: 2 distinct co-authors, with Goksenin Yuksel and Marcel van Gerven recurring. Both reviewed papers rate high for reproducibility (reproMean 100 over 2 papers).
Collaboration pattern
The collaboration footprint is currently small: OpenAlex lists 2 distinct co-authors, and Goksenin Yuksel and Marcel van Gerven recur across the reviewed work. Both 2026 papers place van der Heijden in the last author position without a corresponding-author flag, a pattern consistent with a senior role in a compact spatial-audio ML line at the Donders Institute. The low collaboration score (22/100) likely reflects the small team, possibly compounded by OpenAlex undercounting.
Research keywords
Highlights
Exactly O(3)-equivariant SELD network takes the lead on TAU2021
Self-supervised soundfield embeddings that plug into frozen encoders
A compact research line on first-order Ambisonics for SELD
Both 2026 catalog papers share one thread — first-order Ambisonics representations for sound event localization and detection — with van der Heijden in the last author position and the same small co-author circle (Goksenin Yuksel, Marcel van Gerven).
Timeline
- 2026EquiSELD preprint appears on arXiv: an exactly O(3)-equivariant SELD attention network that wins on TAU2021, with an O(3)-versus-SO(3) ablation showing no clear gain.
- 2026Bearings preprint appears on arXiv: self-supervised soundfield embeddings from first-order Ambisonics, claimed as the first pluggable into frozen encoders.
- 2026OpenAlex places the entire recent-3-year output — 4 works — in this single year, with both Evolve-reviewed papers rated high for reproducibility.
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.
- Bibliometric statistics come from OpenAlex only; the profile is name-resolved from public bibliographic data, so conflation with namesakes cannot be fully ruled out.
- h-index and i10-index are 0 and no citation aggregates are available, so the output and impact radar axes are left unscored.
- OpenAlex counts 4 works in 2026, but only 2 have Evolve reviews; the remaining works are not described in the provided data.
- The reproducibility score of 100 rests on a small base: only 2 reviewed papers (reproMean 100, reproPapers 2).