NLP and spoken-dialogue researcher at KIT bridging full-duplex speech interaction and machine translation evaluation
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
3
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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 scoring is not yet available: indexed volume under this name is too sparse for a reliable score. The axis would measure annual and cumulative publication counts; raw data already show counts rising from 1 (2021) to 30 (2026).
Impact—
Impact scoring is not yet available: citation accumulation is at an early stage (h-index 3, i10-index 1), too thin to score. This axis would measure citation influence — total citations, h-index growth, and uptake of released benchmarks over time.
Focus0/100
The focus score is 0, reflecting a deliberately two-sided record: OpenAlex distributes the work evenly (0.5 machine translation evaluation, 0.5 full-duplex spoken dialogue systems). Topic counts reinforce the spread — Speech Recognition and Synthesis (36), Topic Modeling (30), Natural Language Processing Techniques (26), Speech and dialogue systems (22), Speech and Audio Processing (16). This is breadth across speech and language technology rather than concentration in one niche.
Collaboration83/100
Collaboration scores 83, grounded in 64 distinct coauthors over the last three years and 12 recurring partners, including Vilém Zouhar, Niyati Bafna, Mukund Choudhary, Sara Rajaee, Pinzhen Chen, Jannis Vamvas, and Sara Papi. Both catalog papers are multi-author community efforts — one a first-author systems paper, the other a widely staffed benchmark — indicating work embedded in team settings.
Activity100/100
Activity reaches 100, the strongest signal in this profile: indexed works per year climb from 1 (2021) and 3 (2023, 2024) to 25 (2025) and 30 (2026), with 58 works in the last three years. Both Evolve-cataloged arXiv papers appear in 2026, showing the acceleration is ongoing.
Reproducibility55/100
Reproducibility scores 55, computed from 2 reviewed papers with a mean of 55, both rated medium readiness. Strengths include probing setups, released benchmarks, and rule-based verifiers; the stated limits are that the human study covers only the 7B model, and the translation benchmark relies on unquantized LLM verifiers.
Research summary
Maike Züfle is a researcher at the Karlsruhe Institute of Technology (KIT) whose record straddles two halves of language technology: full-duplex spoken dialogue systems and machine translation evaluation. OpenAlex field distribution attributes the work equally (0.5/0.5) to these two areas, and topic frequencies reinforce the spread: Speech Recognition and Synthesis (36), Topic Modeling (30), Natural Language Processing Techniques (26), Speech and dialogue systems (22), and Speech and Audio Processing (16). The Evolve catalog includes two 2026 arXiv papers. As first author of "Controlling Backchannels in Streamable Full-duplex Models", Züfle proposes a tunable backchannel head for full-duplex models, supported by probing, benchmark, and human evaluations. She also co-authored the community-built "Last Translation Benchmark", which pairs 3,456 difficult exceptions with rule-based verifiers to separate frontier models from human rankings. Her output curve is steep: indexed works rise from 1 (2021) to 25 (2025) and 30 (2026), with 58 works in the last three years and 64 distinct coauthors, yielding an activity score of 100 and a collaboration score of 83. Citation indicators remain early-stage (h-index 3, i10-index 1), so the output and impact axes are not yet scored.
Collaboration pattern
Züfle works in dense team settings: 64 distinct coauthors in three years and a collaboration score of 83. Recurring partners include Vilém Zouhar, Niyati Bafna, Mukund Choudhary, Sara Rajaee, Pinzhen Chen, Jannis Vamvas, Sara Papi, Ona de Gibert, Bhavitvya Malik, Eliya Habba, Orfeas Menis Mastromichalakis, and Patrícia Schmidtová. The two cataloged papers sit at opposite ends of one pattern — a first-author full-duplex systems paper and a large community benchmark — suggesting both leadership and contributing roles.
Research keywords
Highlights
Tunable backchannel control for streamable full-duplex dialogue models
Community-built Last Translation Benchmark with rule-based verification
A profile spanning speech interaction and translation evaluation
Field distribution splits evenly between machine translation evaluation and full-duplex spoken dialogue, with 58 works in three years and 64 distinct coauthors — an early-stage record accelerating across both speech and language technology.
Timeline
- 2021First indexed work appears in OpenAlex (1 work); the record starts thin and citation indicators remain early-stage (h-index 3).
- 2023Indexed output reaches 3 works, a level held again in 2024.
- 2025Output accelerates sharply to 25 indexed works.
- 2026Peak activity: 30 indexed works, plus two arXiv papers in the Evolve catalog — first-author backchannel control for full-duplex models and the co-authored Last Translation Benchmark.
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 are derived from OpenAlex only; identity is resolved from public bibliographic data without an ORCID, so same-name works may be conflated.
- Only two papers are in the Evolve catalog, both 2026 arXiv preprints; peer-review status is not reflected.
- Output and impact dimensions are unscored because citation data are insufficient (h-index 3, i10-index 1).
- Reproducibility scores rest on only two reviewed papers (mean 55), a small base.