Speech-to-speech researcher building full-duplex voice dialogue models with native tool calling for voice agents
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
2
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 metrics are not yet available: the underlying record lacks reliable per-year publication and venue counts for this author, so this axis would measure total works, venue quality, and first or corresponding-author share once sufficient data exists.
Impact—
Impact data is insufficient: citation counts, citation distribution, and h-index trend over time are not available for scoring. This axis would measure citation-based influence, including how often the reviewed speech-to-speech and tool-calling works are cited.
Focus0/100
The focus score is 0, the lowest reading on this axis: the field distribution splits evenly (0.5/0.5) between speech processing and conversational AI and speech processing / spoken dialogue systems (cs.CL). The topic profile is nonetheless narrow — Speech Recognition and Synthesis (5), Speech and Audio Processing (5), Speech and dialogue systems (4) — with one outlier record on Domain Adaptation and Few-Shot Learning (1).
Collaboration83/100
Collaboration is the strongest axis at 83/100: the record lists 64 distinct coauthors, and the repeat-collaborator cluster — including Edresson Casanova, Zhehuai Chen, Jagadeesh Balam, Elena Rastorgueva, Nikhil Srihari, and Travis Bartley — points to sustained work inside a large speech team. Both reviewed 2026 papers are multi-author efforts with Deng in a middle, non-corresponding position.
Activity40/100
Activity scores 40/100: the record shows a sparse cadence — one work in 2020, one in 2021, then one in 2025 and two in 2026 — for 3 works in the last three years. Recent momentum is concentrated in the 2026 arXiv releases on full-duplex speech-to-speech models, following no indexed works in 2022-2024.
Reproducibility38/100
Reproducibility averages 38/100 across the 2 papers Evolve has reviewed. NemotronLabs VoiceChat is rated medium readiness, while the frontend-backend token-delegation architecture is rated low, the review noting strong benchmark results but no significance testing. Untested design claims in the VoiceChat review further limit confidence in reproducing reported behavior.
Research summary
Slyne Deng is a speech processing researcher whose sparse public bibliographic record — resolved from OpenAlex — suggests an early career stage and centers on full-duplex speech-to-speech conversation and tool calling for voice agents. The record shows five works appearing in 2020, 2021, 2025, and 2026 (three in the last three years), with topic mass concentrated in Speech Recognition and Synthesis (5), Speech and Audio Processing (5), and Speech and dialogue systems (4), plus one record on Domain Adaptation and Few-Shot Learning. Field distribution splits evenly between speech processing and conversational AI and spoken dialogue systems (cs.CL). Evolve has reviewed two 2026 arXiv papers: NemotronLabs VoiceChat, an open full-duplex speech-to-speech model with native tool calling that shows strong turn-taking but weak argument accuracy (42.2%; Pass@1 33.0%), and a frontend-backend token-based delegation architecture for adding tool calls to duplex speech frontends. In both, Deng is a middle-position, non-corresponding author, consistent with a contributor role on larger team efforts. Bibliometric impact is modest so far (h-index 2, i10-index 1), and collaboration is the strongest radar signal (83/100) with 64 distinct coauthors. Reproducibility averages 38/100 across the two reviewed papers, with one rated medium and one low; output and impact axes currently lack sufficient data to score.
Collaboration pattern
Slyne Deng works within large, stable speech teams: 64 distinct coauthors appear in the record, and a repeat cluster — Edresson Casanova, Zhehuai Chen, Jagadeesh Balam, Elena Rastorgueva, Nikhil Srihari, Travis Bartley, Viet Anh Trinh, Oluwatobi Olabiyi, Sanjay Chauhan, and Chen Chen — recurs across efforts. In both reviewed 2026 papers Deng is a middle-position, non-corresponding author, consistent with a specialist contributor rather than a lone-PI model.
Research keywords
Highlights
First open full-duplex speech-to-speech model with native tool calling
Turn-taking and streaming stack stand out in review
Token-based delegation architecture for duplex speech frontends
Timeline
- 2020First indexed work appears in the OpenAlex record (1 work).
- 2021One indexed work; the record then shows no indexed works through 2024.
- 2025One indexed work marks a return to publishing.
- 2026Two arXiv papers on full-duplex speech-to-speech tool calling enter the Evolve catalog, both with Deng as a middle-position, non-corresponding author.
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 metrics (h-index 2, i10-index 1, counts by year, coauthor counts) come from an OpenAlex identity record; institution, country, and ORCID are unverified.
- "Slyne Deng" is name-resolved from public bibliographic data; identity boundaries may shift if same-name authors are merged or split.
- Output and impact axes are unscored because the underlying citation and venue data is insufficient.
- Evolve has reviewed only 2 of this author's papers; quality judgments (including reproducibility 38/100) reflect those papers, not the full record.