Tianlun Zuo

Northwestern Polytechnical University · China

Unclaimed · AI-structured profileOpenAlex

NWPU speech researcher: multilingual TTS robustness, diffusion-based TTS, and clinical audio-language modeling.

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

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—

No output score yet: the snapshot lacks a complete, verified publication list beyond 9 works in the recent three-year window and yearly counts of 2 (2025) and 7 (2026). This axis would measure total publication volume across venues once the full record is assembled.

Impact—

No impact score yet: citation data are too thin to score on this early record (h-index 1, i10-index 0 are the only available signals). This axis would measure citation-based influence such as citations per paper and field-normalized impact.

Focus0/100

Scores 0/100 on concentration: the field distribution splits evenly at 0.33 across three clusters — speech processing/NLP for text-to-speech, clinical NLP (audio-language modeling), and machine-learning speech synthesis. Topic counts (Speech Recognition and Synthesis 9, Speech and Audio Processing 7, Speech and dialogue systems 6) reinforce a broad TTS-centered spread rather than a single narrow specialty at this career stage.

Collaboration51/100

Scores 51/100. The recent record lists 12 distinct coauthors across 9 works in the last three years, with recurring collaborators Ziyu Zhang and Lei Xie. Two of the three cataloged papers list Zuo as a middle author and none as corresponding author, indicating active participation within a stable speech team rather than independent leadership so far.

Activity67/100

Scores 67/100. Output is accelerating: yearly counts rise from 2 works in 2025 to 7 in 2026, and the recent three-year window holds 9 works. All three Evolve-reviewed papers appeared as 2026 arXiv preprints, consistent with a high submission-tempo early-career profile, though the observed window is short and peer-reviewed acceptance is not yet documented.

Reproducibility32/100

Scores 32/100 (reproMean over 3 reviewed papers). Evolve rates reproducibility readiness low for the clinical SOAP and DiTAR+ papers and medium for the first-author multilingual TTS benchmark, which at least provides a defined 960-text evaluation set; recurring review flags include untuned baselines and missing significance testing, so independent replication is not yet demonstrated.

Research summary

Tianlun Zuo is an early-career speech researcher at Northwestern Polytechnical University (China), name-resolved from OpenAlex public bibliographic data, whose work centers on text-to-speech synthesis and audio-language modeling. The field distribution splits evenly at about a third each across speech processing/NLP for text-to-speech, clinical NLP (audio-language modeling), and machine-learning-based speech synthesis; topics are led by Speech Recognition and Synthesis (9), Speech and Audio Processing (7), and Speech and dialogue systems (6). All three Evolve-cataloged papers are 2026 arXiv preprints. As first author, Zuo led a 960-text robustness benchmark with a rule-based risk score for low-resource multilingual TTS; as a middle author, Zuo contributed to DiTAR+ (robust autoregressive diffusion TTS) and an end-to-end clinical audio-to-SOAP system reporting 0.52 vs 0.29 Concept F1 over cascaded ASR+LLM baselines. Productivity is accelerating: 9 works in the recent three-year window, 7 indexed in 2026 versus 2 in 2025, with 12 distinct coauthors. Citation signals remain minimal (h-index 1, i10-index 0), and Evolve's reproducibility readiness averages 32/100 across the three papers, leaving tuned baselines and independent replication as open questions.

Collaboration pattern

Tianlun Zuo works within a stable speech team at Northwestern Polytechnical University: 12 distinct coauthors appear across 9 recent-three-year works (collaboration score 51/100). Ziyu Zhang and Lei Xie recur most often, alongside Tingzhi Mao, Zhonghua Fu, Mingchen Shao, Wenjie Tian, Longhao Li, Hanzhao Li, and Haoyu Zhang. Authorship placement — first author on the benchmark paper, middle author on the two modeling papers, corresponding on none — suggests a contributing team member rather than an independent group lead so far.

Research keywords

zero-shot TTSautoregressive diffusion transformermultilingual TTSrobustness evaluationtext normalizationcode-switchinglow-resource languagesclinical SOAP note generationend-to-end speech understandinglong-form audioaudio-language modelsinstruction tuning

Highlights

First-authored robustness benchmark for low-resource multilingual TTS
Leads a 2026 arXiv study providing a 960-text robustness benchmark and a rule-based risk score for diagnosing complex-text failures in low-resource multilingual text-to-speech; Evolve finds the diagnosis useful but notes the benchmark is small, untested, and Whisper-dependent. Reproducibility readiness: medium.
DiTAR+: dual optimization for robust autoregressive diffusion TTS
Co-authors a 2026 zero-shot TTS model combining autoregressive diffusion transformers with dilated context expansion and acoustic attention masking over continuous latent speech representations; Evolve judges the gains plausible but notes its own ablations show single modules approach the full combination. Reproducibility readiness: low.
End-to-end clinical audio to SOAP note generation
Middle-author contribution to a 2026 study bridging long-form clinical audio and audio-language models for SOAP note generation, reporting 0.52 vs 0.29 Concept F1 over cascaded ASR+LLM baselines; Evolve calls the claim plausible but flags untuned baselines, no significance tests, and an unevaluated OPD stage. Reproducibility readiness: low.

Timeline

2025Earliest year in the snapshot: OpenAlex yearly counts record 2 indexed works.
2026Output accelerates to 7 indexed works, the most productive year in the snapshot.
2026First-author robustness benchmark for low-resource multilingual TTS released on arXiv.
2026Co-authored DiTAR+ diffusion TTS and end-to-end clinical SOAP generation papers appear on arXiv.

Papers in the Evolve catalog

Frequent co-authors

Lei Xie×2 Ziyu Zhang×1 Tingzhi Mao×1 Zhonghua Fu×1 Lei Xie×1 Ziyu Zhang×1 Mingchen Shao×1 Wenjie Tian×1 Longhao Li×1 Ziyu Zhang×1 Hanzhao Li×1 Haoyu Zhang×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.

Bibliometric indicators come from OpenAlex only; no ORCID is linked, so this profile is name-resolved from public bibliographic data.
The output and impact radar axes lack sufficient data; the only citation signals available are h-index 1 and i10-index 0 on a very early record.
The catalog covers three 2026 arXiv preprints reviewed by Evolve; peer-reviewed venues and the full publication list are not represented.
The reproducibility score (32/100) reflects Evolve's readiness assessment of the 3 reviewed papers, not independent replication.