Zhejiang University ML researcher: wearable-signal time-series foundation models and RL-based LLM post-training.
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
0
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—
Publication-volume data is insufficient: the OpenAlex snapshot records only 2026 activity (4 works) with no prior-year counts, so cumulative output and per-venue breakdown cannot be computed. This axis would measure total papers and annual productivity.
Impact—
Citation-impact data is unavailable: h-index and i10-index are recorded as 0 with no citation counts in the snapshot, so influence cannot be scored. This axis would measure citations, h-index trajectory, and uptake of the two 2026 preprints.
Focus0/100
Focus scores 0/100, consistent with a deliberately split portfolio: the field distribution divides evenly (0.5/0.5) between wearable-health time-series foundation models and LLM post-training, and the five OpenAlex topics — from reinforcement learning in robotics to vital-sign monitoring — each appear once. Recent work shares a methodological core of foundation-model training and RL fine-tuning while spanning applications.
Collaboration70/100
Collaboration scores 70/100, backed by 33 distinct coauthors across recent works and a recurring inner circle — Sirry Chen and Wei Chen each appear twice in the frequent-coauthor list, alongside Zhongyu Wei, Shuai Dong, Yongfu Zhu, Yuqi Xu, Weichu Xie, Liuwenpu, and Ziyue Wang. Authorship roles vary: first author on SOTER, middle author on the distillation study.
Activity47/100
Activity scores 47/100 on thin evidence: recent3yWorks is 4 and OpenAlex countsByYear lists only 2026 with 4 works, suggesting a concentrated recent burst rather than a sustained multi-year cadence. Both catalog papers are 2026 arXiv preprints, so annual output before 2026 cannot be assessed from this snapshot.
Reproducibility78/100
Reproducibility scores 78/100, the mean over 2 reviewed papers. SOTER rates high readiness, with broad and ablated evidence, while the on-policy distillation paper rates medium. Evolve's verdicts flag the remaining gaps — missing significance tests for SOTER and single-seed runs that limit small margins in the RL study.
Research summary
Fangke Chen is a machine learning researcher at Zhejiang University whose recent work spans two distinct but methodologically linked areas: generative time-series foundation models for wearable physiological signals, and reinforcement-learning-based post-training of large language models. The OpenAlex snapshot records 4 works in the recent window, split evenly (0.5/0.5) between these two field directions, with 33 distinct coauthors. In the first track, Chen leads SOTER, a generative time-series foundation model for wearable human physiological signals, as first author; the paper combines mixture-of-experts and neural CDE ideas and claims zero-shot forecasting gains over generalist TSFMs. In the second track, Chen contributes as a middle author to a study of on-policy distillation as a preparation stage for RL, covering GRPO, reverse/forward KL divergence, Pass@k, and RLVR. Evolve has reviewed both papers: SOTER is rated high for reproducibility readiness and the distillation paper medium, yielding a mean reproducibility score of 78 across 2 papers. Output and impact are marked as missing — h-index and i10-index are recorded as 0 and citation data is not yet available. Collaboration scores 70 and activity 47, reflecting a broad co-author network around a recent, concentrated publication record.
Collaboration pattern
Chen's network spans 33 distinct coauthors in the recent window. The frequent-coauthor list shows repeated entries for Sirry Chen and Wei Chen, plus Zhongyu Wei, Shuai Dong, Yongfu Zhu, Yuqi Xu, Weichu Xie, Liuwenpu, and Ziyue Wang; 'Fangke Chen' itself also appears in the list, so identity disambiguation is not fully settled. Roles rotate — first author on SOTER, middle author on the distillation paper — consistent with a score of 70/100.
Research keywords
Highlights
SOTER: first-author generative foundation model for wearable physiological signals
On-policy distillation as a better RL starting point (middle author)
Evenly split dual-track portfolio
The snapshot's field distribution divides 0.5/0.5 between time-series foundation models for wearable health and LLM post-training, with a focus score of 0/100; five OpenAlex topics each appear once, spanning robotics RL to vital-sign monitoring.
Timeline
- 2026OpenAlex's only year with recorded counts: 4 works appear in 2026, split evenly between the wearable-health TSFM and LLM post-training directions.
- 2026First-author preprint SOTER appears on arXiv: a generative time-series foundation model for wearable human physiological signals, rated high for reproducibility readiness.
- 2026Contributes as a middle author to the on-policy distillation study on better RL starts, rated medium for reproducibility readiness.
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.
- All bibliometrics come from a single OpenAlex snapshot; no ORCID is provided, so author disambiguation for this name is not guaranteed.
- h-index and i10-index are recorded as 0 and citation counts are absent, so the output and impact axes are marked missing rather than scored.
- countsByYear covers only 2026 (4 works), so longer-term productivity and publication-venue history cannot be assessed.
- Reproducibility rests on Evolve's review of 2 arXiv preprints; verdicts flag missing significance tests (SOTER) and single-seed runs (distillation paper).