Time-series forecasting researcher at East China Normal University; wind power benchmarks and model adaptation.
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-13) 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—
Not yet scored: publication counts, venues, and productivity over time are absent from the current snapshot. This axis would measure how much and how often this researcher publishes; the catalog holds only two 2026 arXiv preprints.
Impact—
Impact is not yet scored because citation counts, h-index, and downstream uptake are unavailable. This axis would measure how widely the work is cited and reused; with both papers posted as 2026 preprints, citation evidence has not accumulated in the snapshot.
Focus0/100
The focus score is 0, computed from a field distribution split exactly 50/50 between "machine learning (time series forecasting)" and "machine learning — time series forecasting (renewable energy)". In practice the two halves are adjacent — general multivariate forecasting (MUSE) versus wind power applications (WPBench) — so the dispersion partly reflects labeling rather than truly unrelated directions; both cataloged papers sit within time series forecasting.
Collaboration53/100
Collaboration scores 53 with 13 distinct coauthors recorded across the cataloged work. Both papers are multi-author efforts: MUSE is a first-author contribution and WPBench a middle-position one, consistent with team-based benchmark and method building in a lab setting. The moderate score reflects an established but not yet extensive network; recurring names in the frequent-coauthor list include Yuhan Zhu and Xiangfei Qiu.
Activity—
Activity is not yet scored: yearly publication cadence cannot be computed from a snapshot holding only two 2026 arXiv preprints. This axis would measure publishing rhythm and sustained output over multiple years; no earlier record is available here.
Reproducibility78/100
Reproducibility scores 78, the mean readiness rating across both cataloged papers. MUSE is rated high — SOTA claims on 10 datasets with a clearly specified adapter design — while WPBench is rated medium: it unifies 26 datasets and 19 models but ships no visible code and no significance tests. The Evolve verdicts flag missing statistical testing as the main gap on both.
Research summary
Xinying Cai is a researcher at East China Normal University working on machine learning for time series forecasting. The platform's field distribution splits exactly 50/50 between "machine learning (time series forecasting)" and "machine learning — time series forecasting (renewable energy)", reflecting a dual track: general multivariate forecasting methods on one side and wind power applications on the other. The Evolve catalog contains two 2026 arXiv preprints. The first-authored MUSE adapts a frozen masked-autoencoder (MAE) vision backbone to multivariate forecasting through VCR/TPR adapters and reports state-of-the-art results on 10 datasets, earning a high reproducibility rating, though Evolve notes the gains are often thin and lack significance testing. The co-authored WPBench, in a middle author position, unifies 26 wind power datasets and 19 models from 5 families under shape-aware and structure diagnostics, rated medium for reproducibility because no code is visible. Across both papers the mean reproducibility readiness is 78, and 13 distinct coauthors are recorded, yielding a moderate collaboration score of 53. Citation-based measures of output, impact, and activity are not yet available, so this remains an early-stage profile built from a small but methodologically consistent body of work centered on benchmarking and parameter-efficient forecasting models.
Collaboration pattern
Xinying Cai works in team settings: 13 distinct coauthors appear across two multi-author 2026 preprints. MUSE is first-authored; WPBench is a middle-position contribution. The frequent-coauthor list repeats Yuhan Zhu and Xiangfei Qiu (the latter also appears in a concatenated string "Xiangfei Qiu Linsen Li"), suggesting close collaborators on both the method and benchmark lines. The collaboration score of 53 indicates an established but still-growing network.
Research keywords
Highlights
MUSE: first-author adaptation of a frozen vision backbone for forecasting
WPBench: unifying wind power forecasting evaluation
Early record with above-average reproducibility
Timeline
- 2026First-authored MUSE preprint posted to arXiv, adapting a frozen MAE vision backbone to multivariate time series forecasting.
- 2026Co-authored WPBench benchmark preprint released, unifying 26 wind power datasets and 19 forecasting models from 5 families.
- 2026Both papers enter the Evolve catalog with a mean reproducibility readiness of 78, forming an early-stage profile.
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 dimensions are computed from a public bibliographic snapshot; no citation, h-index, or venue-tier data was provided, so output, impact, and activity remain unscored.
- The profile is name-resolved from public bibliographic data; no ORCID was provided, and "Xinying Cai" may match multiple researchers.
- The Evolve catalog holds only two 2026 arXiv preprints; peer-review status and final publication venues are unknown.
- The frequent-coauthor list contains duplicated and concatenated name strings, so the 13-coauthor count may be slightly overestimated.