Meida Chen — University of Central Florida — post-training quantization for robot action models
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-14) 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 publication-count or output-volume metrics were available in this snapshot, so this axis is unscored. It would measure yearly and per-venue publishing volume; only the two 2026 Evolve-cataloged arXiv papers currently exist as evidence.
Impact—
Citation counts and influence metrics are absent from this snapshot, so impact is unscored. It would measure how often the two quantization-evaluation papers are cited and taken up by the robot-learning and efficient-inference communities.
Focus0/100
The radar records a focus score of 0/100. The underlying field distribution splits evenly (0.5 and 0.5) between two labels that differ only in wording — 'robot learning and efficient model deployment (quantization)' and 'robot learning and efficient inference (model quantization)' — so both cataloged papers sit in one narrow topic. The 0 score should therefore be read alongside this single-topic distribution rather than as evidence of scattered interests.
Collaboration44/100
Collaboration scores 44/100, with 8 distinct coauthors recorded. Recurring collaborators include Jiuyi Xu, Jinjia Guo, Jing Du, Yangming Shi, Qing Jin, Song Wang, and Yang Sui. On both reviewed papers Meida Chen appears in a middle author position without corresponding-author role, suggesting contribution within a team rather than sole leadership, consistent with the moderate score.
Activity—
Publication cadence data is insufficient, so activity is unscored. This axis would measure recent publishing rate and consistency over time; the snapshot only shows two arXiv preprints dated 2026, which cannot establish an activity trend.
Reproducibility78/100
Reproducibility is the strongest scored axis at 78/100 (reproMean 78 over 2 scored papers). VLAQuantBench is rated high readiness, consistent with its 409-run, 94,574-episode closed-loop protocol, while Predict Before You Deploy is rated medium, reflecting Evolve's note that its tiny calibration sets limit scope. The gap between the two papers shapes the mean.
Research summary
Meida Chen is a researcher at the University of Central Florida whose Evolve-reviewed work concentrates on making quantized robot-learning models safe to deploy. Both cataloged papers appeared on arXiv in 2026 and address post-training quantization for embodied action models; the profile's field distribution splits evenly between 'robot learning and efficient model deployment (quantization)' and 'robot learning and efficient inference (model quantization)', effectively one tightly scoped agenda. VLAQuantBench provides a closed-loop evaluation of post-training quantization for vision-language-action models; Evolve's review cites 409 runs and 94,574 episodes, concludes that quantization effects are recipe-dependent, and rates its reproducibility readiness high. Predict Before You Deploy proposes offline prediction of quantization-induced task degradation for world action models using calibrated action-deviation thresholds, with selective-prediction and deferral semantics; Evolve finds it rigorous but limited by tiny calibration sets, with medium readiness. Together the two papers suggest a deliberate pattern: benchmark first, then predict and defer before deployment. Radar scores give reproducibility 78/100 across two scored papers and collaboration 44/100 with 8 distinct coauthors, while output, impact, and activity lack sufficient data to score. This is a name-resolved, early-stage profile built solely from public bibliographic data and Evolve's own reviews.
Collaboration pattern
Evolve records 8 distinct coauthors for Meida Chen, with a collaboration score of 44/100. The frequent-coauthor list includes Jiuyi Xu, Jinjia Guo, Jing Du, Yangming Shi, Qing Jin, Song Wang, and Yang Sui, suggesting stable small-team work around quantization for robot learning. On both reviewed 2026 papers Chen is a middle author and not the corresponding author, so the catalog evidence points to contributor roles within a group rather than solo-led projects.
Research keywords
Highlights
VLAQuantBench: large-scale closed-loop audit of quantization for vision-language-action models
Predicting quantization-induced degradation before deployment
Timeline
- 2026Posts VLAQuantBench to arXiv: a closed-loop evaluation of post-training quantization for vision-language-action models, backed by 409 runs and 94,574 episodes.
- 2026Posts Predict Before You Deploy to arXiv, proposing offline prediction of quantization-induced task degradation for world action models via calibrated action-deviation thresholds.
- 2026Both papers enter the Evolve catalog with reproducibility readiness of high and medium respectively, establishing a benchmark-then-predict research line on quantized robot policies.
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.
- Radar scores are computed from public bibliometric data; no citation totals, h-index, or publication counts were included in this snapshot.
- Only the two Evolve-reviewed 2026 arXiv preprints inform this profile; peer-reviewed venue status and the fuller publication record are not captured.
- Country metadata is absent (institution recorded as University of Central Florida only); this is a name-resolved profile from public bibliographic data tied to ORCID 0000-0002-2436-3848.
- The focus score of 0 sits oddly against a single-topic field distribution (0.5/0.5 across two near-identical labels), so read that axis with caution.