Applies deep reinforcement learning and hybrid deep learning to industrial control and predictive maintenance.
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
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—
Output cannot be scored from this snapshot: no cumulative publication count is available — only the two 2026 works, with h-index and i10-index at 0. This axis would measure the volume and consistency of publications over time, which needs a longer record.
Impact—
Impact cannot be scored yet: h-index and i10-index are 0 and no citation counts exist for these two 2026 arXiv-only preprints. This axis would measure citations and uptake of the work by other researchers once the papers are indexed.
Focus0/100
The focus score is 0/100. The OpenAlex field distribution splits exactly 50/50 between machine learning (industrial AI / predictive maintenance) and reinforcement learning for industrial control, and the topic list shows five topics — machine fault diagnosis, reinforcement learning in robotics, reliability and maintenance optimization, fault detection and control systems, and advanced control systems design — each appearing once. With only two 2026 papers in the snapshot, this dispersion reads as a still-forming agenda rather than established breadth.
Collaboration46/100
Collaboration scores 46/100. The snapshot records 9 distinct coauthors across the two 2026 works, and on both arXiv preprints Cook sits in a middle author position without corresponding-author status. The frequent-coauthor list is dominated by a recurring team — Joseph E. Hernandez, John Burtenshaw, Sean Scott, Chris Couch, and the repeated 'Zhengyang'/'Gu' entries — suggesting small-group machine-learning projects rather than large consortia or solo authorship.
Activity32/100
Activity scores 32/100. The countsByYear series contains a single entry — 2 works in 2026 — and recent3yWorks is also 2, so the visible record is compressed into one year of arXiv preprints. Both items appeared in 2026 with no earlier output in this snapshot, giving a thin but recent publication cadence; a longer observation window would be needed to judge sustained productivity.
Reproducibility38/100
Reproducibility averages 38/100 across the 2 papers Evolve has reviewed. The PID-plus-reinforcement-learning hybrid is rated low readiness — single-episode evaluation and no significance tests — while the predictive-maintenance benchmark reaches medium readiness on the strength of 700+ reported runs, yet is weakened by synthetic linear targets, missing linear baselines, and absent significance testing. Neither paper currently meets a strong reproducibility bar.
Research summary
Thomas Cook is an early-stage researcher whose public bibliographic footprint, name-resolved from OpenAlex records, centers on machine learning for industrial systems: the field distribution splits exactly 50/50 between machine learning (industrial AI / predictive maintenance) and reinforcement learning for industrial control. The snapshot contains two works, both dated 2026 (countsByYear 2026 = 2; recent3yWorks = 2), both arXiv preprints on which Cook is a middle-listed, non-corresponding author. The first, 'Augmenting PID Control with Deep Reinforcement Learning: A Hybrid Approach to the Industrial Benchmark', proposes a hybrid PID controller whose Gain/Shift parameters are discovered by TD3; Evolve's review notes it matches the best DRL agent (-203.46 vs -204.92) while flagging single-episode evaluation, no significance tests, and an unproven MIMO premise. The second, 'Benchmarking Hybrid Deep Learning Architectures for Predictive Maintenance in Industry 4.0', claims an LSTM-Transformer advantage on noise robustness across 700+ runs, with synthetic linear targets and missing linear baselines weakening the evidence. Citation metrics are not yet available — h-index and i10-index are both 0 — so output and impact remain unscored. The record lists 9 distinct coauthors (collaboration 46/100), and the reproducibility mean across the two reviewed papers is 38/100.
Collaboration pattern
Cook appears as a middle-author, non-corresponding contributor on both reviewed 2026 arXiv preprints, working within teams rather than solo — 9 distinct coauthors are recorded and the collaboration score is 46/100. The frequent-coauthor list includes Joseph E. Hernandez, John Burtenshaw, Sean Scott, Chris Couch, and the repeated entries 'Zhengyang' and 'Gu'; 'Thomas Cook' itself also appears there, an apparent name-deduplication artifact in the source data.
Research keywords
Highlights
Hybrid PID controller with TD3-tuned gains matches best DRL agent on Industrial Benchmark
700+ run benchmark of LSTM-Transformer hybrids for predictive maintenance noise robustness
Timeline
- 2026Co-authored 'Augmenting PID Control with Deep Reinforcement Learning: A Hybrid Approach to the Industrial Benchmark' (arXiv) in a middle, non-corresponding position.
- 2026Co-authored 'Benchmarking Hybrid Deep Learning Architectures for Predictive Maintenance in Industry 4.0' (arXiv), again as a middle, non-corresponding author.
- 2026OpenAlex snapshot consolidates the record: 2 works in the recent 3-year window, 9 distinct coauthors, h-index and i10-index at 0, and a reproducibility mean of 38/100 across the two reviewed papers.
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.
- Bibliometrics are computed from OpenAlex only; the record carries no institutional affiliation, country, or ORCID.
- 'Thomas Cook' is a common name: this profile is name-resolved from public bibliographic data, and the frequent-coauthor list even contains a 'Thomas Cook' self-entry, so identity confidence is limited.
- Output and impact are unscored because h-index and i10-index are 0 and no citation counts exist for these two 2026 arXiv-only preprints.
- The frequent-coauthor list contains repeated raw entries ('Zhengyang', 'Gu'), indicating unresolved author-name splitting in the source data.