Industrial AI researcher: deep RL, MLOps, and hybrid deep learning for manufacturing control and predictive maintenance.
3
papers here
3
reviewed by Evolve
0
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—
No aggregate publication-volume data (total papers, per-year counts, output trend) is available in the current snapshot, so this axis is unscored. It would measure sustained research production across venues over time.
Impact—
Citation-based impact metrics (citation counts, h-index, venue influence) are not available for this profile, so the axis is left unscored. It would measure how widely the work is taken up and cited over time.
Focus0/100
The focus score is 0, and the field distribution confirms full breadth: three fields split almost evenly at about 0.33 each — industrial ML/MLOps for cyber-physical manufacturing, machine learning for predictive maintenance, and reinforcement learning for industrial control. These are adjacent areas within one industrial-AI program rather than scattered topics, so the 0 reflects evenly weighted, multi-track applied work rather than a fragmented agenda.
Collaboration48/100
The collaboration score is 41, based on 7 distinct co-authors across the catalog. Recurring collaborators include Zhengyang and Gu, alongside Joseph E. Hernandez, Chris Couch, John Burtenshaw, Sean Scott, and Fredaljohn Rohrbaugh. On all three papers Cook is a middle author and never the corresponding author, indicating contribution within larger teams rather than independent group leadership; the mid-range score is consistent with moderate but team-embedded collaboration.
Activity—
Temporal activity data (publication cadence across years and recent output trend) is insufficient in the current snapshot, so this axis is unscored. It would measure how actively the researcher publishes over time.
Reproducibility32/100
The reproducibility mean is 32 across 3 assessed papers. The digital-twin MLOps paper reports Cpk gains of 28-45% but with internal-only evaluation and no released code or data (low); the PID-plus-TD3 hybrid rests on single-episode results without significance testing (low); the predictive-maintenance benchmark is the strongest at medium, with 700+ runs yet synthetic linear targets and missing linear baselines. Reproducibility is the clearest improvement lever.
Research summary
Thomas Cook is an industrial AI researcher whose visible record in this catalog consists of three 2026 arXiv preprints, each with a middle authorship position and no corresponding-author status. The field distribution splits evenly (about 0.33 each) across industrial MLOps for cyber-physical manufacturing, machine learning for predictive maintenance, and reinforcement learning for industrial control — adjacent tracks of one applied program rather than scattered topics, consistent with a focus score of 0. The papers share a common theme: moving deep reinforcement learning and automated model lifecycle management into production-adjacent industrial settings. A digital-twin MLOps system for automotive manufacturing control reports Cpk gains of 28-45% over uncontrolled operation; a hybrid PID-plus-TD3 controller reportedly matches the best deep-RL agent on the Industrial Benchmark (-203.46 vs -204.92); and a 700+ run benchmark claims LSTM-Transformer noise robustness for predictive maintenance. Evolve's reviews flag recurring methodological gaps — internal-only evaluation, single-episode testing, missing significance tests, and no released code or data — yielding a reproducibility mean of 32 across three assessed papers (two low, one medium).
Collaboration pattern
Cook's visible work is team-embedded: 7 distinct co-authors appear across the three-paper catalog, with Zhengyang and Gu recurring alongside Joseph E. Hernandez, Chris Couch, John Burtenshaw, Sean Scott, and Fredaljohn Rohrbaugh. Collaboration scores 41/100, and Cook holds a middle, non-corresponding authorship position on every paper, consistent with a contributing specialist within engineering-led teams rather than a corresponding group lead.
Research keywords
Highlights
Digital-twin MLOps with RL control reports Cpk gains of 28-45%
Hybrid PID + TD3 controller claims parity with deep RL on the Industrial Benchmark
700+ run benchmark of hybrid LSTM-Transformer for noise-robust predictive maintenance
Timeline
- 2026Publishes autonomous model lifecycle management for digital-twin-based manufacturing control on arXiv, reporting Cpk gains of 28-45% in a deployed automotive MLOps setting.
- 2026Publishes a hybrid PID plus deep reinforcement learning approach to the Industrial Benchmark, reporting parity with the best DRL agent (-203.46 vs -204.92).
- 2026Publishes a benchmark of hybrid deep learning architectures for predictive maintenance in Industry 4.0, based on 700+ runs.
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 scores are computed from a small arXiv-only snapshot (three 2026 preprints); no journal, citation, or venue data is included.
- "Thomas Cook" is a common name; this profile is name-resolved from public bibliographic data and may not refer to a unique individual.
- Institution, country, and ORCID are unavailable, so affiliation-based context could not be verified.
- Output, impact, and activity axes are unscored because the underlying aggregate metrics are missing from the snapshot.