Warwick researcher bridging privacy auditing of graph generative models and large-scale Ethereum DeFi benchmarking
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-09) 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 output metrics — publication counts, venue tiers, or annual volume — are available in this snapshot, so the output axis is unscored. It would measure the size and consistency of the published record over time, beyond the two 2026 arXiv papers reviewed by Evolve.
Impact—
Citation counts, h-index, and downstream adoption are absent from this snapshot, so the impact axis is not scored. It would measure how far the benchmark and privacy-audit papers propagate through and are reused by the wider literature.
Focus0/100
The focus score is 0, the floor of the 0-100 scale, derived from a field distribution split exactly 50/50 between machine learning for blockchain/DeFi analytics and differential privacy / machine-learning privacy auditing. The two reviewed 2026 papers sit one in each line, so the current catalog reads as deliberate breadth across two empirical research programmes rather than concentration in a single subfield.
Collaboration41/100
Collaboration scores a moderate 41, computed from 7 distinct coauthors in the snapshot. Both reviewed papers are team efforts: Maple holds a middle author position on the edge-private graph generator audit and the last position on ETH-TraceBench, neither as corresponding author. Recurring collaborators include Gesine Reinert, Lukasz Szpruch, Andrew Elliott, Kemal Kirtac, Anum Fatima, Stratis Limnios, and James Adams, suggesting stable partnerships across two teams.
Activity—
Recent-activity signals such as publication cadence and preprint frequency are unavailable in this snapshot, so the activity axis is unscored. It would measure ongoing publishing pace beyond the two 2026 arXiv papers reviewed here.
Reproducibility78/100
Reproducibility is the strongest measured axis at 78, a mean computed over 2 papers. Evolve rates the edge-private graph generator audit 'high' for reproducibility readiness, crediting its Clopper-Pearson-bounded attack protocol, and ETH-TraceBench 'medium', reflecting a released 1.35-billion-log-event benchmark alongside missing deep baselines. Both papers share a statistically grounded, artefact-backed evaluation style, which drives this above-average score.
Research summary
Carsten Maple is a researcher at the University of Warwick whose reviewed work bridges two areas in near-equal measure: machine learning for blockchain/DeFi analytics and differential privacy / machine-learning privacy auditing, each accounting for half of the field distribution in the Evolve snapshot. The 2026 ETH-TraceBench paper, where Maple is the last-listed author, releases a 1.35-billion-log-event benchmark for Ethereum DeFi and shows that simple models reach 0.95-0.96 macro-F1 yet lose 0.16-0.19 on unseen protocols, exposing distribution-shift risk; Evolve notes the absence of deep baselines. A 2026 companion audit of edge-private graph generators, with Maple in a middle author position, applies Clopper-Pearson-bounded membership inference attacks to quantify edge-level differential-privacy leakage, judged rigorous and honest though limited to two small networks and two generators. Reproducibility is the strongest measured dimension at 78 across the two papers, earning 'high' and 'medium' readiness ratings, while collaboration registers a moderate 41 across 7 distinct coauthors and focus scores 0, reflecting the even two-field split. Output, impact, and activity axes lack data in the current snapshot and are left unscored. The profile is therefore an early-stage, two-paper view of a broader record not captured here.
Collaboration pattern
Collaboration scores a moderate 41, with 7 distinct coauthors in the snapshot. Recurring collaborators include Gesine Reinert, Lukasz Szpruch, Andrew Elliott, Kemal Kirtac, Anum Fatima, Stratis Limnios, and James Adams. Both reviewed papers are team efforts rather than solo work: Maple holds a middle author position on the edge-private graph generator audit and the last position on ETH-TraceBench, with neither paper listing him as corresponding author — a pattern of senior, distributed authorship across two teams.
Research keywords
Highlights
ETH-TraceBench: 1.35B log events expose DeFi model fragility under protocol shift
Statistically bounded audit of edge-DP leakage in graph generators
A reproducibility-first empirical style across both research lines
Both reviewed 2026 papers pair released artefacts with statistically grounded evaluation — bounded attacks and a large public benchmark — yielding a reproducibility mean of 78 over 2 papers, with readiness ratings of high and medium in the Evolve snapshot.
Timeline
- 2026Releases ETH-TraceBench, a large-scale event-stream benchmark of 1.35 billion log events for Ethereum DeFi under temporal, protocol, and contract shift.
- 2026Publishes an empirical audit of edge-private graph generators using Clopper-Pearson-bounded membership inference attacks.
- 2026Both 2026 arXiv papers enter the Evolve catalog, anchoring a reproducibility mean of 78 across the two reviewed works.
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.
- Dimension scores come from a bibliometric snapshot; the output, impact, and activity axes have insufficient data and are left unscored.
- The reproducibility mean of 78 rests on only 2 reviewed papers, so it may not generalize to the wider record.
- No ORCID or country metadata is present; the profile is name-resolved from public bibliographic data and may not capture the researcher's complete publication list.
- Both reviewed works are 2026 arXiv preprints, so findings and benchmark figures may change in peer-reviewed versions.