TDA×ML researcher building certified multiparameter embeddings and exact permutation tests for neural representations
2
papers here
2
reviewed by Evolve
2
first-author
1
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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 is not scored: the bibliometric snapshot contains no publication-volume or venue record for this researcher beyond the two reviewed arXiv preprints, so there is insufficient data to measure productivity. This axis would count papers and venues over time.
Impact—
Impact is not scored: no citation counts, h-index, or downstream-usage data are present in the input. This axis would measure how often the researcher's work is cited or built upon by others, which cannot be estimated here.
Focus0/100
The focus score is 0. The underlying field distribution splits evenly, 0.5 to 'topological data analysis / machine learning' and 0.5 to 'machine learning / topological data analysis' — two mirror-labeled variants of the same TDA × ML intersection. Read literally, the corpus is a single narrow specialty recorded under inconsistent field strings rather than a genuinely scattered agenda; the zero reflects that 50/50 label split across the reviewed papers.
Collaboration28/100
The collaboration score is 28, based on 3 distinct coauthors in the available record. The frequent-collaborator list is short and stable: Atish Mitra, Žiga Virk, and Pramita Bagchi. Both catalog papers are first-authored by Majhi (one also corresponding), consistent with a small-team, lead-author-driven working style rather than large consortium science. With only two reviewed papers, this low score reflects a thin record as much as a solitary pattern.
Activity—
Activity is not scored: the snapshot provides no multi-year publication timeline, so recent-output cadence cannot be measured. This axis would track publishing frequency and recurrence across years; the record here spans only the 2026 preprints.
Reproducibility100/100
The reproducibility score is 100, the mean over the 2 papers assessed, both rated 'high' readiness. Evolve's reviews describe a careful, honest theory campaign in COMPLEX and a rigorous, transparent evaluation in the topological-interaction paper, with explicitly stated scope limits (small vision models on PCA shadows; operationally empty per-prediction certificates disclosed). This is a strong artifact-and-transparency signal, though computed from only two preprints.
Research summary
Sushovan Majhi is a researcher at George Washington University working at the intersection of topological data analysis (TDA) and machine learning. This profile is name-resolved from public bibliographic data (ORCID 0000-0001-8689-4321) and rests on two 2026 arXiv preprints in the Evolve catalog, on both of which the researcher is first author and on one of which also corresponding author. The first, COMPLEX, proposes a closed-form certified embedding of multiparameter persistence modules, claiming the first two-sided certified mp-persistence embedding; Evolve's review credits careful, honest theory and state-of-the-art Orbit accuracy while flagging that per-prediction certificates are operationally empty. The second supplies exact permutation tests for a topological class-overlap measure (Euler characteristic profiles) in neural representations, with a scope limited to small vision models on PCA shadows. Both papers carry 'high' reproducibility readiness, and the reproducibility dimension scores 100 (mean over 2 papers). The corpus is highly concentrated: both entries sit in mirror-labeled TDA/ML fields (0.5/0.5 split), and recorded collaboration is small, with 3 distinct coauthors (Atish Mitra, Žiga Virk, Pramita Bagchi) and a collaboration score of 28. Output, impact, and activity dimensions are not scored because the underlying bibliometric data is insufficient; no citation or publication-volume figures are quoted here.
Collaboration pattern
The available record shows a compact collaboration footprint: 3 distinct coauthors, with Atish Mitra, Žiga Virk, and Pramita Bagchi as the frequent partners. Both catalog papers are first-authored by Majhi — one also corresponding — suggesting a lead-author, small-team working mode in the TDA × ML niche rather than large distributed collaborations. The collaboration dimension scores 28, though with only two reviewed papers this number should be read as a thin-record signal, not a settled pattern.
Research keywords
Highlights
First closed-form certified embedding for multiparameter persistence
Exact permutation tests for topological class overlap in neural representations
Uniformly high reproducibility readiness across the catalog
Both reviewed papers carry 'high' reproducibility readiness, giving a reproducibility mean of 100 over 2 papers — a consistent artifact-and-disclosure pattern in an early, thin record.
Timeline
- 2026COMPLEX arXiv preprint appears, claiming the first two-sided certified embedding of multiparameter persistence modules (Majhi as first author).
- 2026Second arXiv preprint on certified topological interaction in neural representations appears, with Majhi as first and corresponding author.
- 2026Evolve reproducibility assessment covers both 2026 preprints, both rated 'high' readiness (reproducibility mean 100 over 2 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.
- Bibliometric dimensions are computed from a small public bibliographic snapshot; the output, impact, and activity axes have insufficient underlying data and are left unscored.
- The profile rests on only 2 reviewed items, both 2026 arXiv preprints; no peer-reviewed venue record or citation data is included.
- Identity is name-resolved from public bibliographic data (ORCID 0000-0001-8689-4321); no biographical details beyond the listed institution are verified.
- The focus score of 0 derives from a 0.5/0.5 split between two mirror-labeled field strings describing the same TDA/ML intersection.