Sambit Mishra

University of Southern California · United States

Unclaimed · AI-structured profileOpenAlex

Early-career causal inference researcher at USC proving identifiability of mixed ordinal–exponential causal DAGs

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-09-28) 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 yet: the indexed record lists an h-index and i10-index of 0 with only 2026 arXiv preprints, so publication volume and venue-weighted productivity cannot be assessed. This axis would measure total works, venue quality, and sustained productivity.

Impact—

Impact cannot be scored yet: no citation counts are indexed for the 2026 preprints, and h-index and i10-index are both recorded as 0. This axis would measure citations, citation velocity, and uptake of the identifiability results by the causal discovery community.

Focus100/100

Focus scores a perfect 100: the field distribution lists a single field — causal inference / statistical machine learning — and the topic profile is equally concentrated, with 4 works each in Advanced Causal Inference Techniques and Bayesian Modeling and Causal Inference, against 1 each in statistical methods, AI planning, and explainable AI. Both reviewed preprints target DAG identifiability, confirming a deliberately narrow theoretical agenda.

Collaboration28/100

Collaboration scores 28, the lowest recorded axis. The indexed record shows only 3 distinct co-authors — Urbashi Mitra, Yingying Wang, and Christine K. Johnson — and both cataloged preprints are first-authored submissions with no corresponding-author role flagged. This is a small, tight writing circle typical of a theory-first, preprint-stage researcher, leaving clear scope for a broader experimental or domain-applied network.

Activity47/100

Activity scores 47. All four indexed works carry a 2026 publication year (countsByYear: 2026 — 4; recent3yWorks: 4), so the record shows a concentrated recent burst rather than a multi-year track record. Both cataloged papers appeared on arXiv in 2026 as first-authored preprints, marking an intense but very recent start; a longer time series is needed to read a sustainable output rhythm.

Reproducibility100/100

Reproducibility scores a perfect 100, computed from 1 assessed paper (reproMean 100, reproPapers 1). Within the catalog, the mixed ordinal–exponential identifiability paper carries high reproducibility readiness and the epidemiological-graph paper medium; Evolve's verdicts note synthetic-only experiments and missing baselines, so the score reflects formally checkable mathematical claims rather than released code or real-data pipelines.

Research summary

Sambit Mishra is an early-stage researcher at the University of Southern California working at the intersection of causal discovery and statistical machine learning. The OpenAlex-indexed record (name-resolved; no ORCID linked) lists four works, all dated 2026, with an h-index and i10-index of 0 — consistent with a preprint-stage career. Two arXiv preprints reviewed by Evolve, both first-authored, attack a single well-defined problem: whether edge directions in causal DAGs mixing ordinal and exponential-family variables are identifiable under linear parametric models. One paper proves every-parameter identifiability with sharp converses; the companion extends the program to epidemiological causal graph identification, proving generic identifiability of ordinal–exponential edge directions in mixed DAGs. Evolve's assessments praise the rigor of the theory while flagging synthetic-only validation and missing baselines. Radar scores reflect the profile: a perfect 100 on focus, with indexed topics clustering in Advanced Causal Inference Techniques and Bayesian Modeling and Causal Inference; 100 on reproducibility (computed from one assessed paper); 47 on activity; and 28 on collaboration, with only three distinct co-authors. Output and impact axes are not yet scorable. The natural next step is empirical benchmarking against methods such as DAGMA on real epidemiological data.

Collaboration pattern

The collaboration footprint is deliberately small: only 3 distinct co-authors appear in the indexed record — Urbashi Mitra, Yingying Wang, and Christine K. Johnson — and both reviewed preprints are first-authored with no corresponding role flagged. The radar collaboration score of 28 reflects this tight circle. Output to date is theory-heavy and individually driven, so the co-author network has yet to expand into experimental or domain-applied teams.

Research keywords

causal discoveryDAG identifiabilitydirected acyclic graphsexponential familyordinal regressionordered logitMarkov equivalence classBayesian inferenceDAGMAexplainable AI

Highlights

Every-parameter identifiability for mixed ordinal–exponential causal DAGs
First-authored 2026 arXiv preprint proves that edges mixing ordinal and exponential-family variables are identifiable in causal DAGs under linear parametric models, with sharp converses establishing the limits. Evolve rates the theory rigorous but the validation synthetic-only; reproducibility readiness is high.
Generic identifiability of ordinal–exponential edge directions for epidemiological causal graphs
A companion first-authored 2026 preprint extends the program to epidemiological causal graph identification, proving generic identifiability of ordinal–exponential edge directions in mixed DAGs and cataloging the field's challenges. Evolve notes the theory remains bivariate-only and the experiments lack baselines and real data; reproducibility readiness is medium.

Timeline

2026Posts 'On the Identifiability of Mixed Ordinal and Exponential Family Causal DAGs under Linear Parametric Models' to arXiv as first author, proving every-parameter identifiability with sharp converses.
2026Posts 'Epidemiological Causal Graph Identification: Challenges, Identifiability and Algorithms' to arXiv, extending the identifiability program toward epidemiological applications.
2026OpenAlex-indexed record reaches 4 works in a single year (countsByYear 2026: 4), while h-index and i10-index remain at 0.

Papers in the Evolve catalog

2026-09-18first author
Epidemiological Causal Graph Identification: Challenges, Identifiability and Algorithms
2026-09-17first author
On the Identifiability of Mixed Ordinal and Exponential Family Causal DAGs under Linear Parametric Models

Frequent co-authors

Urbashi Mitra×2 Yingying Wang×1 Christine K. Johnson×1

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 OpenAlex only; the identity is name-resolved from public bibliographic data and no ORCID is linked, so same-name conflation cannot be ruled out.
h-index and i10-index are recorded as 0 and all indexed works are 2026 arXiv preprints, so the output and impact axes remain unscored.
The reproducibility score of 100 rests on a single assessed paper (reproPapers: 1) and reflects theoretical verifiability, not released code or real-data pipelines.
Qualitative findings draw on only two Evolve-reviewed papers; peer-reviewed venue acceptance is not yet recorded.