ML theorist in robust learning under contamination and the algorithmic side of diffusion language models.
2
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
1
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-27) 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 score yet. This axis would measure total publication volume and productivity over time, but the snapshot lacks a complete works count — only 8 recent works are indexed — so overall productivity cannot be fairly scored.
Impact—
No impact score yet. This axis would measure citations, h-index trajectory, and influence on follow-up work; the snapshot reports an h-index of 1 and i10-index of 0 but lacks full citation totals, so influence is not scored.
Focus100/100
Scores 100 for focus. The field distribution is maximally concentrated — machine learning theory (computational learning theory) is the single assigned field — and recent output clusters tightly in time: 8 works in the last three years with 7 indexed in 2026. OpenAlex topics reinforce a narrow algorithmic band around polynomial and algebraic computation, with diffusion language models as the visible second thread in the reviewed papers.
Collaboration44/100
Scores 44 for collaboration. The recent network is real but compact: 8 distinct coauthors in the last three years, including repeat collaborators such as Anton Xue, Litu Rout, Konstantinos Stavropoulos, Sergei Tikhonov, and Arsen Vasilyan. Reviewed authorship spans first-author theory work and a middle-author systems paper, suggesting a mix of leading and supporting roles rather than a large distributed group.
Activity64/100
Scores 64 for activity. countsByYear shows a punctuated rhythm: 1 indexed work in 2016, 2 in 2023, 1 in 2025, then a sharp burst of 7 in 2026 — 8 works within the last three years. The current pace is the highest in the record, driven by 2026 arXiv output that includes both reviewed papers.
Reproducibility55/100
Scores 55 for reproducibility, computed from 1 assessed paper (reproMean 55). Both reviewed papers carry medium reproducibility readiness; Evolve's verdict on the diffusion-LM paper cautions that its core efficiency claim rests on a single confounded comparison pair, a concrete caveat for independent replication.
Research summary
Adam R. Klivans is a machine learning researcher whose profile is name-resolved from OpenAlex bibliographic data, pairing theoretical depth with a recent turn toward language-model architectures. The reviewed record centers on two 2026 arXiv papers: a first-author theoretical result, 'Efficient Robust Learning at the Information-Theoretic Limit,' resolving Blanc's open problem with a polynomial-time ERM-oracle learner that reaches optimal η+ε error under bounded contamination, plus poly-time η+ε Gaussian halfspaces; and a middle-author systems paper, 'dQwen3.5,' adapting hybrid autoregressive backbones (Gated DeltaNet) into hybrid-attention diffusion language models at roughly half the tokens of a matched control. Bibliometric signals show a sharply focused recent agenda — 8 works in the last three years, 7 indexed in 2026 — against a sparse earlier footprint (single indexed works in 2016 and 2025), an h-index of 1, and 8 distinct recent coauthors. OpenAlex topic assignments cluster on algebraic and polynomial computation and analytic number theory alongside learning theory, indicating a broadly algorithmic toolkit. Evolve rates both reviewed papers at medium reproducibility readiness.
Collaboration pattern
Recent work is collaborative but on a compact scale: 8 distinct coauthors in the last three years. Frequent collaborators named in the record include Anton Xue, Litu Rout, Aditya Akella, Sujay Sanghavi, Sanjay Shakkottai, Konstantinos Stavropoulos, Sergei Tikhonov, and Arsen Vasilyan. Authorship spans both roles — first author on the robust learning theory paper and middle author on the diffusion LM paper — indicating he both leads projects and contributes to larger team efforts.
Research keywords
Highlights
Resolves Blanc's open problem in robust learning
Adapts AR backbones into hybrid-attention diffusion LMs
Timeline
- 2016Earliest indexed activity in the OpenAlex snapshot, with 1 work recorded in this year.
- 20232 works indexed in this year, an early sign of resumed output after the sparse earlier record.
- 20251 work indexed in this year, maintaining a low but continuous publication rhythm.
- 2026Sharpest burst in the record: 7 works indexed, including both reviewed arXiv papers — the first-author robust learning result and the middle-author hybrid-attention diffusion LM paper.
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 (h-index, countsByYear, topics, coauthor counts) come from an OpenAlex snapshot; the profile is name-resolved from public bibliographic data and may merge or split identities.
- Output and impact dimensions are unscored because full publication and citation totals are absent from the snapshot.
- Only 2 papers carry Evolve reviews, and the reproducibility score draws on just 1 assessed paper.
- Institution, country, and ORCID are not provided, so affiliation and career context are unknown.