Online learning theorist at Google: minimax-optimal anytime regret and horizon-uniform bounds for inverse optimization
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
1
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-15) 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—
Not yet scored: this snapshot lacks reliable per-paper output data beyond seven works in the trailing three years. This axis would measure total and per-year publication volume, venue spread, and first-author share across the record.
Impact—
Not yet scored: citation evidence is too thin — the snapshot reports an h-index of 1 and an i10-index of 1. This axis would measure citation counts, h-index trajectory, and downstream influence on follow-up research.
Focus0/100
Scores 0 because the field distribution splits evenly between two labels — 50% online learning/learning theory and 50% online learning and optimization theory — rather than concentrating on one. The topic record still shows a coherent cluster: Auction Theory and Applications (4), Optimization and Search Problems (4), Game Theory and Voting Systems (3), Stochastic Gradient Optimization Techniques (3), and Game Theory and Applications (2), all orbiting decision-making under adversarial or strategic uncertainty.
Collaboration50/100
Mid-range score of 50, from eleven distinct coauthors. The frequent-coauthor list is dense with theory researchers — Yang Cai, Anupam Gupta, Aranyak Mehta, Grigoris Velegkas, Renato Paes Leme, Guru Guruganesh, Vineet Gupta, Yanchen Jiang, and Di Wang (Yang Cai and Di Wang each appear twice) — a roster consistent with his algorithmic game theory and online learning topic profile. Both catalog papers carry a middle author position without corresponding authorship, indicating steady participation in multi-author theory teams rather than solo work.
Activity61/100
Score of 61 reflects a recent surge: counts by year show one indexed work in 2017, one in 2024, and six in 2026, with seven works in the trailing three years. Both Evolve-reviewed papers appeared in 2026 on arXiv, so current output is concentrated and accelerating after quiet intervals; the mid-range score accounts for the sparse 2017–2024 record.
Reproducibility100/100
Perfect score of 100, computed from a reproducibility mean of 100 over two reviewed papers. Both 2026 arXiv papers carry high reproducibility readiness and human-verified proofs: Evolve confirmed the anytime-regret proofs check out while noting the overhead is asymptotic, and judged the O(d)-regret inverse optimization proof short, self-contained, and correct on extraction. Explicit constants and deterministic statements make this theory straightforward for others to verify.
Research summary
Christopher Liaw is a researcher at Google (United States) whose bibliometric profile, resolved from OpenAlex records, centers on the theoretical foundations of online learning and optimization. His field distribution splits evenly between online learning/learning theory and online learning and optimization theory, while his topic record extends into auction theory and applications, optimization and search problems, game theory and voting systems, and stochastic gradient optimization techniques. The snapshot lists eleven distinct coauthors, an h-index of 1, and an i10-index of 1, with activity concentrated in 2026: six indexed works that year and seven over the trailing three years. Two of his 2026 arXiv papers are human-verified in the Evolve catalog. The first, on prediction with expert advice, gives an anytime algorithm whose regret with many experts matches the fixed-time constant, removing a √2 factor; Evolve confirmed the proofs while noting the added overhead is asymptotic and vanishes very slowly. The second establishes deterministic O(d)-regret for proper online inverse optimization via a trace-power potential, with a proof judged short, self-contained, and correct on extraction. Both papers carry high reproducibility readiness, and the reproducibility dimension scores 100 across the two verified papers.
Collaboration pattern
The snapshot lists eleven distinct coauthors. The frequent-coauthor roster — Yang Cai, Anupam Gupta, Aranyak Mehta, Grigoris Velegkas, Renato Paes Leme, Guru Guruganesh, Vineet Gupta, Yanchen Jiang, and Di Wang (Yang Cai and Di Wang each appear twice) — clusters around algorithmic game theory, auctions, and online learning, matching his topic profile. On both cataloged papers he holds a middle author position without corresponding authorship, consistent with a collaborative contributor inside multi-author theory groups rather than a solo-line leader.
Research keywords
Highlights
Anytime regret with many experts matches the fixed-time constant
Deterministic O(d) regret for online inverse optimization
Timeline
- 2017Earliest work in this OpenAlex-resolved snapshot: one indexed publication recorded for 2017.
- 2024One indexed work recorded for 2024, a quiet interval before the 2026 output cluster.
- 2026Six indexed works recorded in 2026, with seven works over the trailing three years — an activity peak.
- 2026Evolve human-verifies two arXiv papers: anytime regret with many experts matching the fixed-time constant, and deterministic O(d)-regret online inverse optimization, both rated high reproducibility readiness.
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 come from an OpenAlex-resolved identity; Christopher Liaw is not a rare name, so records may include or exclude works by other researchers with the same name.
- The output and impact axes are unscored (dimMissing): citation evidence in this snapshot is limited to an h-index of 1 and an i10-index of 1.
- Only two papers, both 2026 arXiv preprints, are human-verified in the Evolve catalog; venue acceptance is not reflected.
- The reproducibility score of 100 is computed over just two reviewed papers and may not generalize.