MIT Media Lab veteran pairing long-run mobility and network research with new LLM-agent social simulation work.
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
122
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—
No output score has been computed yet. This axis would measure publication volume and venue distribution over time; the snapshot lists raw per-year counts (63 works in the last three years) but Evolve has not normalized them into a score.
Impact—
No impact score has been computed yet. This axis would measure normalized citation influence on the field; the snapshot reports h-index 122 and i10-index 479, but a field-weighted impact score is not part of this snapshot.
Focus0/100
Focus scores 0, reflecting exceptional topical breadth rather than narrowing. The snapshot splits the reviewed work evenly (0.5 / 0.5) between LLM-based social simulation and human-computer interaction, while the five largest topic clusters — human mobility (121), complex networks (113), activity recognition (76), vision and imaging (74), and human-technology interaction (72) — span several distinct research communities.
Collaboration44/100
The collaboration axis scores 44. Within the Evolve-reviewed snapshot the co-author graph is compact: 8 distinct coauthors, with recurring partners including Hafsa Akbar, Daniel Platnick, Marjan Alirezaie, Hossein Rahnama, José Ramón Enríquez, and Jiaxin Pei. Both catalog papers list the profile subject in the final author position without a corresponding-author flag, consistent with a senior-supervisor pattern on small, theory-driven teams rather than large consortia.
Activity100/100
Activity scores 100, the strongest axis in this snapshot. The record lists 63 works over the last three years, with per-year counts of 37 in 2023, 37 in 2024, and 21 in 2025, plus 5 already logged for 2026. Output remains high across a record stretching back decades, peaking at 51 works in 2021 and 47 in 2018 within the provided counts.
Reproducibility38/100
Reproducibility scores 38, computed as a mean over the 2 papers reviewed so far. The 2026 Bayesian belief-layer paper is rated medium readiness — theory-matched results but a single toy topic and only N=20 — while the deliberation position paper is rated low, its evidence resting on the authors' own unpublished working papers.
Research summary
Alex 'Sandy' Pentland, based at the MIT Media Lab (US), is profiled here from OpenAlex-resolved bibliographic data together with two papers reviewed in the Evolve catalog. The record is long and unusually productive: per-year counts run from the 1970s to the present, with a 68-work peak in 2002 and a late-career acceleration reaching 47 works in 2018 and 51 in 2021; the last three years list 63 works, with 37 logged in each of 2023 and 2024. Snapshot indicators include h-index 122 and i10-index 479. Topical mass concentrates in human mobility and location-based analysis (121 works), complex network analysis (113), context-aware activity recognition (76), advanced vision and imaging (74), and human-technology interaction (72), which is why the focus axis scores 0 — an even 0.5/0.5 split between LLM-based social simulation and human-computer interaction in the reviewed set. The recent Evolve-reviewed work pivots toward LLM agents: a 2026 Bayesian belief layer giving agents controllable, recoverable stubbornness within Friedkin–Johnsen opinion dynamics, and a 2026 position paper arguing that AI should facilitate democratic deliberation at scale under four design principles. Evolve's verdicts praise the former's theory-matched results (one toy topic, N=20) and find the latter timely but reliant on the authors' unpublished working papers.
Collaboration pattern
The reviewed work is team science but small-team science: 8 distinct coauthors appear in the snapshot, and the frequent-coauthor list centers on Hafsa Akbar, Daniel Platnick, Marjan Alirezaie, Hossein Rahnama, José Ramón Enríquez, and Jiaxin Pei. On both 2026 catalog papers the profile subject sits in the last author position without the corresponding-author flag, a pattern typical of a senior collaborator advising earlier-career first authors.
Research keywords
Highlights
Controllable stubbornness for LLM agents
A design case for AI-scale democratic deliberation
Timeline
- 1976The earliest cluster in the OpenAlex record appears, with 3 works logged in 1976.
- 2002Highest single-year output in the provided counts: 68 works logged in 2002.
- 201847 works logged in 2018, the start of a sustained late-career run that continues at 45 in 2019.
- 2021Annual count reaches 51 works, the highest since the 2002 peak.
- 2026Two arXiv papers enter the Evolve catalog with the researcher in the last author position: a Bayesian belief layer for controllable opinion dynamics in LLM agents, and a position paper on AI-facilitated democratic deliberation at scale.
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.
- All bibliometrics come from an OpenAlex-resolved snapshot; h-index and i10-index are quoted as provided and not independently verified.
- The per-year counts include an isolated 1946 entry (1 work), likely a record-linkage artifact, and the frequent-coauthor list contains two name variants of the subject themself.
- Evolve's output and impact composite scores are not yet computed, and only 2 papers have been reviewed, so the reproducibility mean (38) rests on a very small base.
- This profile is name-resolved from public bibliographic data only and draws on no biography, award, or non-publication sources.