Co-author on 2026 arXiv work in video-model visual benchmarking and autonomous LLM research agents
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-19) 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—
Publication and citation aggregates are not yet available in this snapshot, so output is unscored. This axis would measure cumulative publication volume across venues and years; two Evolve-reviewed 2026 arXiv papers are too few to characterize it.
Impact—
Citation-based impact data (citations, h-index, i10-index) is insufficient in this snapshot, so impact is unscored. The axis would measure scholarly uptake and influence of the two 2026 papers; it is too early to assess.
Focus0/100
The focus score of 0 reflects an evenly split record: OpenAlex assigns 50% to computer vision (multimodal machine learning) and 50% to artificial intelligence (autonomous LLM research agents), and each of three indexed topics carries a count of one. With one cataloged paper per field, breadth currently exceeds depth in this early profile.
Collaboration56/100
A collaboration score of 56 rests on 16 distinct co-authors in the recent window. Recurring names include Rwiddhi Chakraborty (Oliver), Cheng Zhang, Zhuoran You, Michael Kampffmeyer, Yong Jae Lee, Fernando De la Torre, and Robert Jenssen, suggesting sustained work within multi-author teams. On both cataloged papers Fan Bai holds a middle authorship position without corresponding-author status, consistent with a contributing rather than leading role at this stage.
Activity20/100
An activity score of 20 comes from thin indexed volume: countsByYear lists a single work in 2026, and recent3yWorks is 1. The Evolve catalog independently holds two 2026 arXiv papers, indicating OpenAlex coverage is incomplete or lagging. All indexed output falls in one year, so publication cadence cannot yet be characterized beyond this single recent burst.
Reproducibility55/100
Reproducibility averages 55 across the two scored papers, both rated medium for reproducibility readiness. The limits are statistical: the video-benchmark study's compelling finding — Gemini dropping from 74% to 30% — rests partly on small datasets, including a 20-video CMD split, while PrimeScientist's attempt-efficiency gains look solid but its headline reward gains come from single runs.
Research summary
Fan Bai (name resolved from OpenAlex public bibliographic data; a relatively common name) appears in the Evolve catalog as a co-author — middle position, non-corresponding — on two 2026 arXiv papers sitting in two distinct fields. The first, "Not Another Text Benchmark: Putting the "Visual" Back in Visual Question Answering for Large Video Models," belongs to multimodal machine learning and reports frontier video models such as Gemini collapsing from 74% to 30% when multiple-choice options are presented as visual frames. The second, "PrimeScientist: Strategic Allocation of Research Effort in Autonomous Research," belongs to autonomous LLM research agents and applies budget-adaptive Monte Carlo tree search over executable plan trees. OpenAlex's field distribution splits evenly, 50/50, between these two areas, and each of three indexed topics carries a single paper. The recent-window record lists 16 distinct co-authors, with recurring names including Rwiddhi Chakraborty (Oliver), Cheng Zhang, Michael Kampffmeyer, and Fernando De la Torre. Evolve rates both papers medium for reproducibility readiness, averaging 55; the benchmark study is limited by small datasets, and the agent study's headline reward gains rest on single runs. Citation aggregates are not yet available, making this an early-stage profile.
Collaboration pattern
Fan Bai's indexed collaboration network spans 16 distinct co-authors in the recent window, with recurring names including Rwiddhi Chakraborty (Oliver), Cheng Zhang, Zhuoran You, Michael Kampffmeyer, Yong Jae Lee, Fernando De la Torre, and Robert Jenssen. On both Evolve-cataloged papers Fan Bai appears in a middle authorship position without corresponding-author status, indicating a contributor role within larger multi-author teams rather than an independent research line so far.
Research keywords
Highlights
Visualizing VQA: exposing text shortcuts in large video models
Budget-aware MCTS for autonomous research effort allocation
Timeline
- 2026Co-authored (middle position) the arXiv benchmark study "Not Another Text Benchmark: Putting the "Visual" Back in Visual Question Answering for Large Video Models."
- 2026Co-authored (middle position) the arXiv paper "PrimeScientist: Strategic Allocation of Research Effort in Autonomous Research."
- 2026Both papers entered the Evolve review catalog, each rated medium for reproducibility readiness with a reproducibility mean of 55.
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 are computed from OpenAlex only; citation-based aggregates (citations, h-index, i10-index) are not yet sufficient for this profile.
- The name Fan Bai is common; this profile is name-resolved from public bibliographic data, with no institution, country, or ORCID confirmed.
- Only two Evolve-reviewed papers (both 2026 arXiv, middle authorship) underpin this profile — an early-stage record.
- OpenAlex indexed volume (one 2026 work) undercounts the two cataloged arXiv papers, suggesting incomplete author disambiguation or indexing lag.