TUM Klinikum researcher working across neuroimaging AI (MRI-to-PET translation) and LLM agent-harness efficiency
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
0
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—
Output metrics are insufficient here: the OpenAlex snapshot reports h-index 0 and no multi-year publication totals beyond 2026, so this axis is unscored. It would measure cumulative publication volume and per-year productivity across the researcher's career.
Impact—
Impact data is not yet available: with h-index 0 and i10-index 0, no citation-based signal exists in the snapshot, so this axis is unscored. It would measure citation counts, field-weighted influence, and uptake of the researcher's methods over time.
Focus0/100
Focus scores 0 because the record is evenly split: fieldDist assigns 0.5 to AI agents / LLM systems and 0.5 to medical image analysis (neuroimaging). Topic clusters — advanced MRI techniques (2) alongside dementia and cognitive impairment, medical image segmentation, medical imaging applications, and software engineering (1 each) — confirm a dual-track early portfolio rather than a consolidated specialty.
Collaboration61/100
Collaboration scores 61, grounded in 20 distinct coauthors within the recent three-year window. Twelve frequent coauthors are listed in the catalog — including Haozhe Liu, Tian Ye, Mingchen Zhuge, Ping Luo, and Enze Xie — pointing to dense, team-scale projects: Li was first author on the neuroimaging paper and a middle author on the larger SoL-Pi effort.
Activity40/100
Activity scores 40. The snapshot records 3 works in the last three years, all concentrated in 2026 (countsByYear: 2026 = 3), with no earlier output visible. Both catalog papers are 2026 arXiv releases — a first-author neuroimaging method and a middle-author agent-harness study — indicating a recent, front-loaded start rather than a sustained multi-year cadence.
Reproducibility55/100
Reproducibility scores 55, computed from 2 reviewed papers (reproMean 55), each rated medium for reproducibility readiness. SoL-Pi is credited with a solid isolation protocol but single-run statistics; the MRI-to-PET paper shows strong results with caveats — narrow gains over bridge baselines and different readers for MRI versus synthetic PET.
Research summary
Yitong Li, affiliated with TUM Klinikum in Germany, is an early-stage researcher whose public OpenAlex record (ORCID 0009-0009-3874-6055) shows a young but distinctly dual-track profile. Field distribution splits evenly at 0.5/0.5 between medical image analysis (neuroimaging) and AI agents / LLM systems, with topic clusters in advanced MRI techniques (2 records) plus dementia and cognitive impairment, medical image segmentation, medical imaging applications, and software engineering (1 each). Three works are recorded in the last three years, all dated 2026. In the Evolve-reviewed catalog, Li is first author of 'Bridging Modalities on the Cortex', a surface-based MRI-to-PET translation method using a diffusion bridge and a spherical Transformer for FDG-PET synthesis in dementia diagnosis — assessed as the first surface-based approach of its kind, with strong results but narrow gains over bridge baselines. Li also appears as a middle author on 'SoL-Pi', which recursively scales auto-research loops and reports a harness matching Pi at roughly 45% fewer tokens on EdgeBench. Citation metrics (h-index 0, i10-index 0) are too early to be meaningful, while the collaboration score (61) rests on 20 distinct coauthors and reproducibility (55) on two medium-readiness papers.
Collaboration pattern
Collaboration (score 61) is team-centric: 20 distinct coauthors appear in the recent window, and 12 frequent coauthors are listed in the catalog — Haozhe Liu, Tian Ye, Sensen Gao, Qihang Cao, Mingchen Zhuge, Duomin Wang, Ruihua Zhang, Ping Luo, Jiawang Bian, Lei Zhu, Ligeng Zhu, and Enze Xie. Li both leads (first author on the MRI-to-PET paper) and supports (middle author on SoL-Pi) within large arXiv-era teams, a pattern typical of fast-moving agent-systems and imaging work.
Research keywords
Highlights
First surface-based MRI-to-PET diffusion bridge
Auto-research loops cut agent-harness token cost by ~45%
Dual-track 2026 debut with medium reproducibility readiness
Both 2026 arXiv releases — the neuroimaging diffusion bridge (first author) and the SoL-Pi harness (middle author) — are rated medium for reproducibility readiness, yielding a reproMean of 55 across 2 papers: an early signal of method-level rigor rather than a mature open-source replication track record.
Timeline
- 2026First-author arXiv release of 'Bridging Modalities on the Cortex', a surface-based MRI-to-PET translation method built on a diffusion bridge and spherical Transformer.
- 2026Middle-author arXiv release of 'SoL-Pi', recursively scaling auto-research loops for an efficient agent harness.
- 2026OpenAlex snapshot records 3 works dated 2026 spanning neuroimaging and LLM-agent topics, with 20 distinct coauthors.
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; h-index and i10-index are 0 and no per-paper citation counts are in the snapshot, so the output and impact axes are marked missing rather than scored.
- The reviewed catalog covers just 2 arXiv papers (both 2026); the record is early-stage and may not reflect a complete publication history.
- 'Yitong Li' is a common name; this profile is name-resolved from public bibliographic data (OpenAlex, ORCID 0009-0009-3874-6055) and the listed institution, so identity ambiguity is possible.
- Assessment claims (e.g., ~45% token savings, first surface-based bridge) come from Evolve's own paper reviews, not independent citation or replication evidence.