Research Constellation

How to read this

Each sparkling star is one of the 38 papers on the Abbasi Lab publications page (2016–2026). Each glowing orb is a person who co-wrote at least one of them. A filament joins every author to their paper, so two people are connected when they share a paper. Papers are grouped into seven research topics, and each topic has its own color and nebula.

• CenterReza Abbasi-Asl, the principal investigator, who is an author on every paper.
• Ringed orbsAbbasi Lab members: current team, alumni and trainees, as listed on the People page.
• Orb colorA person’s main topic (where most of their papers are). Orb size grows with their number of papers.

Explore

Drag to orbit, scroll or pinch to zoom. Click a paper for its authors and links, a person for their papers and closest collaborators, or a topic to see all of its papers. The timeline replays how the network grew. Press space to play or pause, / to search and Esc to reset. Links such as #topic=vision, #paper=12 or #p=Bin%20Yu open a view directly.

How the data was built

Papers are the 38 entries on abbasilab.org/publications (5 preprints and 33 peer-reviewed papers), captured on October 6, 2026. Full author names, DOIs and dates were checked against publisher and preprint-server records. Lab roles come from the People page.

• The Nature Communications paper lists 9 authors in the journal. The lab page shows 7, so Michael Kunst and Lydia Ng were added.
• Missing or broken links on the lab page were replaced with the paper’s DOI or OpenReview page (RECITE, the foundation-model robustness paper, the protein language model workshop paper and the DBS-entrained gamma paper).
• Where a listed preprint has since been published, “Read paper” opens the published version and the preprint is linked as well.
• Q. Cui on the papers is Gavin Cui on the People page. “Y. Wang” on the 2024 PNAS paper (Yu Wang) and the 2026 preprint (Yujing Wang) are two different people.
• Each paper has one topic, assigned by hand from its subject and the themes on the lab’s Research page.

Data snapshot: October 6, 2026. The layout is a force-directed 3D embedding computed once; distances are only roughly meaningful.