AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

SPATIA: Multimodal Generation and Prediction of Spatial Cell Phenotypes

Zhenglun Kong, Mufan Qiu, John Boesen, xiang lin, Sukwon Yun, Tianlong Chen, Manolis Kellis, Marinka Zitnik

ICML 2026 · Image generation › Autoregressive and latent generation · approach: Retrieval and matching with generative features · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add confidence-aware OT reweighting to flow matching (ablation)SPATIA base · Conditional morphology generationFID: 59.5 -> 59.1Improved
add morphology-profile alignment loss (ablation)SPATIA + Reweight · Conditional morphology generationWasserstein correlation: 0.91 -> 0.94Improved
add morphology-profile alignment loss (ablation)SPATIA + Reweight · Conditional morphology generationKS: 0.62 -> 0.65Improved
add reconstruction (MAE) loss (ablation)cell-level backbone · Cell classificationaccuracy: 0.93 -> 0.94Improved
add multi-level (niche and tissue) hierarchy (ablation)cell-level model with MAE loss · Cell classificationaccuracy: 0.94 -> 0.97Improved
add cross-attention fusion (ablation)multi-level SPATIA · Cell classificationaccuracy: 0.97 -> 0.98Improved
add niche-level context (ablation)SPATIA generation · Conditional morphology generationFID: 60.3 -> 59.2Improved
corrupt 10-20% of OT pairs randomly (ablation)SPATIA flow module · Conditional morphology generationFID: 59.2 -> 61.0 -> 63.8Hurt
use SPATIA instead of single-cell baselinescell clustering embeddings · CosMxNMI: 0.560 -> 0.490Mixed
use SPATIA embeddings instead of UNIBCNB receptor status prediction · BCNB ER AUCAUC: 0.891 -> 0.902Improved
use SPATIA instead of CellPLMcell annotation · MS datasetF1: 0.709 -> 0.725Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy