AI Glossary
What is Multimodal?
A multimodal model accepts or produces multiple kinds of data: reading screenshots and PDFs, hearing audio, generating images or video. Frontier chat models are now multimodal on input by default; generation across modalities is typically handled by specialized models (image and video generators).
News momentum
Mentions in recent AI news titles and summaries, refreshed from the ingestion stream.
106
+33 vs prior 7d
- CRANE: Post-Hoc Counterfactual Retrieval for Multimodal Classifiers
arxiv-ai · 15h
- Rethinking Sequence Compression with Test-time Regression
arxiv-ai · 15h
- LACToS: Local Anomaly Detection via Conditional Typicality of Superpixels
arxiv-ai · 16h
- Just added to the catalog:
sglang-omni: Serve Interactive Multimodal Models with SGLang
https://osrepos.com/repo/sgl-project-sglang-omni?u…
bluesky-ai · 17h