JHU Study: Few-Shot In-Context Learning Emerges Convergently Across Six Modalities
jhu-clsp · hf · 2026-09-16
Researchers at JHU CLSP built a controlled cross-modality framework to test the 'Convergent Emergence Hypothesis': when few-shot in-context learning (ICL) emerges, it should share a common cross-modality difficulty profile.
- The same task suite was instantiated across six modalities: language, genome, integer sequences, time series, images, and proteins
- Paired-mapping ICL emerged in all six modalities and surpassed controlled baselines
- Per-task effects were correlated across five modalities, partially supporting the hypothesis
The work extends ICL research beyond text, suggesting the ability to infer abstract patterns from examples may be a general emergent property of autoregressive training rather than a quirk of human-language data.
Related event: JHU study finds in-context learning emerges across six modalities(4 posts)→
More from Research
- Adobe & UCLA's Tev: verifiable progress rewards lift long-horizon agent Pass@5 from 0.17 to 0.28 — jiqizhixin · 2026-11-02
- Old trick still works: portal sampling cuts path-tracing noise dramatically at 64spp — ssh4net · 2026-11-02
- Marcel Böhme's paper asks why cybersecurity can never be solved, proposes vulnerability-guided hardening — mboehme_ · 2026-11-02
- Perplexity open-sources WANDR: 500 research tasks, 170,495 evidence points, URL-rechecking grader — thisguyknowsai · 2026-11-02
- AI researcher: we have no reliable way to tell whether an AI system is conscious — dioscuri · 2026-11-02
- Ant Group's Venus-DefakerOne2 hits 96% accuracy detecting AI-generated videos — inclusionAI · 2026-11-02