icml-2026-4764f538·2 events·first seen Aliases: ICML 2026
A keynote address delivered at ICML 2026, published via the AI Snake Oil newsletter, examines the question of what meaningful research and work remains for humans as AI capabilities expand. The piece comes from a credible academic voice in the AI criticism and evaluation space. The framing targets the research community directly, making it relevant to how practitioners and researchers are thinking about the trajectory of the field.
AREA is a new method for CLIP-based Class-Incremental Learning (CIL) that decomposes the classification process into attribute extraction and aggregation stages to combat catastrophic forgetting. Extraction is stabilized by anchoring visual and textual attributes on a hyperspherical embedding space via principal geodesic analysis, while aggregation uses lightweight task-specific experts regularized by a variational information bottleneck. Inference employs optimal transport routing over task attribute manifolds. The method is reported to consistently outperform state-of-the-art CIL approaches and is accepted at ICML 2026.