{"type":"ScholarlyArticle","title":"Semantic Geometric Calibration in Randomized Neural Feature Space","authors":[{"name":"Itay Abuhazera","slug":"itay-abuhazera","given":"Itay","family":"Abuhazera","affiliation":""},{"name":"Liron Cohen","slug":"liron-cohen","given":"Liron","family":"Cohen","affiliation":""},{"name":"Gil Einziger","slug":"gil-einziger","given":"Gil","family":"Einziger","affiliation":""}],"abstract":"Modern neural networks can be accurate yet poorly calibrated. We present Random Geometric Calibration (RGC), a post hoc method that augments confidence estimation with a geometric signal derived from distances to training data in an aggregated semantic representation space. Prior geometric calibration methods in feature space often rely on architecture-specific layer choices, which can be unstable across models and datasets. To avoid this dependence, RGC samples a fixed number of intermediate blocks uniformly at random and reuses this set at both calibration and test time.\r\nWe instantiate RGC with RGCL, which applies spatial pyramid pooling to each sampled activation, projects features to a fixed dimension, and computes a nearest-neighbor separation score that is mapped to probabilities via isotonic regression.\r\nWe also present RGCC, which replaces pooling and projection with uniform coordinate sampling for improved computational efficiency.\r\nWe evaluate RGCL and RGCC across multiple datasets, including CIFAR-10, CIFAR-100, and Tiny-ImageNet, and across a diverse set of architectures such as ResNet, DenseNet, and DINOv2. Our results show that RGCL achieves state-of-the-art or near state-of-the-art calibration performance, reducing expected calibration error by approximately 78% on average and by up to 97% for some models.\r\nFinally, we provide representation analysis that explains why RGCL and RGCC succeed without layer selection and characterize their computational overheads. We observe that layers exhibiting favorable geometric structure tend to emerge in later network blocks, and that randomized aggregation implicitly emphasizes these layers, leading to robust and reliable calibration performance.","doi":"10.24963/ijcai.2026/458","url":"https://ijcai.org/proceedings/2026/458","pdf":"https://ijcai.org/proceedings/2026/0458.pdf","markdown":"https://ijcai.org/proceedings/2026/458.md","bibtex":"https://ijcai.org/proceedings/2026/bibtex/458","volume":{"slug":"2026","title":"Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2026","url":"https://ijcai.org/proceedings/2026/","year":2026,"location":"Bremen, Germany","dates":"15-21 August 2026","editors":"Diego Calvanese","isbn":"978-1-956792-09-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2026"},"section":"Machine Learning","track":"Main Track","number":458,"pages":{"first":4116,"last":4124},"keywords":["Machine Learning: Geometric learning","Uncertainty in AI: Uncertainty representations","Machine Learning: Classification","Machine Learning: Evaluation","Machine Learning: Trustworthy machine learning"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}