Skip to content

What Changed? Drift Detection with Real, Virtual, and Incomparable Diagnosis

Aug 2026 · 0 citations · 6 references
Computer Science

TL;DR

It is shown that cross-evaluated heads on a frozen shared representation inherit the extrapolation confound of shallow exchange scores: pure input rotations with fixed labels inflate a deep exchange score from about 0 to 0.80, while representation-novelty scores are blind in the complementary direction.

Abstract

Sharing a deep encoder does not, by itself, fix the central confound of task-comparison scores. We show that cross-evaluated heads on a frozen shared representation inherit the extrapolation confound of shallow exchange scores: pure input rotations with fixed labels inflate a deep exchange score from about 0 to 0.80, while representation-novelty scores are blind in the complementary direction (flat under label permutations that change the task completely). Transplanting a conditional two-discriminator discrepancy into the embedding space resolves both blind spots: the functional axis stays within +-0.001 under rotations and tracks label-permutation drift mass monotonically. Built into a mixture-of-heads lifecycle, the two-axis gate attains better decision quality with fewer heads than exchange or novelty triggers at a matched training budget. On generalized category discovery, the same chunk-level functional axis separates semantic novelty from photometric shift with AUROC 0.98-0.99 where per-input OOD scores (MSP, Energy, Mahalanobis, KNN) sit near chance for that distinction. All findings replicate across frozen ImageNet-21k ViT-B/16 and self-supervised DINOv2 backbones on CIFAR-100, and extend to residual adapter pools with recurrence, where a null-calibrated novelty trigger never fires on mechanism changes while the two-axis gate handles them with full recurrence reuse. We state explicitly the common-factoring condition under which embedding-space conclusions transfer to the original mechanism.

View source

Similar papers

#artificial intelligence Preprint Aug 2026

A Shared Encoder Is Not a Shared Task: Conditional Comparison for Deep Expert Pools

It is shown that cross-evaluated heads on a frozen shared representation inherit the extrapolation confound of shallow exchange scores: pure input rotations with fixed labels inflate a deep exchange score from about 0 to 0.80, while representation-novelty scores are blind in the complementary direction.

Kentaro Oda · 0 citations
#artificial intelligence Preprint Sep 2026

What Do Scan-Derived Class Prototypes Add? Disentangling Supervision, Prototype Content and Query Protocol in Recognition over Frozen Foundation Features

Across DINOv2-giant and MetaCLIP-H with real-background queries, the largest fused-accuracy advantage of the real prototypes over either control is one percentage point; larger differences favor controls, by up to 2.8 points in arm means.

Chen-Xi Tao, Hong-in Won, Seung-Kyum Choi · 1 citation
Preprint Aug 2026

DiD It in 87 Minutes: A Label-Free Softmax-to-Linear Adaptation of Vision Transformers for Object Detection

DiD is introduced, a label-free conversion method that exclusively trains the linear-attention backbone by aligning detector-facing interface tensors with those of a frozen Softmax teacher, and substantially outperforms established baselines and matches supervised, fully trained linear models.

Huai-Yuan Qin, Gabriel James Goenawan, Zihang Lin et al. · 1 citation
Preprint Sep 2026

Reassessing Global Gradient-Norm Imbalance in BLIP Fine-Tuning Across Physical Domains

Imbalanced gradient magnitudes between the visual and language pathways of a vision-language model are often treated as a defect to be corrected. We test that premise for one family of correction, deliberately excluding adaptive, signal-driven schemes (e.g. BalGrad, OGM, PMR, CGGM), which are a mechanistically distinct...

Kiran Naseer, Samreen Azhar, Dwarikanath Mahapatra · 0 citations
Open access Aug 2026

Label-Free Visual Concept Drift Detection via Classifier Two-Sample Tests and Characteristic Function Embeddings

This work proposes a real-time, label-free concept drift detection framework that couples the Classifier Two-Sample Test (C2ST) with a novel Characteristic Function Embedding (CFE) bottleneck, operating directly on deep convolutional representations.

Gurgen Hovakimyan, Jorge Miguel Bravo · 0 citations

Which Tasks Survive Self-Supervised Learning?

Same-instance self-supervised learning (SSL) learns representations by enforcing consistency across two views of the same underlying instance. This principle alone, however, does not determine which downstream tasks remain recoverable from the learned representation. We study this question through \emph{semantic recove...

Achleshwar Luthra, Lucas Bryant, T. Zhu et al. · 0 citations

Related blog posts

MIT News · Artificial Intelligence Sep 29, 2026

Who we become when we talk to machines

Professor Sherry Turkle’s new book, “Artificial Intimacy,” offers a withering critique of chatbots and the antisocial dynamics she believes they encourage.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.