AI & ML
DEFTS: Data-Efficient, Training-Free Few-Shot Learning for CLIP in the Text Subspace
arXiv stat.ML0 views
arXiv:2611.21993v1 Announce Type: cross Abstract: Contrastive Language-Image Pre-training (CLIP) achieves impressive zero-shot image classification performance by comparing image embeddings against class-specific text embeddings. While recent few-shot methods leverage labeled training samples to improve upon this zero-shot baseline, they often predominantly rely on image embeddings and underutilize the semantic structure of the text embeddings. In this paper, we introduce DEFTS, a data-efficient, training-free few-shot classification method strictly in the text subspace. We first demonstrate empirically that while raw text embeddings can be suboptimal as direct classifiers, the linear subspace they span remains highly discriminative. DEFTS therefore proposes to reduce the image-text modality gap by projecting estimated image prototypes directly into this subspace. Importantly, by expressing these prototypes as linear combinations of text embeddings, DEFTS reduces a high-dimensional estimation problem to the derivation of just a few coefficients. These coefficients are then computed via a Bayesian posterior mean, using a prior naturally centered on the original text embedding coordinates. Extensive experiments across 12 image classification benchmarks demonstrate that DEFTS significantly improves upon existing training-free methods.
Source:https://arxiv.org/abs/2611.21993PreviousCancellation-Free Constrained SquareCB for Safe Controller SelectionNextLocally Linear Generalized Single Index Regression
Related posts
Kumo Tabular: How NVIDIA Built a Foundation Model That Predicts on Any Table Without Training
0dev.to

Asus TUF Gaming F16 (2026), A15 & T700 Mini: Specs, Pricing & The ₹25,000 GPU Catch
0dev.to
I Stopped Chasing Every New AI Tool — And My Blog Finally Started Growing
0dev.to

200 OK is dead as an AI agent health signal. Here’s what comes next.
0The New Stack

AI needs fresh data. Vector indexing is getting in the way
0The New Stack
Electron fingerprints help AI predict complex alloy properties with little training data
0Phys.org Tech