Low-Rank Prompt Learning for Vision-Language Models with Fixed-Token Bases

• URL: http://arxiv.org/abs/2609.09462v1
• Date: Tue, 08 Sep 2026 21:26:37 GMT
• Title: Low-Rank Prompt Learning for Vision-Language Models with Fixed-Token Bases
• Abstract summary: We study a dense prompt matrix $mathbfPinmathbbRmtimes d$ trained from only a few examples per class.<n>We find that the token-side factor need not be learned at all.<n>In the CLIP prompt setting, the embedding-side coefficients carry the adaptation while the token basis can simply be fixed.
• Score: 11.437689514839333
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Prompt learning adapts CLIP to downstream recognition by replacing hand-written templates with learned continuous context vectors, which in Context Optimization (CoOp) form a dense prompt matrix $\mathbf{P}\in\mathbb{R}^{m\times d}$ trained from only a few examples per class. We study whether this matrix is over-parameterized by factorizing it as $\mathbf{P}=\mathbf{B}\mathbf{A}$, which cuts the trainable prompt parameters from $md$ to $r(m+d)$, and to $rd$ once the token-side factor $\mathbf{B}$ is fixed. Across seven few-shot benchmarks and two CLIP backbones, low-rank prompts match or improve dense CoOp at far fewer parameters, with the clearest gains on low-shot base-to-new generalization. We then find that the token-side factor need not be learned at all: fixing $\mathbf{B}$ to a Gaussian, orthogonal, SVD-derived, or even random basis and training only the embedding-side factor $\mathbf{A}$ stays on par with the fully trainable factorization, and a source-trained $\mathbf{B}$ offers no advantage over a random one. A prompt-factor asymmetry and a local update-space dimension gap show why fixing $\mathbf{B}$ is far less restrictive than fixing $\mathbf{A}$, and a smoothness-only guarantee certifies that optimizing $\mathbf{A}$ over a fixed $\mathbf{B}$ converges. In the CLIP prompt setting, the embedding-side coefficients carry the adaptation while the token basis can simply be fixed.

Related papers

• Robust Learning of a Group DRO Neuron [21.632698901872843]
  We study the problem of learning a primal neuron under standard squared loss in the presence of arbitrary label noise and group-level distributional shifts.<n>Our framework provides robust learning guarantees in the face of arbitrary label corruptions and group-specific distributional shifts.
  arXiv  Detail & Related papers  (2026-01-26T04:00:53Z)
• Sample and Computationally Efficient Robust Learning of Gaussian Single-Index Models [37.42736399673992]
  A single-index model (SIM) is a function of the form $sigma(mathbfwast cdot mathbfx)$, where $sigma: mathbbR to mathbbR$ is a known link function and $mathbfwast$ is a hidden unit vector. We show that a proper learner attains $L2$-error of $O(mathrmOPT)+epsilon$, where $
  arXiv  Detail & Related papers  (2024-11-08T17:10:38Z)
• Compressing Large Language Models using Low Rank and Low Precision Decomposition [46.30918750022739]
  This work introduces $rm CALDERA$ -- a new post-training LLM compression algorithm. It harnesses the inherent low-rank structure of a weight matrix $mathbfW$ by approximating it via a low-rank, low-precision decomposition. Results show that compressing LlaMa-$2$ $7$B/$13B$/$70$B and LlaMa-$3$ $8$B models using $rm CALDERA$ outperforms existing post-training compression techniques.
  arXiv  Detail & Related papers  (2024-05-29T08:42:30Z)
• Provably learning a multi-head attention layer [55.2904547651831]
  Multi-head attention layer is one of the key components of the transformer architecture that sets it apart from traditional feed-forward models. In this work, we initiate the study of provably learning a multi-head attention layer from random examples. We prove computational lower bounds showing that in the worst case, exponential dependence on $m$ is unavoidable.
  arXiv  Detail & Related papers  (2024-02-06T15:39:09Z)
• Statistical Learning under Heterogeneous Distribution Shift [71.8393170225794]
  Ground-truth predictor is additive $mathbbE[mathbfz mid mathbfx,mathbfy] = f_star(mathbfx) +g_star(mathbfy)$.
  arXiv  Detail & Related papers  (2023-02-27T16:34:21Z)
• Learning a Single Neuron with Adversarial Label Noise via Gradient Descent [50.659479930171585]
  We study a function of the form $mathbfxmapstosigma(mathbfwcdotmathbfx)$ for monotone activations. The goal of the learner is to output a hypothesis vector $mathbfw$ that $F(mathbbw)=C, epsilon$ with high probability.
  arXiv  Detail & Related papers  (2022-06-17T17:55:43Z)
• Model Selection with Near Optimal Rates for Reinforcement Learning with General Model Classes [27.361399036211694]
  We address the problem of model selection for the finite horizon episodic Reinforcement Learning (RL) problem. In the model selection framework, instead of $mathcalP*$, we are given $M$ nested families of transition kernels. We show that textttARL-GEN obtains a regret of $TildemathcalO(d_mathcalE*H2+sqrtd_mathcalE* mathbbM* H2 T)$
  arXiv  Detail & Related papers  (2021-07-13T05:00:38Z)
• Agnostic Learning of a Single Neuron with Gradient Descent [92.7662890047311]
  We consider the problem of learning the best-fitting single neuron as measured by the expected square loss. For the ReLU activation, our population risk guarantee is $O(mathsfOPT1/2)+epsilon$. For the ReLU activation, our population risk guarantee is $O(mathsfOPT1/2)+epsilon$.
  arXiv  Detail & Related papers  (2020-05-29T07:20:35Z)
• Taking a hint: How to leverage loss predictors in contextual bandits? [63.546913998407405]
  We study learning in contextual bandits with the help of loss predictors. We show that the optimal regret is $mathcalO(minsqrtT, sqrtmathcalETfrac13)$ when $mathcalE$ is known.
  arXiv  Detail & Related papers  (2020-03-04T07:36:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.