Copertina

Pruning the Paradox: How CLIP's Most Informative Heads Enhance Performance While Amplifying Bias

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Madasu, Avinash, Lal, Vasudev, Howard, Phillip
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Computer Vision and Pattern Recognition Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Quantifying and Enabling the Interpretability of CLIP-like Models
di: Madasu, Avinash, et al.
Pubblicazione: (2024)

Cultural Awareness in Vision-Language Models: A Cross-Country Exploration
di: Madasu, Avinash, et al.
Pubblicazione: (2025)

ICSVR: Investigating Compositional and Syntactic Understanding in Video Retrieval Models
di: Madasu, Avinash, et al.
Pubblicazione: (2023)

SocialCounterfactuals: Probing and Mitigating Intersectional Social Biases in Vision-Language Models with Counterfactual Examples
di: Howard, Phillip, et al.
Pubblicazione: (2023)

Training-Free Mitigation of Language Reasoning Degradation After Multimodal Instruction Tuning
di: Ratzlaff, Neale, et al.
Pubblicazione: (2024)

Is Your Paper Being Reviewed by an LLM? Benchmarking AI Text Detection in Peer Review
di: Yu, Sungduk, et al.
Pubblicazione: (2025)

Is Your Paper Being Reviewed by an LLM? Investigating AI Text Detectability in Peer Review
di: Yu, Sungduk, et al.
Pubblicazione: (2024)

DPO Learning with LLMs-Judge Signal for Computer Use Agents
di: Luo, Man, et al.
Pubblicazione: (2025)

Locating Demographic Bias at the Attention-Head Level in CLIP's Vision Encoder
di: Yasser, Alaa, et al.
Pubblicazione: (2026)

Uncovering Bias in Large Vision-Language Models with Counterfactuals
di: Howard, Phillip, et al.
Pubblicazione: (2024)

LVLM-Compress-Bench: Benchmarking the Broader Impact of Large Vision-Language Model Compression
di: Kundu, Souvik, et al.
Pubblicazione: (2025)

Decoupling Template Bias in CLIP: Harnessing Empty Prompts for Enhanced Few-Shot Learning
di: Zhang, Zhenyu, et al.
Pubblicazione: (2025)

Debiasing Classifiers by Amplifying Bias with Latent Diffusion and Large Language Models
di: Ko, Donggeun, et al.
Pubblicazione: (2024)

Learning from Reasoning Failures via Synthetic Data Generation
di: Stan, Gabriela Ben Melech, et al.
Pubblicazione: (2025)

Enhanced Structured Lasso Pruning with Class-wise Information
di: Liu, Xiang, et al.
Pubblicazione: (2025)

Geometry-Aware CLIP Retrieval via Local Cross-Modal Alignment and Steering
di: Prakash, Nirmalendu, et al.
Pubblicazione: (2026)

MoPE-CLIP: Structured Pruning for Efficient Vision-Language Models with Module-wise Pruning Error Metric
di: Lin, Haokun, et al.
Pubblicazione: (2024)

Automatic Channel Pruning for Multi-Head Attention
di: Lee, Eunho, et al.
Pubblicazione: (2024)

AA-CLIP: Enhancing Zero-shot Anomaly Detection via Anomaly-Aware CLIP
di: Ma, Wenxin, et al.
Pubblicazione: (2025)

CLIP-MUSED: CLIP-Guided Multi-Subject Visual Neural Information Semantic Decoding
di: Zhou, Qiongyi, et al.
Pubblicazione: (2024)

CLIP Tricks You: Training-free Token Pruning for Efficient Pixel Grounding in Large VIsion-Language Models
di: Lee, Sangin, et al.
Pubblicazione: (2026)

Omni-NegCLIP: Enhancing CLIP with Front-Layer Contrastive Fine-Tuning for Comprehensive Negation Understanding
di: Xu, Jingqi
Pubblicazione: (2026)

Data Pruning by Information Maximization
di: Tan, Haoru, et al.
Pubblicazione: (2025)

SafeR-CLIP: Mitigating NSFW Content in Vision-Language Models While Preserving Pre-Trained Knowledge
di: Yousaf, Adeel, et al.
Pubblicazione: (2025)

FiVL: A Framework for Improved Vision-Language Alignment through the Lens of Training, Evaluation and Explainability
di: Aflalo, Estelle, et al.
Pubblicazione: (2024)

Grounding-Aware Token Pruning: Recovering from Drastic Performance Drops in Visual Grounding Caused by Pruning
di: Chien, Tzu-Chun, et al.
Pubblicazione: (2025)

PLPHP: Per-Layer Per-Head Vision Token Pruning for Efficient Large Vision-Language Models
di: Meng, Yu, et al.
Pubblicazione: (2025)

One Head Eight Arms: Block Matrix based Low Rank Adaptation for CLIP-based Few-Shot Learning
di: Zhou, Chunpeng, et al.
Pubblicazione: (2025)

Enhancing Multimodal Understanding with CLIP-Based Image-to-Text Transformation
di: Che, Chang, et al.
Pubblicazione: (2024)

Synergy and Diversity in CLIP: Enhancing Performance Through Adaptive Backbone Ensembling
di: Rodriguez-Opazo, Cristian, et al.
Pubblicazione: (2024)

How Bias Binds: Measuring Hidden Associations for Bias Control in Text-to-Image Compositions
di: Li, Jeng-Lin, et al.
Pubblicazione: (2025)

Enhancing Compositional Reasoning in CLIP via Reconstruction and Alignment of Text Descriptions
di: Kwon, Jihoon, et al.
Pubblicazione: (2025)

InterCLIP-MEP: Interactive CLIP and Memory-Enhanced Predictor for Multi-modal Sarcasm Detection
di: Chen, Junjie, et al.
Pubblicazione: (2024)

Cross-Cultural Value Awareness in Large Vision-Language Models
di: Howard, Phillip, et al.
Pubblicazione: (2026)

AdaptCLIP: Adapting CLIP for Universal Visual Anomaly Detection
di: Gao, Bin-Bin, et al.
Pubblicazione: (2025)

DesignCLIP: Multimodal Learning with CLIP for Design Patent Understanding
di: Wang, Zhu, et al.
Pubblicazione: (2025)

Semantic Relation-Enhanced CLIP Adapter for Domain Adaptive Zero-Shot Learning
di: Yu, Jiaao, et al.
Pubblicazione: (2025)

Enhancing Layer Attention Efficiency through Pruning Redundant Retrievals
di: Li, Hanze, et al.
Pubblicazione: (2025)

UNSEEN: Enhancing Dataset Pruning from a Generalization Perspective
di: Xu, Furui, et al.
Pubblicazione: (2025)

The Reasoning Boundary Paradox: How Reinforcement Learning Constrains Language Models
di: Nguyen, Phuc Minh, et al.
Pubblicazione: (2025)