Research map / Vision-language and multimodal / Visual question answering
MyVLM: Personalizing VLMs for User-Specific Queries
Yuval Alaluf, Elad Richardson, Sergey Tulyakov, Kfir Aberman, Danny Cohen-Or
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add regularization and augmentations to concept embedding training (ablation) | MyVLM · MyVLM dataset subset (10 concepts) | Recall (%): 25.88 -> 84.87 | Improved |
| add image and caption augmentations (ablation) | MyVLM · MyVLM dataset subset (10 concepts) | Recall (%): 72.77 -> 84.87 | Improved |
| increase training images from 1 to 4 (ablation) | MyVLM on BLIP-2 · MyVLM dataset | Recall (%): 75.42 -> 87.11 | Improved |
| increase training images from 1 to 4 (ablation) | MyVLM on LLaVA · MyVLM dataset | Recall (%): 88.93 -> 95.97 | Improved |
| learn a concept embedding with concept heads instead of LLM-guided keyword replacement | LLaVA captioning (all) · MyVLM dataset | Recall (%): 46.23 -> 95.97 | Improved |
| learn a concept embedding instead of keyword replacement for people on BLIP-2 | BLIP-2 captioning (people) · MyVLM dataset | Recall (%): 84.33 -> 79.76 | No clear effect |
| learn a concept embedding instead of in-context prompting (OpenFlamingo) for objects | personalized captioning (objects) · MyVLM dataset | Recall (%): 49.77 -> 95.10 | Improved |
| apply MyVLM personalization on BLIP-2 captioning | BLIP-2 · Standard captioning metrics (all) | CIDEr: 1.89 -> 1.28 | Hurt |
| apply MyVLM personalization on LLaVA captioning | LLaVA · Standard captioning metrics (all) | CIDEr: 0.17 -> 0.58 | Improved |
Later papers that compared with this method
2 later accepted papers report MyVLM as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Training-Free Personalization via Retrieval and Reasoning on Fingerprints (ICCV 2025 · better on 2 of 2 shared results)
- Yo'LLaVA: Your Personalized Language and Vision Assistant (NeurIPS 2024 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- BLIP-2 — BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation (ICML 2022) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- RAP: Retrieval-Augmented Personalization for Multimodal Large Language Models (CVPR 2025 · similarity 0.90)
- Contextualized Visual Personalization in Vision-Language Models (ICML 2026 · similarity 0.88)
- MMPB: It’s Time for Multi-Modal Personalization (NeurIPS 2025 · similarity 0.88)
- Personalized Visual Instruction Tuning (ICLR 2025 · similarity 0.87)
- MM-R1: Unleashing the Power of Unified Multimodal Large Language Models for Personalized Image Generation (AAAI 2026 · similarity 0.87)
- MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMs (ICLR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
