Saved in:
Bibliographic Details
Main Authors: Leach, William, He, Ru, Ma, Sizhuo, Jia, Yizhen, Cao, Min, Wang, Jian, Cao, Rick
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.27737
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914606721531904
author Leach, William
He, Ru
Ma, Sizhuo
Jia, Yizhen
Cao, Min
Wang, Jian
Cao, Rick
author_facet Leach, William
He, Ru
Ma, Sizhuo
Jia, Yizhen
Cao, Min
Wang, Jian
Cao, Rick
contents Vision-language models (VLMs) are increasingly attractive for multimodal quality assessment, but their default reliance on autoregressive text generation and dynamic visual processing is poorly matched to scalar regression under strict latency budgets. We present a bounded-compute adaptation of SmolVLM2-256M-Video-Instruct for product-rating prediction in the LoViF 2026 Efficient VLM challenge. Motivated by recent multimodal engagement-prediction results showing that feature-based regression can outperform token-based score generation, we replace the language-modeling head with a lightweight two-layer MLP fed by pooled decoder states, and we enforce deterministic inputs through fixed 384x384 images and truncated metadata. Across controlled ablations, static global image processing slightly outperforms dynamic tiling, and scaling from 100K to 16M training examples substantially improves validation correlation. Under the official held-out evaluation, our 228M-parameter model achieves 0.39 PLCC and 0.40 CES, providing a strong and reproducible baseline for resource-constrained multimodal regression.
format Preprint
id arxiv_https___arxiv_org_abs_2605_27737
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Bounded-Compute Multimodal Regression for Product-Rating Prediction
Leach, William
He, Ru
Ma, Sizhuo
Jia, Yizhen
Cao, Min
Wang, Jian
Cao, Rick
Computer Vision and Pattern Recognition
Vision-language models (VLMs) are increasingly attractive for multimodal quality assessment, but their default reliance on autoregressive text generation and dynamic visual processing is poorly matched to scalar regression under strict latency budgets. We present a bounded-compute adaptation of SmolVLM2-256M-Video-Instruct for product-rating prediction in the LoViF 2026 Efficient VLM challenge. Motivated by recent multimodal engagement-prediction results showing that feature-based regression can outperform token-based score generation, we replace the language-modeling head with a lightweight two-layer MLP fed by pooled decoder states, and we enforce deterministic inputs through fixed 384x384 images and truncated metadata. Across controlled ablations, static global image processing slightly outperforms dynamic tiling, and scaling from 100K to 16M training examples substantially improves validation correlation. Under the official held-out evaluation, our 228M-parameter model achieves 0.39 PLCC and 0.40 CES, providing a strong and reproducible baseline for resource-constrained multimodal regression.
title Bounded-Compute Multimodal Regression for Product-Rating Prediction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.27737