AI Paper Atlas

Research map / Video / Video generation

LOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation

Jiarui Wang, Huiyu Duan, Ziheng Jia, Zicheng Zhang, Yu Zhao, Juntong Wang, Guangtao Zhai, Xiongkuo Min

ICML 2026 · Video › Video generation · approach: Compositional text-to-video diffusion · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
add temporal features (full LOVE vs. row 5 variant) (ablation)LOVE · AIGVE-60K perceptionSRCC: 0.734 -> 0.793Improved
add text-defined quality-level initialization (ablation)LOVE · AIGVE-60K perceptionSRCC: 0.642 -> 0.734Improved
add LoRA fine-tuning on vision encoder and LLM (ablation)LOVE · AIGVE-60K QAaccuracy: 65.8% -> 77.3%Improved
fine-tune LMM on AIGVE-60K instead of zero-shot (InternVL3 9B)InternVL3-9B · AIGVE-60K perception instance-levelSRCC: 0.2731 -> 0.6421Improved
use LOVE reward (perception + correspondence) vs. perception only (ablation)Wan2.1 GRPO · HPSv2 evaluationHPSv2: 1.6132 -> 1.6250Improved
use LOVE-Reward with GRPO on Wan2.1Wan2.1 · LOVE correspondence evaluationLOVE correspondence score: 0.5175 -> 0.5268Improved
use CLIPScore as GRPO reward instead of LOVEWan2.1 GRPO · HPSv2 evaluationHPSv2: 1.5576 -> 1.5576No clear effect
use HPSv2 or ImageReward as GRPO reward, evaluated on LOVE perceptionWan2.1 GRPO · LOVE perception evaluationLOVE perception score: 0.5118 -> 0.5091 (HPSv2+GRPO)Hurt
train on AIGVE-60K and test zero-shot on other datasets vs. VQAScore trained on GenAI-BenchLOVE · GenAI-BenchSRCC: 0.632 -> 0.682Improved
use an LMM-based evaluator vs. handcrafted metricsAIGV evaluation · AIGVE-60K T2V correspondence (model-level)SRCC: 0.7321 (BMPRI) -> 0.9778 (LOVE)Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy