# SLAP: The Semantic Least Action Principle for Variational Video-Language Modeling

Source: [SLAP: The Semantic Least Action Principle for Variational Video-Language Modeling](<https://www.sota2.com/research/paper/6a1d1214dbd7150506f84010>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-06-02T05:00:05.014Z

Authors: Xiang Fang, Wanlong Fang

Published: 2026-05-29T02:28:33.000Z

## Abstract

In the era of Large Video-Language Models (LVLMs), the computational necessity of sparse frame sampling creates a fundamental \`\`temporal gap'', rendering models blind to critical causal transitions. Existing solutions relying on generative hallucination (e.g., latent diffusion) or autoregressive extrapolation often fail to maintain semantic consistency over long horizons, suffering from object vanishing and energetic instability. We propose a paradigm shift from probabilistic generation to variational mechanics with the \\textbf{Semantic Least Action Principle (SLAP)}. Drawing a rigorous isomorphism between classical mechanics and semantic dynamics, we model the latent video trajectory as a path on a Riemannian manifold governed by a Semantic Lagrangian. By formulating the interpolation task as a Boundary Value Problem (BVP) solved via the discrete Euler-Lagrange equations, SLAP naturally enforces object persistence without pixel-level rendering. Extensive experiments show the effectiveness of our proposed SLAP.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Object Persistence on Tunnel (test)](<https://www.sota2.com/research/sota/object-persistence-on-tunnel-test.md>): Object Persistence; Tunnel (test); Accuracy: 83.9.

- [Video Question Answering on MSR-VTT (Frame Sampling Accuracy)](<https://www.sota2.com/research/sota/video-question-answering-on-msr-vtt-frame-sampling-accuracy.md>): Video Question Answering; MSR-VTT; Accuracy (50% Frames): 45.2.

- [Video Interpolation on Tunnel (val)](<https://www.sota2.com/research/sota/video-interpolation-on-tunnel-val.md>): Video Interpolation; Tunnel (val); FLOPs (T): 0.15.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2605.30750>)

[PDF](<https://arxiv.org/pdf/2605.30750.pdf>)