AI Paper Atlas

Research map / Video / Video generation

Past as State, Present as Attention: Persistent-State Blockwise Flow Matching for Long-Horizon Co-Speech Motion Generation

Yi Liu, Xiangyue Zhang, Jia Ma, Jianfang Li, Yisheng He, Jianqiang Ren

NeurIPS 2026 · Video › Video generation · approach: Autoregressive long video generation · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a persistent-state blockwise flow matching model for long-horizon co-speech motion generation, carrying a compact persistent state across segments instead of short seeds or ever-longer explicit histories.

Problem

Long-horizon co-speech motion generation suffers from temporal drift when conditioning on short seed frames, and from high cost and poor learnability when retaining long explicit motion histories.

Approach

Separates the past as a persistent state carried across blocks from the present handled by attention within each block, generating motion blockwise with flow matching; the abstract only partially describes this.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add persistent-state history (State past, Attn current) instead of no history (ablation)PASPA · BEAT2 1 SpeakerFGD: 7.065 -> 5.216Improved
Use persistent state for past instead of full token-attention history (ablation)PASPA · BEAT2 1 SpeakerFGD: 5.710 -> 5.216Improved
Use attention for current block instead of state-based modeling (ablation)PASPA · BEAT2 1 SpeakerFGD: 7.652 -> 5.216Improved
Switch from single-block to multi-block supervision naively (ablation)PASPA · BEAT2 1 SpeakerFGD: 7.410 -> 8.456Hurt
Add weakly monotonic timesteps (ablation)PASPA multi-block · BEAT2 1 SpeakerFGD: 8.456 -> 6.346Improved
Add adaptive loss weighting (ablation)PASPA multi-block · BEAT2 1 SpeakerFGD: 6.346 -> 5.614Improved
Add stop-gradient (SG) (ablation)PASPA · BEAT2 1 SpeakerFGD: 5.614 -> 5.216Improved
Add logit-normal time sampling (ablation)PASPA · BEAT2 1 SpeakerFGD: 5.216 -> 4.450Improved
Add condition dropout during training (ablation)PASPA · BEAT2 1 SpeakerFGD: 4.450 -> 3.937Improved
Add history dropout without guidance (ablation)PASPA · BEAT2 1 SpeakerFGDHurt
Use full Hybrid CFG instead of condition-only CFG (ablation)PASPA · BEAT2 1 SpeakerFGD: 3.883 -> 3.532Improved
Add prefill-decode separation (ablation)PASPA · synthetic ultra-long rolloutslatencyImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

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.

Experimental setup

Builds on: flow matching

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