AI Paper Atlas

Research map / Probabilistic and scientific modeling / Dynamical systems

ModeRNN: Harnessing Spatiotemporal Mode Collapse in Unsupervised Predictive Learning

Zhiyu Yao, Yunbo Wang, Haixu Wu, Jianmin Wang, Mingsheng Long

TPAMI 2023 · Probabilistic and scientific modeling › Dynamical systems · approach: Low-rank and diffusion neural dynamics · 18 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
train on entire mixed dataset instead of single-mode subsetConvLSTM · Mixed Moving MNIST Mode-1MSE: 26.6 -> 33.4Hurt
train on entire mixed dataset instead of single-mode subsetConvLSTM · Mixed Moving MNIST Mode-3MSE: 120.5 -> 137.1Hurt
train on entire mixed dataset instead of single-mode subsetModeRNN · Mixed Moving MNIST Mode-1MSE: 21.9 -> 17.1Improved
train on entire mixed dataset instead of single-mode subsetModeRNN · Mixed Moving MNIST Mode-3MSE: 83.1 -> 74.8Improved
train on entire mixed dataset instead of subsetRIM · Mixed Moving MNIST Mode-2MSE: 53.3 -> 52.1Mixed
increase spatiotemporal slots from 1 to 4 (ablation)ModeRNN · Mixed Moving MNISTper-frame MSE: 69.3 -> 44.7Improved
increase slots from 4 to 5 (ablation)ModeRNN · Mixed Moving MNISTper-frame MSE: 44.7 -> 45.9Hurt
add adaptive slot fusion instead of equal-weight sum (ablation)ModeRNN · Mixed Moving MNISTper-frame MSE: 60.5 -> 44.7Improved
add slot binding (multiple slots instead of N=1) (ablation)ModeRNN · Mixed Moving MNISTper-frame MSE: 69.3 -> 44.7Improved
replace baseline predictors with ModeRNNModeRNN vs SVG · KTH actionPSNR: 27.73 -> 28.22Improved
replace TrajGRU with ModeRNNModeRNN · Radar echoCSI30: 0.357 -> 0.428Improved
train on entire KTH dataset instead of subsetsPredRNN · KTH Mode-1 (torso movement)PSNR: 27.45 -> 26.92Hurt

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