AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

QPKO: Differentiable QP-Embedded Deep Koopman Framework for Modeling Nonlinear Systems

Runze Tian, Peng Kou

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Differentiable optimization layers

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
use QP-based mapping to compute the global linear model instead of the second-best Koopman frameworkdeep Koopman model · TSR (T=150, T=200)NMAE: 46.9% and 41.1% reductionImproved
use QP-based mapping instead of the second-best frameworkdeep Koopman model · CartPole (T=150, T=200)NMAE: 26.5%, 27.9% reductionImproved
use QP-based mapping instead of the second-best frameworkdeep Koopman model · VDP (T=150, T=200)NMAE: 22.7%, 13.3% reductionImproved
use QP-based mapping instead of the second-best frameworkdeep Koopman model · WT (T=150, T=200)NMAE: 7.9%, 9.3% reductionImproved
use end-to-end QP-based mapping instead of the second-best frameworkdeep Koopman model · CartPole training timewall-clock time: 3.1x speedupImproved
use end-to-end QP-based mapping instead of the second-best frameworkdeep Koopman model · VDP training timewall-clock time: 2.6x speedupImproved
use end-to-end QP-based mapping instead of the second-best frameworkdeep Koopman model · TSR training timewall-clock time: 1.9x speedupImproved
use QP-based mapping instead of alternating LS training (EDMD-DL)deep Koopman model · WT training timewall-clock time: 0.23±0.27 -> 0.56±0.02Hurt
use QP-based mapping instead of second-best baseline in MPCKoopman MPC · CartPolefinal setpoint error: 61.8% reductionImproved
use QP-based mapping instead of second-best baseline in MPCKoopman MPC · VDPfinal setpoint error: 27.3% reductionImproved
use QP-based mapping instead of second-best baseline in MPCKoopman MPC · TSRfinal setpoint error: 36.4% reductionImproved
use QP-based mapping instead of LS, end-to-end, or independent linear model in MPCKoopman MPC · TSRcontrol success: baselines fail; QPKO 0.51±0.39Improved

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