AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Boosting Verification of Deep Reinforcement Learning via Piece-Wise Linear Decision Neural Networks

Jiaxu Tian, Dapeng Zhi, Si Liu, Peixin Wang, Cheng Chen, Min Zhang

NeurIPS 2023 · Reinforcement learning › Visual and representation RL · approach: Scalable RL systems and environments · 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
use linear policy per partition instead of constant action (ablation)PLDNN · B1number of partitions: 4 -> 1Improved
use linear policy per partition instead of constant action (ablation)PLDNN · B2number of partitions: 100 -> 4Improved
use linear policy per partition instead of constant action (ablation)PLDNN · CartPolenumber of partitions: 254 -> 16Improved
use linear policy per partition instead of constant action (ablation)PLDNN · QUADnumber of partitions: 412 -> 1Improved
use PLDNN instead of standard DNN controllerDNN policy · B1training time (s): 11.0 -> 14.3Hurt
use PLDNN instead of standard DNN controllerDNN policy · QUADtraining time (s): 781.5 -> 871.1Hurt
use PLDNN instead of standard DNN controllerDNN policy · B1, B2, CartPole, QUADcumulative rewardNo clear effect
use PLDNN instead of standard DNN controller under Gaussian state noiseDNN policy · B1, B2, CartPole, QUADrobustness (reward decline)No clear effect
verify with LinCon on PLDNN instead of Polar on DNNPolar · Table 2 benchmarksverification speedup: up to 47.7×Improved
verify with LinCon on PLDNN instead of Verisig 2.0 on DNNVerisig 2.0 · Table 2 benchmarksverification speedup: up to 438.6×Improved
compare LinCon with 20-core Verisig 2.0Verisig 2.0 (20 cores) · Table 2 benchmarksverification speedup: up to 58.1×Improved
verify with LinCon instead of Polar on easy ReLU casePolar · B1 ReLU2×20verification time (s): 3 -> 2.11Improved

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