AI Paper Atlas

Research map / Reinforcement learning / Imitation learning

Imitation with Neural Density Models

Kuno Kim, Akshat Jindal, Yang Song, Jiaming Song, Yanan Sui, Stefano Ermon

NeurIPS 2021 · Reinforcement learning › Imitation learning · approach: Offline imitation from imperfect data · 4 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 NDI+EBM rather than the strongest listed baselineImitation learning with one demonstration · Hoppertask reward: 3261 \pm 533 -> 3458 \pm 210Improved
use NDI+EBM rather than the strongest listed baselineImitation learning with one demonstration · Half-Cheetahtask reward: 3456 \pm 401 -> 4511 \pm 569Improved
use NDI+EBM rather than the strongest listed baselineImitation learning with one demonstration · Walkertask reward: 3957 \pm 253 -> 5061 \pm 135Improved
use NDI+EBM rather than the strongest listed baselineImitation learning with one demonstration · Anttask reward: 2299 \pm 519 -> 4293 \pm 431Improved
use NDI+MADE rather than the strongest listed baselineImitation learning with one demonstration · Humanoidtask reward: 364 \pm 50 -> 6088 \pm 689Improved
use NDI+MADE density estimation instead of NDI+EBMNDI · Ant, one demonstrationtask reward: 4293 \pm 431 -> 555 \pm 311Hurt
set lambda_f to 0.0001 rather than 0.005 (ablation)NDI+EBM · Mujoco taskstask reward: Hopper: 3458 \pm 210 -> 3506 \pm 188; Half-Cheetah: 4511 \pm 569 -> 5697 \pm 805; Walker: 5061 \pm 135 -> 5171 \pm 157; Mixed
set lambda_f to 0.0001 rather than 0.005 (ablation)NDI+EBM · Mujoco tasksimitation KL (lower is better): Hopper: 0.11 \pm 0.02 -> 0.15 \pm 0.05; Half-Cheetah: 0.17 \pm 0.09 -> 0.32 \pm 0.15; Walker: 0.22 \pm 0.14 -> 0.25 \pm Hurt
set lambda_f to 0.1 rather than 0.005 (ablation)NDI+EBM · Mujoco taskstask reward: Hopper: 3458 \pm 210 -> 1057 \pm 29; Half-Cheetah: 4511 \pm 569 -> 103 \pm 59; Walker: 5061 \pm 135 -> 2710 \pm 501; AntHurt
set lambda_f to 0.1 rather than 0.005 (ablation)NDI+EBM · Mujoco tasksimitation KL (lower is better): Hopper: 0.11 \pm 0.02 -> 0.78 \pm 0.13; Half-Cheetah: 0.17 \pm 0.09 -> 1.41 \pm 0.51; Walker: 0.22 \pm 0.14 -> 0.41 \pm Hurt
use lambda_f=0.005 rather than lambda_f=0.0001 (ablation)NDI+EBM · Mujoco tasksimitation KL (lower is better): Hopper: 0.15 \pm 0.05 -> 0.11 \pm 0.02; Half-Cheetah: 0.32 \pm 0.15 -> 0.17 \pm 0.09; Walker: 0.25 \pm 0.04 -> 0.22 \pm Improved
add BC pretrainingNDI+MADE · Ant, one demonstrationtask reward: 555 \pm 311 -> 4014 \pm 105Improved

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