AI Paper Atlas

Research map / Reinforcement learning / Imitation learning

SafeMIL: Learning Offline Safe Imitation Policy from Non-Preferred Trajectories

Returaj Burnwal, Nirav Pravinbhai Bhatt, Balaraman Ravindran

AAAI 2026 · Reinforcement learning › Imitation learning · approach: Offline imitation from imperfect data · 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 SafeMIL instead of BC on mixed unlabeled dataoffline safe imitation · Walker/Swimmer/Ant-Velocitynormalized costImproved
use SafeMIL instead of baselinesoffline safe imitation · Point-Goal1normalized costImproved
use SafeMIL instead of T-REXoffline safe imitation · Point-Circle2normalized costHurt
use SafeMIL instead of baselinesoffline safe imitation · Point-Button1normalized costNo clear effect
increase bag size K from 1 up to 128 (ablation)SafeMIL · Swimmer-Velocitynormalized costImproved
use partial trajectories of length H=1, 5, 10 (K=128) instead of full trajectories (ablation)SafeMIL · Swimmer-Velocitynormalized cost and returnNo clear effect
weight BC loss per transition instead of per trajectory (ablation)SafeMIL · Swimmer-Velocity and Point-Goal1cost and returnNo clear effect

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