AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

Vision-Based Navigation With Language-Based Assistance via Imitation Learning With Indirect Intervention

Khanh Nguyen, Debadeepta Dey, Chris Brockett, Bill Dolan

CVPR 2019 · Vision-language and multimodal › Robot action and navigation · approach: Pretraining and data augmentation for VLN · 107 citations (OpenAlex)

Check 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
Learn a policy that requests language subgoal assistance when neededVNLA agent · AskNav Test seensuccess rate (%): 42.98 -> 52.09Improved
Learn a policy that requests language subgoal assistance when neededVNLA agent · AskNav Test unseensuccess rate (%): 25.05 -> 34.50Improved
Replace never requesting help with continuous requests from the beginningVNLA agent · AskNav Test seensuccess rate (%): 28.39 -> 40.33Improved
Replace never requesting help with continuous requests from the beginningVNLA agent · AskNav Test unseensuccess rate (%): 6.36 -> 20.00Improved
Use learned help requests instead of random requestsVNLA agent · AskNav Test seensuccess rate (%): 42.98 -> 52.09Improved
Use learned help requests instead of random requestsVNLA agent · AskNav Test unseensuccess rate (%): 25.05 -> 34.50Improved
Use the learned help-request policy instead of the help-requesting teacherVNLA agent · AskNav Test seensuccess rate (%): 52.26 -> 52.09No clear effect
Add language subgoals to direct interventionsDirect-advisor VNLA agent · AskNav Test unseensuccess rate (%): 32.19 -> 34.56Improved
Replace direct decision overwriting with indirect subgoal interventionsVNLA agent · AskNav Test seensuccess rate (%): 52.09 -> 52.09No clear effect
Replace direct decision overwriting with indirect subgoal interventionsVNLA agent · AskNav Test unseensuccess rate (%): 34.56 -> 34.50No clear effect
Train with room-type goals, then evaluate object finding without room typesVNLA agent with learned help requests · NoRoom Test seensuccess rate (%): 43.69 -> 53.85Improved
Train with room-type goals, then evaluate object finding without room typesVNLA agent with learned help requests · NoRoom Test unseensuccess rate (%): 33.41 -> 41.63Improved

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