Research map / Vision-language and multimodal / Robot action and navigation
Episodic Transformer for Vision-and-Language Navigation
Alexander Pashevich, Cordelia Schmid, Chen Sun
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 changed | Where | Measured | Effect |
|---|---|---|---|
| replace LSTM with episodic transformer | LSTM agent · ALFRED validation seen | task success (%): 23.2 -> 33.8 | Improved |
| replace LSTM with episodic transformer | LSTM agent · ALFRED validation unseen | task success (%): 2.4 -> 3.2 | Improved |
| replace LSTM with episodic transformer | LSTM agent · ALFRED validation sub-goals seen | sub-goal success (%): 75.5 -> 77.3 | No clear effect |
| give the LSTM the E.T. language encoder (ablation) | LSTM agent · ALFRED validation seen | task success (%): 23.2 -> 27.8 | Improved |
| add access to the full history of visual frames instead of the last frame (ablation) | E.T. · ALFRED validation seen | task success (%): 28.9 -> 33.8 | Improved |
| add access to more than the last action (all actions instead of 1) (ablation) | E.T. · ALFRED validation seen | task success (%): 33.8 -> 27.1 | Hurt |
| add the last action as input (versus none) (ablation) | E.T. · ALFRED validation seen | task success (%): 23.7 -> 33.8 | Improved |
| use 2 transformer blocks instead of 1 (ablation) | E.T. · ALFRED validation seen | task success (%): 25.0 -> 33.8 | Improved |
| use 3 blocks instead of 2 (ablation) | E.T. · ALFRED validation seen | task success (%): 33.8 -> 28.6 | Hurt |
| add joint training with subgoal-action synthetic instructions (ablation) | E.T. · ALFRED validation seen | task success (%): 33.8 -> 38.5 | Improved |
| add joint training with expert-frame instructions (ablation) | E.T. · ALFRED validation seen | task success (%): 33.8 -> 28.5 | Hurt |
| add BERT embeddings for language pretraining (ablation) | E.T. · ALFRED validation seen | task success (%): 33.8 -> 32.3 | Hurt |
Closest papers by meaning
- A New Path: Scaling Vision-and-Language Navigation With Synthetic Instructions and Imitation Learning (CVPR 2023 · similarity 0.91)
- History Aware Multimodal Transformer for Vision-and-Language Navigation (NeurIPS 2021 · similarity 0.90)
- Multimodal Transformer with Variable-Length Memory for Vision-and-Language Navigation (ECCV 2022 · similarity 0.90)
- SOAT: A Scene- and Object-Aware Transformer for Vision-and-Language Navigation (NeurIPS 2021 · similarity 0.89)
- NavBench: Probing Multimodal Large Language Models for Embodied Navigation (NeurIPS 2025 · similarity 0.88)
- Planning from Imagination: Episodic Simulation and Episodic Memory for Vision-and-Language Navigation (AAAI 2025 · similarity 0.88)
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.
