Pretrained Language Models as Visual Planners for Human Assistance
Authors:
- 1 October 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 15256-15268
Abstract
In our pursuit of advancing multi-modal AI assistants capable of guiding users to achieve complex multi-step goals, we propose the task of ‘Visual Planning for Assistance (VPA)’. Given a succinct natural language goal, e.g., "make a shelf", and a video of the user’s progress so far, the aim of VPA is to devise a plan, i.e. a sequence of actions such as "sand shelf", "paint shelf", etc. to realize the specified goal. This requires assessing the user’s progress from the (untrimmed) video, and relating it to the requirements of natural language goal, i.e. which actions to select and in what order? Consequently, this requires handling long video history and arbitrarily complex action dependencies. To address these challenges, we decompose VPA into video action segmentation and forecasting. Importantly, we experiment by formulating the forecasting step as a multimodal sequence modeling problem, allowing us to leverage the strength of pre-trained LMs (as the sequence model). This novel approach, which we call Visual Language Model based Planner (VLaMP), outperforms baselines across a suite of metrics that gauge the quality of the generated plans. Furthermore, through comprehensive ablations, we also isolate the value of each component – language pre-training, visual observations, and goal information. We have open-sourced all the data, model checkpoints, and training code.Keywords
All Related Versions
This publication has 46 references indexed in Scilit:
- Rescaling Egocentric Vision: Collection, Pipeline and Challenges for EPIC-KITCHENS-100International Journal of Computer Vision, 2021
- Long-Term Anticipation of Activities with Cycle Consistency2021
- Procedure Planning in Instructional Videos2020
- SoundSpaces: Audio-Visual Navigation in 3D Environments2020
- A Cordial Sync: Going Beyond Marginal Policies for Multi-agent Embodied Tasks2020
- Temporal Aggregate Representations for Long-Range Video Understanding2020
- Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification2018
- Next-active-object prediction from egocentric videosJournal of Visual Communication and Image Representation, 2017
- Anticipating Visual Representations from Unlabeled Video2016
- A Hierarchical Representation for Future Action Prediction2014