Improving Long-Horizon Imitation through Instruction Prediction

2Citations
Citations of this article
10Readers
Mendeley users who have this article in their library.

Abstract

Complex, long-horizon planning and its combinatorial nature pose steep challenges for learning-based agents. Difficulties in such settings are exacerbated in low data regimes where over-fitting stifles generalization and compounding errors hurt accuracy. In this work, we explore the use of an often unused source of auxiliary supervision: language. Inspired by recent advances in transformer-based models, we train agents with an instruction prediction loss that encourages learning temporally extended representations that operate at a high level of abstraction. Concretely, we demonstrate that instruction modeling significantly improves performance in planning environments when training with a limited number of demonstrations on the BabyAI and Crafter benchmarks. In further analysis we find that instruction modeling is most important for tasks that require complex reasoning, while understandably offering smaller gains in environments that require simple plans. More details and code can be found at https://github.com/jhejna/instruction-prediction.

References Powered by Scopus

Planning and acting in partially observable stochastic domains

3122Citations
N/AReaders
Get full text

Vision-and-Language Navigation: Interpreting Visually-Grounded Navigation Instructions in Real Environments

896Citations
N/AReaders
Get full text

ALFRED: A benchmark for interpreting grounded instructions for everyday tasks

414Citations
N/AReaders
Get full text

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Cite

CITATION STYLE

APA

Hejna, J., Abbeel, P., & Pinto, L. (2023). Improving Long-Horizon Imitation through Instruction Prediction. In Proceedings of the 37th AAAI Conference on Artificial Intelligence, AAAI 2023 (Vol. 37, pp. 7857–7865). AAAI Press. https://doi.org/10.1609/aaai.v37i7.25951

Readers over time

‘23‘2402468

Readers' Seniority

Tooltip

PhD / Post grad / Masters / Doc 5

63%

Researcher 2

25%

Professor / Associate Prof. 1

13%

Readers' Discipline

Tooltip

Computer Science 5

56%

Engineering 3

33%

Mathematics 1

11%

Save time finding and organizing research with Mendeley

Sign up for free
0