Timezone: »

Poster
Ask4Help: Learning to Leverage an Expert for Embodied Tasks
Kunal Pratap Singh · Luca Weihs · Alvaro Herrasti · Jonghyun Choi · Aniruddha Kembhavi · Roozbeh Mottaghi

Tue Nov 29 09:00 AM -- 11:00 AM (PST) @ Hall J #639
Embodied AI agents continue to become more capable every year with the advent of new models, environments, and benchmarks, but are still far away from being performant and reliable enough to be deployed in real, user-facing, applications. In this paper, we ask: can we bridge this gap by enabling agents to ask for assistance from an expert such as a human being? To this end, we propose the Ask4Help policy that augments agents with the ability to request, and then use expert assistance. Ask4Help policies can be efficiently trained without modifying the original agent's parameters and learn a desirable trade-off between task performance and the amount of requested help, thereby reducing the cost of querying the expert. We evaluate Ask4Help on two different tasks -- object goal navigation and room rearrangement and see substantial improvements in performance using minimal help. On object navigation, an agent that achieves a $52\%$ success rate is raised to $86\%$ with $13\%$ help and for rearrangement, the state-of-the-art model with a $7\%$ success rate is dramatically improved to $90.4\%$ using $39\%$ help. Human trials with Ask4Help demonstrate the efficacy of our approach in practical scenarios.

#### Author Information

##### Jonghyun Choi (Yonsei University)

I received a Ph.D. degree from University of Maryland, College Park, under the supervision of Prof. Larry S. Davis, and a B.S. and a M.S. degree from Seoul National University, under the supervision of Prof. Kyoung-Mu Lee in SNU Computer Vision Lab. I was an assistant professor in GIST and a researcher in Allen Institute for Artificial Intelligence (AI2) (2016-2018) and Comcast Labs, DC (2015). During my PhD, I was fortunate to work as a research intern in Microsoft Research, Redmond (2014 Summer), Disney Research, Pittsburgh (2014 Spring), Adobe Research, San Jose (2013 Summer) and U.S. Army Research Lab. (Adelphi, MD, Summer 2011).