Several recent works have found the emergence of grounded com-positional language in the communication protocols developed bymostly cooperative multi-agent systems when learned end-to-endto maximize performance on a downstream task. However, humanpopulations learn to solve complex tasks involving communicativebehaviors not only in fully cooperative settings but also in scenar-ios where competition acts as an additional external pressure forimprovement. In this work, we investigate whether competitionfor performance from an external, similar agent team could actas a social influence that encourages multi-agent populations todevelop better communication protocols for improved performance,compositionality, and convergence speed. We start fromTask &Talk, a previously proposed referential game between two coopera-tive agents as our testbed and extend it intoTask, Talk & Compete,a game involving two competitive teams each consisting of twoaforementioned cooperative agents. Using this new setting, we pro-vide an empirical study demonstrating the impact of competitiveinfluence on multi-agent teams. Our results show that an externalcompetitive influence leads to improved accuracy and generaliza-tion, as well as faster emergence of communicative languages thatare more informative and compositional.
more »
« less
Towards Situated Communication in Multi-Step Interactions: Time is a Key Pressure in Communication Emergence
Enabling efficient communication in artificial agents brings us closer to machines that can cooperate with each other and with human partners. Hand-engineered approaches have substantial limitations, leading to increased interest in methods for communication to emerge autonomously between artificial agents. Most of the research in the field explores unsituated communication in one-step referential tasks. The tasks are not temporally interactive and lack time pressures typically present in natural communication and language learning. In these settings, agents can successfully learn what to communicate but not when or whether to communicate. Here, we extend the literature by assessing emergence of communication between reinforcement learning agents in a temporally interactive, cooperative task of navigating a gridworld environment. We show that, through multi-step interactions, agents develop just-in-time messaging protocols that enable them to successfully solve the task. With memory—which provides flexibility around message timing—agent pairs converge to a look-ahead communication protocol, finding an optimal solution to the task more quickly than without memory. Lastly, we explore situated communication, enabling the acting agent to choose when and whether to communicate. With the opportunity cost of forgoing an action to communicate, the acting agent learns to solicit information sparingly, in line with the Gricean Maxim of quantity. Our results point towards the importance of studying language emergence through situated communication in multi-step interactions.
more »
« less
- Award ID(s):
- 1837515
- NSF-PAR ID:
- 10471793
- Publisher / Repository:
- Cognitive Science Society, https://escholarship.org/uc/item/61k7486v
- Date Published:
- Journal Name:
- Annual Meeting of the Cognitive Science Society
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)A means to communicate by touch is established when two humans grasp a common rigid object, and such communication is thought to play a role in the superior performance two humans acting together are able to demonstrate over either agent acting alone. But the superior performance demonstrated by dyads, whether in making point-to-point movements or tracking unpredictable targets, is strictly empirical to date. Mechanistic accounts for the performance improvement and explanations relying on haptic communication have been lacking. In this paper we develop a model of haptic communication across a linkage connecting two agents that provides an explicit means for the dyad to achieve a higher loop gain than either agent acting alone and higher than the two agents acting together without haptic feedback. We show that haptic communication closes an additional feedback loop through the linkage and the sensorimotor control systems of both agents. This feedback loop contributes a new factor to the loop gain and thus a definitive mechanism for the dyad to improve performance. Our model predicts higher internal forces with haptic communication, which have previously been observed. Additional testable hypotheses emerge from the model and create a promising future means to transfer human-human dyad behaviors to human-robot teams.more » « less
-
Regardless of how much data artificial intelligence agents have available, agents will inevitably encounter previously unseen situations in real-world deployments. Reacting to novel situations by acquiring new information from other people—socially situated learning—is a core faculty of human development. Unfortunately, socially situated learning remains an open challenge for artificial intelligence agents because they must learn how to interact with people to seek out the information that they lack. In this article, we formalize the task of socially situated artificial intelligence—agents that seek out new information through social interactions with people—as a reinforcement learning problem where the agent learns to identify meaningful and informative questions via rewards observed through social interaction. We manifest our framework as an interactive agent that learns how to ask natural language questions about photos as it broadens its visual intelligence on a large photo-sharing social network. Unlike active-learning methods, which implicitly assume that humans are oracles willing to answer any question, our agent adapts its behavior based on observed norms of which questions people are or are not interested to answer. Through an 8-mo deployment where our agent interacted with 236,000 social media users, our agent improved its performance at recognizing new visual information by 112%. A controlled field experiment confirmed that our agent outperformed an active-learning baseline by 25.6%. This work advances opportunities for continuously improving artificial intelligence (AI) agents that better respect norms in open social environments.more » « less
-
null (Ed.)We summarize our past five years of work on designing, building, and studying Sugilite, an interactive task learning agent that can learn new tasks and relevant associated concepts interactively from the user’s natural language instructions and demonstrations leveraging the graphical user interfaces (GUIs) of third-party mobile apps. Through its multi-modal and mixed-initiative approaches for Human- AI interaction, Sugilite made important contributions in improving the usability, applicability, generalizability, flexibility, robustness, and shareability of interactive task learning agents. Sugilite also represents a new human-AI interaction paradigm for interactive task learning, where it uses existing app GUIs as a medium for users to communicate their intents with an AI agent instead of the interfaces for users to interact with the underlying computing services. In this chapter, we describe the Sugilite system, explain the design and implementation of its key features, and show a prototype in the form of a conversational assistant on Android.more » « less
-
Li, Yang ; Hilliges, Otmar (Ed.)We summarize our past five years of work on designing, building, and studying Sugilite, an interactive task learning agent that can learn new tasks and relevant associated concepts interactively from the user’s natural language instructions and demonstrations leveraging the graphical user interfaces (GUIs) of third-party mobile apps. Through its multi-modal and mixed-initiative approaches for Human-AI interaction, Sugilite made important contributions in improving the usability, applicability, generalizability, flexibility, robustness, and shareability of interactive task learning agents. Sugilite also represents a new human-AI interaction paradigm for interactive task learning, where it uses existing app GUIs as a medium for users to communicate their intents with an AI agent instead of the interfaces for users to interact with the underlying computing services. In this chapter, we describe the Sugilite system, explain the design and implementation of its key features, and show a prototype in the form of a conversational assistant on Android.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
