Unsupervised Translation of Emergent Communication

Emergent Communication (EC) provides a unique window into the language systems that emerge autonomously when agents are trained to jointly achieve shared goals. However, it is difficult to interpret EC and evaluate its relationship with natural languages (NL). This study employs unsupervised neural machine translation (UNMT) techniques to decipher ECs formed during referential games with varying task complexities, influenced by the semantic diversity of the environment. Our findings demonstrate UNMT's potential to translate EC, illustrating that task complexity characterized by semantic diversity enhances EC translatability, while higher task complexity with constrained semantic variability exhibits pragmatic EC, which, although challenging to interpret, remains suitable for translation. This research marks the first attempt, to our knowledge, to translate EC without the aid of parallel data.

Authors

• Ido Levy Technion - Israel Institute of Technology
• Orr Paradise University of California, Berkeley
• Boaz Carmeli Technion - Israel Institute of Technology
• Ron Meir Technion - Israel Institute of Technology
• Shafi Goldwasser University of California, Berkeley
• Yonatan Belinkov Technion, Technion

Keywords

No keywords are indexed for this paper.