Mergent T Ranslation in M Ulti - a Gent C Ommunication
نویسندگان
چکیده
While most machine translation systems to date are trained on large parallel corpora, humans learn language in a different way: by being grounded in an environment and interacting with other humans. In this work, we propose a communication game where two agents, native speakers of their own respective languages, jointly learn to solve a visual referential task. We find that the ability to understand and translate a foreign language emerges as a means to achieve shared goals. The emergent translation is interactive and multimodal, and crucially does not require parallel corpora, but only monolingual, independent text and corresponding images. Our proposed translation model achieves this by grounding the source and target languages into a shared visual modality, and outperforms several baselines on both word-level and sentence-level translation tasks. Furthermore, we show that agents in a multilingual community learn to translate better and faster than in a bilingual communication setting.
منابع مشابه
Mergent C Omplexity via M Ulti - a Gent
Reinforcement learning algorithms can train agents that solve problems in complex, interesting environments. Normally, the complexity of the trained agent is closely related to the complexity of the environment. This suggests that a highly capable agent requires a complex environment for training. In this paper, we point out that a competitive multi-agent environment trained with self-play can ...
متن کاملFatemeh Tavakoli C ULTURAL S PECIFICATION IN T RANSLATION S TUDY OF “ T HE C ONFERENCE OF THE B IRDS ”
متن کامل
O Rganizational M Ulti - a Gent S Ystems : a P Rocess D Riven a Pproach
There exists a need in industrial and business applications to intelligently integrate data, information and knowledge from a diverse range of sources, particularly during product design, or policy formation. Optimizing decision making requires the expertise of many agents, both computational and human, to be combined and coordinated. The concept of an ’organization’ has emerged as central to t...
متن کاملNformation S Ecurity a Pproach in O Pen D Istributed M Ulti - a Gent V
This paper presented the main information, security problems and threats in open multi-agent distributed e-learning information systems and Proposed various approaches to solve information security attacks in virtual learning environment using service oriented architecture which based on multi-agent information systems architecture, the solution on the multi-agent learning information system im...
متن کامل