Visual Referential Games Further the Emergence of Disentangled Representations

要約

タイトル：ビジュアルリファレンスゲームは、Disentangled Representationsの出現を促進する

要約：
-自然言語は、情報を伝えるために人間が使用する強力なツールです。
-深層学習で汎化性が高い学習表現を実現するために、Disentangledな研究が重要となってきました。
-この論文では、ビジュアルリファレンスゲームにおける合成性、システマティック性、分解能の関係について調査しています。
-結果、Obverterアーキテクチャを基にした視覚的リファレンスゲームは、多くのDisentanglementメトリックを上回る結果を出すことがわかりました。
-また、新しいPosDisメトリックが提案され、Mutual Information Gap（MIG）メトリックからの情報性と完全性の懸念を補完することが影響を与えると発見され、異なる合成言語の区別が可能になりました。
-最後に、Zero-shot組成学習テストでの結果について調査し、これまで提案された分解能と合成性メトリックとの相関関係を検討しました。訓練プロセスを通じて、静的に有意な相関係数が、測定の瞬間によって正か負かで現れます。

要約(オリジナル)

Natural languages are powerful tools wielded by human beings to communicate information. Among their desirable properties, compositionality has been the main focus in the context of referential games and variants, as it promises to enable greater systematicity to the agents which would wield it. The concept of disentanglement has been shown to be of paramount importance to learned representations that generalise well in deep learning, and is thought to be a necessary condition to enable systematicity. Thus, this paper investigates how do compositionality at the level of the emerging languages, disentanglement at the level of the learned representations, and systematicity relate to each other in the context of visual referential games. Firstly, we find that visual referential games that are based on the Obverter architecture outperforms state-of-the-art unsupervised learning approach in terms of many major disentanglement metrics. Secondly, we expand the previously proposed Positional Disentanglement (PosDis) metric for compositionality to (re-)incorporate some concerns pertaining to informativeness and completeness features found in the Mutual Information Gap (MIG) disentanglement metric it stems from. This extension allows for further discrimination between the different kind of compositional languages that emerge in the context of Obverter-based referential games, in a way that neither the referential game accuracy nor previous metrics were able to capture. Finally we investigate whether the resulting (emergent) systematicity, as measured by zero-shot compositional learning tests, correlates with any of the disentanglement and compositionality metrics proposed so far. Throughout the training process, statically significant correlation coefficients can be found both positive and negative depending on the moment of the measure.

arxiv情報

提供元, 利用サービス

arxiv.jp, OpenAI