abstract

AAAI2021/paper.tex

@@ -167,19 +167,19 @@
 
 \begin{abstract}
  The natural emergence of symbolic languages with high compositionality has
-  attracted extensive attentions from a broad range of communities. Existing
+  attracted extensive attention from a broad range of communities. Existing
   studies achieve high compositionality through \emph{deliberately handcrafted}
   inductions (e.g., small vocabulary sizes, carefully constructed distractors, 
   and ease-of-teaching) in multi-agent learning, which are unnatural. 
-  Yet, few studies investigate the emergence of symbolic language with high compositionality in
-  \emph{``natural''} environments, i.e., without any deliberately handcrafted
+  Yet, few studies investigate the emergence of symbolic language with high
+  compositionality  \emph{``naturally''}, i.e., without any deliberately handcrafted
   inductions. 
   
   In this paper, we are the first to successfully achieve high compositional symbolic
-  language in a purely \emph{natural} environment.
-  Initially, by thoroughly investigating the compositionality of symbolic
-  language emerged after removing the \emph{deliberately handcrafted}
-  inductions, we observe that the agent capacity plays the key role in
+  language in a purely \emph{natural} manner.
+  Initially, by thoroughly investigating the compositionality of emerged symbolic
+  language after removing the \emph{deliberately handcrafted}
+  inductions, we observe that the agent capacity plays a key role in
   compositionality. We further reveal and characterize the quantitative relationship
   between the agent capacity and the compositionality of symbolic language both
   theoretically and experimentally. The theoretical analysis is built on the MSC
@@ -189,8 +189,8 @@
   with eliminated external environment factors. Both theoretical analysis and
   experimental results lead to a counter-intuitive conclusion that lower agent
   capacity facilitates the emergence of symbolic language with higher
-  compositionality. Based on our conclusion, we are able to generate higher
-  compositional symbolic language with a high probability. 
+  compositionality. \note{Based on our conclusion, we can generate higher
+  compositional symbolic language with a high probability.}
 
 
 %  The natural emergence of symbolic languages with high compositionality has

AAAI2021/tex/introduction.tex

@@ -13,14 +13,15 @@ learning~\cite{}. For example, \note{XXXX}
 %the environment setting.  
 
 
-To evaluate the emerged symbolic language, compositionality is widely used and
-taken as an important metric. Roughly, compositionality is a principle that the
-meaning of a complex expression (e.g, phase), which is assembled out of the
-given set of simple components (e.g., symbols), 
-is determined by its constituent components and the rules that combines them~\cite{}.
-For example, the expression "AAAI is a conference'' is consists of two
-meaningful words "'' 
-
+Compositionality is widely used and
+taken as an important metric to evaluate the emerged symbolic language.
+Originally, compositionality is a principle that
+whether the meaning of a complex expression (e.g, phase), which is assembled out of the
+given set of simple components (e.g., symbols), can be determined by its
+constituent components and the rules that combines them~\cite{}.
+\note{For example, the expression "AAAI is a conference'' consists of two
+meaningful words ``AAAI'' and ``conference'', and a rule of definition (``is'').}
+More recently, measuring the compositionality \note{xxxxx}.
 
 
 %It
@@ -33,26 +34,47 @@ meaningful words "''
 %extract information from a single symbol.
 %
 %
-%\begin{figure}[t]
-%  \centering
-%  \includegraphics[width=0.9\columnwidth]{fig/occupy}
-%  \caption{(a): The correspondence between symbol sequences ($s_0$, $s_1$) and (shape,
-%color) pairs in a perfectly compostional language. $s_0$, $s_1$ in {a, b, c}, shape
-%in {circle, square} and color in {red, blue, green}; (b): The correspondence
-%between symbol sequences ($s_0$, $s_1$) and (shape, color) pairs in a language with
-%low compostionality.}
-%  \label{fig:symbols}
-%  \end{figure}
-
-Prior studies focus on investigating how to affect the compositionality of the
-emergent language. Researchers have found that various environmental pressures
-would affect compositionality, e.g., small vocabulary sizes[3], memoryless[4],
-carefully constructed rewards[5] and ease-of-teaching[6]. However, these works
-only consider \emph{nurture} [7] (i.e., environmental factors), rather than
-\emph{nature} (i.e., hereditary factors from agents), when inducing or exploring
-the emergent language without exception. Moreover, some environmental pressures,
-like regrading the entropy as an item of additional rewards, may be too ideal to
-exist in the real world. 
+\begin{figure}[t]
+  \centering
+  \includegraphics[width=0.9\columnwidth]{fig/occupy}
+  \caption{\rmk{compositionality.}}
+  \label{fig:symbols}
+  \end{figure}
+
+Prior studies focus on achieving high compositionality of the emergent language
+through \emph{deliberately handcrafted} inductions unnaturally, e.g., small vocabulary
+sizes~\cite{}, memoryless~\cite{}, carefully constructed rewards~\cite{}, and
+ease-of-teaching~\cite{}. \note{xxxxxxx}
+However, these unnatural inductions prevent us to better understand the mystery of
+the emergence of language and even intelligence among our pre-human ancestors.
+
+
+Yet, few works investigate the emergence of symbolic language with high
+compositionality in \emph{naturally} (i.e., without \emph{deliberately
+  handcrafted} inductions). As a results, it is never clear whether \emph{natural}
+environment and agent are sufficient for compositionality. 
+In this work, we focus on generating high compositional symbolic language
+naturally without any ``human'' induction.
+Initially, we thoroughly analyze the compositionality of emerged symbolic
+language after removing the \emph{deliberately handcrafted}
+inductions. Figure~\ref{fig:comp} reports the compositionality when train two
+agents in a listener-speaker referential game. It can be observed that \note{xxxxxxxx}.
+Thus, it is challenging to achieve high compositionality without induction.
+
+
+================
+
+
+%Prior studies focus on investigating how to affect the
+%compositionality of the emergent language. Researchers
+%have found that various environmental pressures would affect compositionality,
+%e.g., small vocabulary sizes[3], memoryless[4], 
+%carefully constructed rewards[5] and ease-of-teaching[6]. However, these works
+%only consider \emph{nurture} [7] (i.e., environmental factors), rather than
+%\emph{nature} (i.e., hereditary factors from agents), when inducing or exploring
+%the emergent language without exception. Moreover, some environmental pressures,
+%like regrading the entropy as an item of additional rewards, may be too ideal to
+%exist in the real world. 
 
 In contrast to prior work, we investigate the compositionality of emergent
 language from a new perspective, i.e., the agent capacity. Different from

NIPS2020/main.tex

@@ -166,7 +166,6 @@ works can be roughly classified into two categories, referential game and
 multi-agent reinforcement learning (MARL), based on the environment setting.
 However, previous works, no matter referential game related or multi-agent reinforcement
 learning related, ignore the independence of agents in \note{training or
-inference.} Agents usually share one or more of the model parameters, loss functions,
 observation of environments, and thusly can be taken as one huge brain with
 multiple connected sensors (agents). In other words, previous works did not
 really achieve the emergence of symbolic language among \emph{multiple} agents.