add some doc

com.unity.ml-agents.extensions/Runtime/MultiAgent/BaseMultiAgentGroup.cs

             }
         }
 
-        /// <summary>
-        /// Register the agent to the MultiAgentGroup.
-        /// Registered agents will be able to receive group rewards from the MultiAgentGroup
-        /// and share observations during training.
-        /// </summary>
+        /// <inheritdoc />
         public virtual void RegisterAgent(Agent agent)
         {
             if (!m_Agents.Contains(agent))
             }
         }
 
-        /// <summary>
-        /// Remove the agent from the MultiAgentGroup.
-        /// </summary>
+        /// <inheritdoc />
         public virtual void UnregisterAgent(Agent agent)
         {
             if (m_Agents.Contains(agent))
             }
         }
 
+        /// <inheritdoc />
         public int GetId()
         {
             return m_Id;
-        /// Get list of all agents registered to this MultiAgentGroup.
+        /// Get list of all agents currently registered to this MultiAgentGroup.
-        /// List of agents belongs to the MultiAgentGroup.
+        /// List of agents registered to the MultiAgentGroup.
         /// </returns>
         public List<Agent> GetRegisteredAgents()
         {
         /// <summary>
-        /// Add group reward for all agents under this MultiAgentGroup.
-        /// Disabled agent will not receive this reward.
+        /// Increments the group rewards for all agents in this MultiAgentGroup.
+        /// <remarks>
+        /// This function increase or decrease the group rewards by given amount for all agents
+        /// in the group. Use <see cref="SetGroupReward(float)"/> to set the group reward assigned
+        /// to the current step with a specific value rather than increasing or decreasing it.
+        ///
+        /// A positive group reward indicates the whole group's accomplishments or desired behaviors.
+        /// Every agent in the group will receive the same group reward no matter whether the
+        /// agent's act directly leads to the reward. Group rewards are meant to reinforce agents
+        /// to act in the group's best interest instead of indivisual ones.
+        /// Group rewards are treated differently than individual agent rewards during training, so
+        /// calling AddGroupReward() is not equivalent to calling agent.AddReward() on each agent in the group.
+        /// </remarks>
+        /// <param name="reward">Incremental group reward value.</param>
         public void AddGroupReward(float reward)
         {
             foreach (var agent in m_Agents)
         }
 
         /// <summary>
-        /// Set group reward for all agents under this MultiAgentGroup.
-        /// Disabled agent will not receive this reward.
+        /// Set the group rewards for all agents in this MultiAgentGroup.
+        /// <remarks>
+        /// This function replaces any group rewards given during the current step for all agents in the group.
+        /// Use <see cref="AddGroupReward(float)"/> to incrementally change the group reward rather than
+        /// overriding it.
+        ///
+        /// A positive group reward indicates the whole group's accomplishments or desired behaviors.
+        /// Every agent in the group will receive the same group reward no matter whether the
+        /// agent's act directly leads to the reward. Group rewards are meant to reinforce agents
+        /// to act in the group's best interest instead of indivisual ones.
+        /// Group rewards are treated differently than individual agent rewards during training, so
+        /// calling SetGroupReward() is not equivalent to calling agent.SetReward() on each agent in the group.
+        /// </remarks>
+        /// <param name="reward">The new value of the group reward.</param>
         public void SetGroupReward(float reward)
         {
             foreach (var agent in m_Agents)
         }
 
         /// <summary>
-        /// Returns the current step counter (within the current episode).
+        /// End episodes for all agents in this MultiAgentGroup.
+        /// <remarks>
+        /// This should be used when the episode can no longer continue, such as when the group
+        /// reaches the goal or fails at the task.
+        /// </remarks>
         public void EndGroupEpisode()
         {
             foreach (var agent in m_Agents)
         }
 
         /// <summary>
-        /// End Episode for all agents under this MultiAgentGroup.
+        /// Indicate that the episode is over but not due to the "fault" of the group.
+        /// This has the same end result as calling <see cref="EndGroupEpisode"/>, but has a
+        /// slightly different effect on training.
+        /// <remarks>
+        /// This should be used when the episode could continue, but has gone on for
+        /// a sufficient number of steps.
+        /// </remarks>
         public void GroupEpisodeInterrupted()
         {
             foreach (var agent in m_Agents)

com.unity.ml-agents/Runtime/Agent.cs

 
         internal void SetMultiAgentGroup(IMultiAgentGroup multiAgentGroup)
         {
-            // unregister from current group if this agent has been assigned one before
+            // Unregister from current group if this agent has been assigned one before
             UnregisterFromGroup?.Invoke(this);
 
             m_GroupId = multiAgentGroup.GetId();

com.unity.ml-agents/Runtime/IMultiAgentGroup.cs

         void RegisterAgent(Agent agent);
 
         /// <summary>
-        /// UnRegister agent from the MultiAgentGroup.
+        /// Unregister agent from the MultiAgentGroup.
         /// </summary>
         void UnregisterAgent(Agent agent);
     }