Viewing profile — xidong_wu

[flagged]

This paper optimizes a controller/policy which will be used to agent itself in the next round

Your understanding is basically correct. "how applicable the search controller is when applied to new problems". We need meta-agent thinking pattern. Self-evolving agent has been v…