Here, we report an artificial thermal recognition memory system mainly consisting of programmable polymer memristors and resistive temperature detectors. The memristor has a two-terminal structure with an insulating polymer sandwiched between a nanowrinkled graphene electrode and a nanopillar metal electrode, exhibiting novel voltage-controlled programmable rewritable and nonerasable nonvolatile memory effects. The integrated system is capable of detecting diverse weak and strong thermal stimuli and encoding them into binary memory signals for the selective recording of human thermal perception recognition. This work offers an effective strategy for constructing programmable memristive devices via interfacial nanoengineering and provides a new architecture of recognition memory for artificial intelligence.

在这里，我们报告了一种人工热识别记忆系统，主要由可编程聚合物忆阻器和电阻温度检测器组成。忆阻器具有两端结构，绝缘聚合物夹在纳米皱纹石墨烯电极和纳米柱金属电极之间，表现出新颖的电压控制可编程可重写和不可擦除非易失性记忆效应。该集成系统能够检测各种弱和强热刺激，并将其编码为二进制记忆信号，用于选择性记录人体热感知识别。这项工作为通过界面纳米工程构建可编程忆阻器件提供了一种有效的策略，并为人工智能提供了一种新的识别记忆架构。