Abstract
Living systems have inspired research on non-biological dynamic materials and systems chemistry to mimic specific complex biological functions. Upon pursuing ever more complex life-inspired non-biological systems, mimicking even the most elementary aspects of learning is a grand challenge. We demonstrate a programmable hydrogel-based model system, whose behaviour is inspired by associative learning, i.e., conditioning, which is among the simplest forms of learning. Algorithmically, associative learning minimally requires responsivity to two different stimuli and a memory element. Herein, nanoparticles form the memory element, where a photoacid-driven pH-change leads to their chain-like assembly with a modified spectral behaviour. On associating selected light irradiation with heating, the gel starts to melt upon the irradiation, originally a neutral stimulus. A logic diagram describes such an evolution of the material response. Coupled chemical reactions drive the system out-of-equilibrium, allowing forgetting and memory recovery. The findings encourage to search non-biological materials towards associative and dynamic properties.
Highlights
Living systems have inspired research on non-biological dynamic materials and systems chemistry to mimic specific complex biological functions
To address the above question, we explore whether artificial non-biological materials could be designed to show algorithmic responses inspired by associative learning[20,21]
In a reductionistic approach, a synthetic material may algorithmically show some aspects of associative learning, provided that the system is capable of responding to two different stimuli, combined with a built-in memory element
Summary
Living systems have inspired research on non-biological dynamic materials and systems chemistry to mimic specific complex biological functions. Adopting a still more reductionist approach, a generic question could be posed: can even inanimate materials be designed to show programmed responses inspired by some algorithmic forms of learning, in resemblance to systems based on synaptic electronics or biochemical circuits showing associatively responsive behaviours14–19?. The system is designed such that associating such an irradiation (the neutral stimulus) and heating (the unconditioned stimulus) triggers a memory, whereafter a subsequent sole irradiation with the given spectral composition allows the gel melting (algorithmically inspired by the conditioning process in dogs). The two stimuli are not fully orthogonal and the response pathways (heating and light-induced heating) are interdependent, which deviates from the ideal concept, the programmed hydrogel system algorithmically shows many aspects of biological learning, and encourages research in synthetic materials towards more associative and dynamic functionalities
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
