Abstract: The disclosure discloses a preparation method and application of a non-noble metal single atom catalyst, and belongs to the technical fields of chemistry, chemical engineering and material science. According to the disclosure, cheap raw materials and simple method are used to prepare the single atom catalyst. In essence, metal is anchored on light-absorbing carrier in a single atom form under irradiation to produce the single atom catalyst. In the disclosure, the non-noble metal single atom catalyst is prepared by using a photochemical synthetic route for the first time. The single atom catalyst synthesized in the disclosure is dispersed on the surface of photoactive substance. Using nickel single atom as a co-catalyst in photocatalytic water splitting to produce hydrogen, the cost is low and the catalytic efficiency is greatly improved compared with other types of non-noble metal modified composite photocatalysts.

Abstract: The disclosure discloses a preparation method and application of a non-noble metal single atom catalyst, and belongs to the technical fields of chemistry, chemical engineering and material science. According to the disclosure, cheap raw materials and simple method are used to prepare the single atom catalyst. In essence, metal is anchored on light-absorbing carrier in a single atom form under irradiation to produce the single atom catalyst. In the disclosure, the non-noble metal single atom catalyst is prepared by using a photochemical synthetic route for the first time. The single atom catalyst synthesized in the disclosure is dispersed on the surface of photoactive substance. Using nickel single atom as a co-catalyst in photocatalytic water splitting to produce hydrogen, the cost is low and the catalytic efficiency is greatly improved compared with other types of non-noble metal modified composite photocatalysts.

Abstract: In some implementations of this invention, the performance of a network of reinforcement learning agents is maximized by optimizing the communication topology between the agents for the communication of gradients, weights or rewards. For instance, a sparse Erdos-Renyi network may be employed, and network density may be selected in such a way as to maximize reachability and to minimize homogeneity. In some cases, a sparse network topology is employed for massively distributed learning, such as across entire fleets of autonomous vehicles or mobile phones that learn from each other instead of requiring a master to coordinate learning.

Abstract: In some implementations of this invention, the performance of a network of reinforcement learning agents is maximized by optimizing the communication topology between the agents for the communication of gradients, weights or rewards. For instance, a sparse Erdos-Renyi network may be employed, and network density may be selected in such a way as to maximize reachability and to minimize homogeneity. In some cases, a sparse network topology is employed for massively distributed learning, such as across entire fleets of autonomous vehicles or mobile phones that learn from each other instead of requiring a master to coordinate learning.

Abstract: In some implementations of this invention, the performance of a network of reinforcement learning agents is maximized by optimizing the communication topology between the agents for the communication of gradients, weights or rewards. For instance, a sparse Erdos-Renyi network may be employed, and network density may be selected in such a way as to maximize reachability and to minimize homogeneity. In some cases, a sparse network topology is employed for massively distributed learning, such as across entire fleets of autonomous vehicles or mobile phones that learn from each other instead of requiring a master to coordinate learning.

Abstract: In some implementations of this invention, the performance of a network of reinforcement learning agents is maximized by optimizing the communication topology between the agents for the communication of gradients, weights or rewards. For instance, a sparse Erdos-Renyi network may be employed, and network density may be selected in such a way as to maximize reachability and to minimize homogeneity. In some cases, a sparse network topology is employed for massively distributed learning, such as across entire fleets of autonomous vehicles or mobile phones that learn from each other instead of requiring a master to coordinate learning.