An Efficient and Secure Privacy-Preserving Federated Learning Framework Based on Multiplicative Double Privacy Masking

With the increasing awareness of privacy protection and the improvement of relevant laws,federal learning has gradually become a new choice for cross-agency and cross-device machine learning.In order to solve the problems of privacy leakage,high computational overhead and high traffic in some federated learning schemes,this paper proposes amultiplicative double privacymask algorithm which is convenient for homomorphic addition aggregation.The combination of homomorphic encryption and secret sharing ensures that the server cannot compromise user privacy from the private gradient uploaded by the participants.At the same time,the proposed TQRR(Top-Q-Random-R)gradient selection algorithm is used to filter the gradient of encryption and upload efficiently,which reduces the computing overhead of 51.78%and the traffic of 64.87%on the premise of ensuring the accuracy of themodel,whichmakes the framework of privacy protection federated learning lighter to adapt to more miniaturized federated learning terminals.

Linking the hierarchical decision- making process of fruit choice and the phenotypic selection strength on fruit traits by birds

Aims Animals in search of fleshy fruits forage mostly according to the number of available fruits and then select individual fruits based on reward quality or advertised subtle traits.This hierarchical pat-tern of fruit choice would be translated into patterns of selection strength mediated by frugivores on fruit display traits.Thus,frugi-vores would exert higher selection pressures on fruit crop size and lower selection pressures on within-plant variation of phenotypic traits(infructescence,fruit and seed size).However,no attempt to link this behavioral mechanism of hierarchical trait selection to natural selection patterns has been made.Therefore,we sought to determine the relationship between the hierarchical decision-mak-ing process of fruit choice and patterns of natural selection on fruit traits.Methods We recorded bird visits and measured fruit-related traits(fruit crop size,fruit diameter and seed weight)in a natural population of Psychotria carthagenensis,a bird-dispersed treelet,in a Yungas for-est from Argentina.To assess phenotypic selection patterns on fruit display traits,we performed multivariate selection analysis,and to explicitly identify a hierarchy of fruit trait choice we used a classifi-cation tree as a predictive model.Important Findings Selection patterns on fruit display traits were in agreement with a hierarchical process of fruit choice made by birds.The strength of directional selection on the total number of fruits in a plant(i.e.fruit crop size)was nearly two times higher than on fruit size,and the classification tree analysis supported this hierarchical pattern.Our results support previous evidence that seed dispersers shape fruit crop size with higher intensity than subindividual fruit traits.Also,high levels of subindividual phenotypic variation of fruit display traits may be explained by relaxed selection pressures exerted by frugivores.Empirical studies also show that this pattern may consti-tute a general phenomenon among other plant-animal interactions.