BlockFLow: Decentralized, Privacy-Preserving, and Accountable Federated Machine Learning

Mugunthan, Vaikkunth; Rahman, Ravi; Kagal, Lalana

doi:10.1007/978-3-030-86162-9_23

Cited by 19 publications

(36 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With few exceptions (i.e. attacks on reputation [45], [50] and [66]), both security and privacy issues are rooted in local model updates. The security issue is related to the model poisoning which we mentioned above, while the privacy issue is related to sensitive information that might be leaked from the local updates.…”

Section: Rq 1-2: What Problems Were Solved?mentioning

confidence: 99%

“…Blockchains are open access where participants can deploy contracts pseudoanonymously 2) Private Blockchains do not allow access for clients outside the private network and require an entity that controls who is permitted to participate 3) Permissioned blockchains are private blockchains with a decentralized committee which controls the onboarding process Note that the FLFs that utilize open-source public blockchains such as Ethereum [32], [33], [35], [39], [41], [43], [50], [54], [56], [57], [64], Stellar [51] and EOS [60] were not deployed on the respective public blockchain in the experiments due to the enormous costs this would incur.…”

Section: ) Publicmentioning

confidence: 99%

“…aggregating millions of parameters, calculate expansive contribution measurements like Shapley Value or privacy-preserving methods hinder the framework to scale beyond a small group of entrusted entities towards mass adoption. To overcome the scalability-bottleneck of storage, some FLF applied an Interplanetary-File-System (IPFS) [71], where data is stored off-chain in a distributed file system, using the content-address as a unique pointer to each file in a global namespace over all computing devices [32], [34], [39], [42], [45], [50], [51], [60], [64]. Other FLF are based on novel design choices to tackle the scalability-issues: Witt et al [30] applied compressed Federated Knowledge Distillation, storing only 1-bit compressed soft-logits onchain.…”

Section: Rq 2-3: Is the Framework Scalable?mentioning

confidence: 99%

“…[32], [65], [69]), (v) workers (e.g. [50], [67], [68]) and (vi) smart contracts (e.g. [30], [35], [47]).…”

Section: Rq 3-2: How Are the Contributions Of Workers Measured?mentioning

confidence: 99%

“…Instead the distributed functionality was simulated or its impact estimated. For instance, Mugunthan et al [50] focus on the evaluation of the frameworks' contribution scoring procedure by simulating collusion attacks on the FL procedure. The effect of introducing blockchain to the FL framework was accounted for by estimating the per-agent gas consumption.…”

Section: Rq 4-2: How Comprehensive Are the Experiments?mentioning

confidence: 99%

See 4 more Smart Citations

Decentral and Incentivized Federated Learning Frameworks: A Systematic Literature Review

Witt¹,

Heyer²,

Toyoda³

et al. 2022

Preprint

View full text Add to dashboard Cite

The advent of Federated Learning (FL) has ignited a new paradigm for parallel and confidential decentralized Machine Learning (ML) with the potential of utilizing the computational power of a vast number of IoT, mobile and edge devices without data leaving the respective device, ensuring privacy by design. Yet, in order to scale this new paradigm beyond small groups of already entrusted entities towards mass adoption, the Federated Learning Framework (FLF) has to become (i) truly decentralized and (ii) participants have to be incentivized. This is the first systematic literature review analyzing holistic FLFs in the domain of both, decentralized and incentivized federated learning. 422 publications were retrieved, by querying 12 major scientific databases. Finally, 40 articles remained after a systematic review and filtering process for in-depth examination. Although having massive potential to direct the future of a more distributed and secure AI, none of the analyzed FLF is production-ready. The approaches vary heavily in terms of use-cases, system design, solved issues and thoroughness. We are the first to provide a systematic approach to classify and quantify differences between FLF, exposing limitations of current works and derive future directions for research in this novel domain.

show abstract

Section: Rq 1-2: What Problems Were Solved?mentioning

confidence: 99%