ML Security PapersEnhancing Split Learning with Sharded and Blockchain-Enabled SplitFed Approaches
Amirreza Sokhankhosh 1, Khalid Hassan 1, Sara Rouhani 2
Published on arXiv
2509.25555
Data Poisoning Attack
OWASP ML Top 10 — ML02
Key Finding
BSFL increases resilience to data poisoning attacks by 62.7% while SSFL improves scalability by 85.2% and performance by 31.2% over baseline SplitFed Learning.
BSFL (Blockchain-enabled SplitFed Learning)
Novel technique introduced
Collaborative and distributed learning techniques, such as Federated Learning (FL) and Split Learning (SL), hold significant promise for leveraging sensitive data in privacy-critical domains. However, FL and SL suffer from key limitations -- FL imposes substantial computational demands on clients, while SL leads to prolonged training times. To overcome these challenges, SplitFed Learning (SFL) was introduced as a hybrid approach that combines the strengths of FL and SL. Despite its advantages, SFL inherits scalability, performance, and security issues from SL. In this paper, we propose two novel frameworks: Sharded SplitFed Learning (SSFL) and Blockchain-enabled SplitFed Learning (BSFL). SSFL addresses the scalability and performance constraints of SFL by distributing the workload and communication overhead of the SL server across multiple parallel shards. Building upon SSFL, BSFL replaces the centralized server with a blockchain-based architecture that employs a committee-driven consensus mechanism to enhance fairness and security. BSFL incorporates an evaluation mechanism to exclude poisoned or tampered model updates, thereby mitigating data poisoning and model integrity attacks. Experimental evaluations against baseline SL and SFL approaches show that SSFL improves performance and scalability by 31.2% and 85.2%, respectively. Furthermore, BSFL increases resilience to data poisoning attacks by 62.7% while maintaining superior performance under normal operating conditions. To the best of our knowledge, BSFL is the first blockchain-enabled framework to implement an end-to-end decentralized SplitFed Learning system.
Key Contributions
- Sharded SplitFed Learning (SSFL) that distributes the SL server workload across parallel shards, improving scalability by 85.2% and performance by 31.2% over baseline SFL
- Blockchain-enabled SplitFed Learning (BSFL) that replaces the centralized FL server with a committee-driven blockchain consensus mechanism for decentralized, tamper-resistant aggregation
- An evaluation mechanism in BSFL to detect and exclude poisoned or tampered model updates, achieving 62.7% greater resilience to data poisoning attacks than baseline SFL
🛡️ Threat Analysis
BSFL explicitly defends against data poisoning and model integrity attacks from malicious clients submitting corrupted updates in a federated/split learning setting — a canonical Byzantine-fault-tolerant FL defense. The paper evaluates resilience to poisoning attacks (62.7% improvement), and the blockchain consensus mechanism functions as a robust aggregation/poisoned-update exclusion scheme, squarely within ML02's scope for Byzantine-fault-tolerant FL protocols.