Hanieh Totonchi Asl, Ebrahim Nouri, Zhuoran Li, Danella Zhao · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.27431 · cited by 0 works (OpenAlex) · open-access PDF
A interligação de dados de operação da BERNet (PPML) oferece um caminho promissor para trazer serviços de IA poderosos para a borda sem expor dados de usuário sensíveis. No entanto, sua implantação prática permanece restrita pelo alto custo de comunicação, computação e dados de movimento de computação segura multi-parte (MPC), particularmente durante a inferência não linear. Os sistemas MPC existentes geralmente desconectam a geração aleatória independente de entrada correlata da execução on-line, criando uma pressão de memória substancial para buffer dados criptográficos de uso único. Além disso, o armazenamento limitado pode forçar um padrão de execução de stop-and-refill que interrompe repetidamente o pipeline para reabastecer esses tampões, trazendo a geração aleat
Machine translation (Meta M2M100); the English original follows.
Privacy-preserving machine learning (PPML) offers a promising path toward bringing powerful AI services to the edge without exposing sensitive user data. However, its practical deployment remains constrained by the high communication, computation, and data-movement costs of secure multi-party computation (MPC), particularly during nonlinear inference. Existing MPC systems typically decouple input-independent correlated randomness generation from online execution, creating substantial memory pressure to buffer single-use cryptographic data. Moreover, limited storage can force a stop-and-refill execution pattern that repeatedly stalls the pipeline to replenish these buffers, bringing randomness generation back onto the critical path. We propose Interflow, an algorithm-hardware co-design that rethinks this separation by batch-wise processing and pipelining correlated-randomness generation with online execution: while the current batch consumes pre-generated randomness to execute secure nonlinear operations, correlation generation proceeds concurrently for the next batch. By leveraging interleaved execution, efficient data movement, and hardware acceleration, the design transforms preprocessing from a serialized, storage-intensive offline phase into a continuously overlapped pipeline that generates correlated randomness closer to the point of consumption, reducing memory footprint and data movement overhead. Implemented on the Zynq UltraScale+ ZCU102, Interflow improves nonlinear evaluation throughput by up to 1.55x for CNN operators and 1.48x for transformer operators. It achieves up to 1.26x speedup for ResNet-50 and up to 1.21x speedup for BERT-base, and reduces memory footprint by up to 37%, compared with baseline frameworks.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Computer science, Speedup, Randomness, Pipeline (software), Memory footprint, Throughput
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.27431, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange