Split learning for health: Distributed deep learning without sharing raw patient data
read the original abstract
Can health entities collaboratively train deep learning models without sharing sensitive raw data? This paper proposes several configurations of a distributed deep learning method called SplitNN to facilitate such collaborations. SplitNN does not share raw data or model details with collaborating institutions. The proposed configurations of splitNN cater to practical settings of i) entities holding different modalities of patient data, ii) centralized and local health entities collaborating on multiple tasks and iii) learning without sharing labels. We compare performance and resource efficiency trade-offs of splitNN and other distributed deep learning methods like federated learning, large batch synchronous stochastic gradient descent and show highly encouraging results for splitNN.
This paper has not been read by Pith yet.
Forward citations
Cited by 34 Pith papers
-
Image Prompt Reconstruction Attacks on Distributed MLLM Inference Frameworks
First study of image prompt reconstruction attacks on distributed MLLM inference, proposing MPAA for pixel-level and IEDA for semantic reconstruction with 100% embedding extraction accuracy on four model families.
-
Do Vision Models Truly Forget? New Findings from Representation-Level Certification of Visual Unlearning in Vertical Federated Learning
Mirage auditing reveals that VFL unlearning methods passing output-level checks still retain substantial class structure in representations across multiple datasets and baselines.
-
WISV: Wireless-Informed Semantic Verification for Distributed Speculative Decoding in Device-Edge LLM Inference
WISV uses a channel-aware semantic acceptance policy on hidden representations to boost accepted sequence length by up to 60.8% and cut interaction rounds by 37.3% in distributed speculative decoding, with under 1% ac...
-
Mixture of Predefined Experts: Maximizing Data Usage on Vertical Federated Learning
Split-MoPE integrates split learning with predefined-expert routing to maximize usable data in vertical federated learning under sample misalignment, delivering state-of-the-art accuracy in one communication round plu...
-
HOSL: Hybrid-Order Split Learning for Memory-Constrained Edge Training
HOSL reduces client memory up to 3.7x versus full first-order split learning while staying within 0.20-4.23% accuracy on OPT models by pairing client zeroth-order estimation with server first-order optimization.
-
Pattern-Calibrated Multimodal Prediction under Blockwise Missingness
MOSAIC learns overlap-aware shared-specific representations, fits a first-stage predictor on overlapping data, and calibrates the gap using target-pattern samples, with non-asymptotic error bounds decomposing overlap ...
-
TL++: Accuracy and Privacy Preserving Traversal Learning for Distributed Intelligent Systems
TL++ recovers centralized mini-batch gradients via virtual batches in split learning and adds secret sharing for cut-layer tensors, achieving 91.41% accuracy on CIFAR-10 with 13x lower communication than full-model sync.
-
Why SGD is not Brownian Motion: A New Perspective on Stochastic Dynamics
SGD is reformulated via a master equation from discrete updates, producing a discrete Fokker-Planck equation that predicts non-stationary variance growth proportional to learning rate in flat Hessian directions.
-
Towards Large Model Feature Coding
LaMoFCBench is a new benchmark covering 4 categories and 16 scenarios that exposes misalignment between mainstream feature codecs and the heterogeneous statistics of large-model activations.
-
Do Vision Models Truly Forget? New Findings from Representation-Level Certification of Visual Unlearning in Vertical Federated Learning
Mirage auditing framework reveals that VFL unlearning methods passing output-level certification retain substantial class structure in representations, with no method achieving high utility plus both output and repres...
-
Q-LocalAdam: Memory-Efficient Client-Side Adaptive Optimization for Edge Federated Learning
Q-LocalAdam reduces optimizer memory by 3.37x via tailored 8-bit quantization for Adam states while maintaining or improving accuracy under data heterogeneity in edge federated learning.
-
Federated Imputation under Heterogeneous Feature Spaces
FedHF-Impute enables federated imputation across heterogeneous feature spaces by using a shared global feature graph and message passing for indirect cross-client knowledge transfer, reporting RMSE gains on SECOM and ...
-
Keyed Nonlinear Transform: Lightweight Privacy-Enhancing Feature Sharing for Medical Image Analysis
KNT applies key-conditioned nonlinear obfuscation to split-inference features, cutting re-identification AUC from 0.635 to 0.586 with 0.15 ms overhead and under 1 pp accuracy loss.
-
SplitFed-CL: A Split Federated Co-Learning Framework for Medical Image Segmentation with Inaccurate Labels
SplitFed-CL improves segmentation performance in privacy-preserving federated settings by having a global teacher refine unreliable local labels via weighted student-teacher correction, consistency regularization, and...
-
Modulated learning for private and distributed regression with just a single sample per client device
Introduces modulated learning for private distributed regression allowing one sample per client via calibrated noise injection on samples and aggregation of transformed representations to achieve unbiased gradients in...
-
HARMONY: Bridging the Personalization-Generalization Gap by Mitigating Representation Skew in Heterogeneous Split Federated Learning
HARMONY mitigates representation skew in heterogeneous hybrid split federated learning via meta-learning to simulate diverse extractors and server-side contrastive learning to align features, delivering up to 43% accu...
-
Networked Information Aggregation for Binary Classification
Sequential prediction passing on DAGs for logistic regression yields O(M/sqrt(D)) excess loss when M-agent windows cover all features, with Omega(k/D) lower bound identifying depth as the fundamental limit.
-
Application-Aware Twin-in-the-Loop Planning for Federated Split Learning over Wireless Edge Networks
TiLP integrates network, training, and task sub-twins into a digital twin and uses receding-horizon cross-entropy planning with actor-critic guidance to jointly optimize resource allocation in federated split learning...
-
Efficient Federated RLHF via Zeroth-Order Policy Optimization
Par-S²ZPO matches centralized RLHF sample complexity while converging faster in policy updates and outperforming FedAvg on MuJoCo tasks.
-
PrivaDE: Privacy-preserving Data Evaluation for Blockchain-based Data Marketplaces
PrivaDE is a privacy-preserving protocol for jointly computing data utility scores in ML using secure computation, with optimizations for efficiency and blockchain integration via smart contracts.
-
Recovering Input Text from Hidden States: Study of Gradient-Based Inversion of Decoder-Only Language Models
Continuous optimization recovers input text from GPT-2 last-layer hidden states at up to 97.5% exact match on short prompts, with failures concentrated on high-frequency function words.
-
AC$^2$P$^2$SL: Adaptive Communication-Computation Pipeline Parallel Split Learning over Edge Networks
AC²P²SL pipelines communication and computation across micro-batches in split learning, jointly optimizes splits and pre-allocation under constraints, and adds adaptive re-allocation for dynamic UEs to reduce overall ...
-
Adaptive Joint Compression and Synchronisation in Federated Split Learning for IoT Rainfall Prediction
A latency-driven scheduler jointly tunes activation compression and synchronization frequency in federated split learning, delivering up to 87% payload reduction and 54% less synchronization traffic on rainfall predic...
-
Defense Against Prompt Inversion Attacks: An Information-Theoretic Approach for LLM Collaborative Inference
Develops an information-theoretic defense using privacy adapters and low-dimensional bottlenecks to minimize mutual information for protecting prompts in collaborative LLM inference, with claimed theoretical bounds an...
-
QSplitFL: Capability Aware Deep Q-Learning for Optimal Split Point Selection in Split Federated Learning
QSplitFL is a DQN framework that selects split points in split federated learning from hardware metrics with a decayed loss-drop reward and committee voting, reporting faster convergence and higher accuracy than basel...
-
LightSplit: Practical Privacy-Preserving Split Learning via Orthogonal Projections
LightSplit uses non-invertible orthogonal projections as an information bottleneck in split learning to reduce transmitted dimensionality by 32x while retaining more than 95% accuracy and limiting reconstruction risk.
-
Modulated learning for private and distributed regression with just a single sample per client device
Single-sample clients add one calibrated noisy perturbation to their data point and share transformed representations, allowing the server to recover unbiased gradients for private distributed regression.
-
SplitFT: An Adaptive Federated Split Learning System For LLMs Fine-Tuning
SplitFT adapts cut-layer selection and reduces LoRA rank per client in federated split learning to improve efficiency and performance when fine-tuning LLMs on heterogeneous devices and data.
-
A Survey on Split Learning for LLM Fine-Tuning: Models, Systems, and Privacy Optimizations
A survey that introduces a unified training pipeline and taxonomizes split learning approaches for LLM fine-tuning across model, system, and privacy dimensions.
-
FedProxy: Federated Fine-Tuning of LLMs via Proxy SLMs and Heterogeneity-Aware Fusion
FedProxy replaces weak adapters with a proxy SLM for federated LLM fine-tuning, outperforming prior methods and approaching centralized performance via compression, heterogeneity-aware aggregation, and training-free fusion.
-
Secure and Privacy-Preserving Vertical Federated Learning
Three optimized MPC protocols for privacy-preserving vertical federated learning that support global and global-local updates while reducing computation versus naive full-MPC delegation.
-
Reproducibility in Machine Learning for Health
Systematic evaluation of over 100 ML4H papers finds poorer reproducibility than other ML fields, driven by limited data and code access, and offers recommendations to data providers, publishers, and researchers.
-
Towards Value-Constrained Credit Assignment in Fully Delegated AI Cooperatives
Proposes value-constrained credit assignment via gradient filtering and traversal learning for fully delegated AI cooperatives.
-
Split and Aggregation Learning for Foundation Models Over Mobile Embodied AI Network (MEAN): A Comprehensive Survey
The paper surveys split and aggregation learning for foundation models in 6G networks to improve efficiency, resource use, and data privacy in distributed AI.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.