aiqbench analysis
This study challenges the assumption that sparsifying activations in collaborative inference inherently enhances privacy by demonstrating that transmission positions alone enable high-fidelity input reconstruction, particularly for sensitive data like faces. The findings indicate a significant shift in risk vectors where positional metadata becomes a critical vulnerability even when communication costs and task utility are minimized. Consequently, developers must treat these position indices as sensitive transmitted data requiring rigorous auditing rather than relying on sparsity as an automatic privacy safeguard.
A Privacy Study of Sparse Collaborative Inference
arXiv:2608.16236v1 Announce Type: new Abstract: Collaborative inference (CI) splits a model between an edge device and a server, whereby the client computes an intermediate activation, transmits it, and the server completes the computation. This raises two concerns, the communication cost of the transmission and the risk that it reveals private information about the input. Recent work reduces this cost by sparsifying activations and entropy-coding the result. Sparsity has also been argued to improve privacy, on the intuition that transmitting fewer values reveals less about the input. We test this claim by decomposing the sparse activation into the retained values and the set of positions they occupy, and by reconstructing inputs from each component in isolation. We find that sparsification reduces the leakage far less than it reduces the transmission cost, and that the remaining risk shifts to the positions, which prior analyses treat as side information for decoding. Across natural-image and face datasets, the positions alone constitute a serious privacy risk, enabling high-fidelity reconstructions and re-identification of individuals. The leakage from the positions persists even when both the transmission cost and the task utility are low. We conclude that the positions of sparse activations should be treated as sensitive transmitted data and audited carefully in the context of collaborative inference. Code is available at https://github.com/an7123/Privacy-Study-Sparse-CI.
Abstract:Collaborative inference (CI) splits a model between an edge device and a server, whereby the client computes an intermediate activation, transmits it, and the server completes the computation. This raises two concerns, the communication cost of the transmission and the risk that it reveals private information about the input. Recent work reduces this cost by sparsifying activations and entropy-coding the result. Sparsity has also been argued to improve privacy, on the intuition that transmitting fewer values reveals less about the input. We test this claim by decomposing the sparse activation into the retained values and the set of positions they occupy, and by reconstructing inputs from each component in isolation. We find that sparsification reduces the leakage far less than it reduces the transmission cost, and that the remaining risk shifts to the positions, which prior analyses treat as side information for decoding. Across natural-image and face datasets, the positions alone constitute a serious privacy risk, enabling high-fidelity reconstructions and re-identification of individuals. The leakage from the positions persists even when both the transmission cost and the task utility are low. We conclude that the positions of sparse activations should be treated as sensitive transmitted data and audited carefully in the context of collaborative inference. Code is available at this https URL.
Machine Learning (cs.LG)
arXiv:2608.16236 [cs.LG]
(or arXiv:2608.16236v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.16236 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 306 words
Related coverage
Convolutional Sparse Coding via the Locally Competitive Algorithm on Loihi 2
arXiv · Jun 9, 2026
Spike-Aware C++ INT8 Inference for Sparse Spiking Language Models on Commodity CPUs
arXiv · Jun 3, 2026
Accelerating Hierarchical Sparse Predictive Coding with Hybrid Amortized Inference
arXiv · Jul 28, 2026
FedMental: Evaluating Federated Learning for Mental Health Detection from Social Media Data
arXiv · Jun 5, 2026