arXiv Machine Learning September 25, 2026 By Sudip Bhujel, Shanghao Shi, Ruiquan Huang, Ning Zhang, Yang Xiao

Temporal Gradient Inversion for Private Trajectory Reconstruction in Embodied Reinforcement Learning

Read the original on arXiv Machine Learning →

The paper introduces TRACE, an amortized temporal gradient‑inversion attack that reconstructs private observation‑action trajectories from per‑step policy gradients in embodied reinforcement‑learning agents. TRACE exploits cross‑time correlation between gradients and exact action recovery from policy‑head gradients, achieving high reconstruction quality (18.8 dB PSNR) and near‑perfect action recovery with minimal computation. The study demonstrates TRACE’s effectiveness across various neural architectures and input modalities, and suggests that protecting temporal gradient streams may require sequence‑aware privacy mechanisms.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Hugging Face Trending Papers

Sep 24

Temporal Gradient Inversion for Private Trajectory Reconstruction in Embodied Reinforcement Learning

Distributed learning in embodied reinforcement-learning agents offers a degree of privacy by retaining raw sensor data on-device and transmitting only policy gradients to the server. Yet temporal stru...

llms agents reinforcement-learning robotics multimodal safety

arXiv Machine Learning

Jul 7

Trajectory-Aware Information Matching for Multi-Step Gradient Inversion in Federated Learning

arXiv:2509. 22082v3 Announce Type: replace Abstract: Federated learning enables distributed information sharing and collaborative model training without exposing raw client data.

By Li Xia, Jing Yu, Zheng Liu, Sili Huang, Wei Tang, Xuan Liu

benchmarks safety

arXiv AI

Sep 11

Cascading Gradient Inversion via LT-Code Inspired Peeling in Federated Learning

The paper introduces a new gradient inversion attack for federated learning that leverages concepts from erasure‑correcting codes to recover entire training batches and their labels from a single FedSGD round. Unlike previous analytic attacks, this method can exactly reconstruct batches of up to 128 samples on ImageNet and achieves over 90% recovery even when the attacker actively manipulates the model. The study demonstrates that federated learning’s privacy leakage is far greater than previously estimated.

By Saeed Shariati, Mohsen Alambardar Meybodi

robotics benchmarks safety

arXiv AI

4d ago

A Sharp Transition in Data Reconstruction under Differential Privacy

arXiv:2609.37344v1 Announce Type: cross Abstract: Data reconstruction attacks have empirically been successful in recovering training samples from learned models, raising privacy concerns and motivat...

By Max Cairney-Leeming, Simone Bombari, Marco Mondelli

safety

arXiv Machine Learning

Jun 18

TIGER: Inverting Transformer Gradients via Embedding-Subspace Distance Optimization

arXiv:2606. 18312v1 Announce Type: cross Abstract: Federated learning allows multiple clients to jointly train a shared model by sending gradient updates to a central server while keeping raw inputs local.

By William Kalikman, Ivo Petrov, Dimitar I. Dimitrov, Martin Vechev

llms rag efficiency safety

arXiv AI

Jun 10

Test-time Adversarial Takeover: A Real-time Hijacking Interface against Robotic Diffusion Policies

arXiv:2606. 10371v1 Announce Type: cross Abstract: Diffusion-based action generation has become a foundational component of embodied AI, but its reliance on visual conditioning leaves deployed visuomotor policies vulnerable to adversarial manipulation.

By Zi Yin, Peilin Chai, Siyuan Huang, Zhanhao Hu

llms diffusion robotics safety

One email a morning, machine-written

Related stories

Temporal Gradient Inversion for Private Trajectory Reconstruction in Embodied Reinforcement Learning

Trajectory-Aware Information Matching for Multi-Step Gradient Inversion in Federated Learning

Cascading Gradient Inversion via LT-Code Inspired Peeling in Federated Learning

A Sharp Transition in Data Reconstruction under Differential Privacy

TIGER: Inverting Transformer Gradients via Embedding-Subspace Distance Optimization

Test-time Adversarial Takeover: A Real-time Hijacking Interface against Robotic Diffusion Policies