Researchers introduce Mid-Harness, a method that samples and verifies candidate actions at the model-harness boundary to improve execution reliability in terminal agents. This approach allocates test-time compute by verifying multiple options before forwarding one for execution, keeping the generator and harness unchanged.
The article introduces Human2Any, a method for human-to-robot transfer that utilizes constraint-aware compositional planning. No further details or body text are provided in the source.
Researchers present ShareMMU, an IOMMU design that allows multiple untrusted accelerators to securely reuse pre-translated addresses within a shared virtual address space. This approach mitigates side-channel risks associated with large Shared Translation Lookaside Buffers (TLBs) while maintaining high performance.
Researchers introduce Parallel Decoding Distillation (PDD), a simplified trajectory-based method to accelerate inference in diffusion and flow matching models. Unlike current approaches that rely on hard-to-optimize variational score distillation and adversarial losses, PDD predicts multiple denoising steps per network evaluation.
Nvidia researcher N. Horrocks introduces extended detector error models (EDEMs) designed to handle fault-tolerant quantum programs with stabilizer circuits and stochastic Pauli faults. These models allow error analysis to compose in sequence and parallel, mirroring the structure of physical circuit realizations.
Researchers propose Onyx, a cost-efficient approach to disk-oblivious approximate nearest neighbor (ANN) search that balances bandwidth and access count by inverting the design of state-of-the-art ORAM-ANN systems.
Researchers introduce Privatar, a framework that offloads avatar reconstruction from headsets to untrusted local devices while safeguarding against data interception. The system leverages domain-specific knowledge of avatar reconstruction to achieve provably private offloading with minimal performance cost.
Researchers from NVIDIA have introduced GPIR, a system that accelerates Private Information Retrieval (PIR) on GPUs by addressing the high server-side computation and memory traffic inherent in lattice-based protocols. The system employs a stage-aware hybrid execution model that dynamically switches between operation-level and stage-level kernels to maximize on-chip data reuse.
A new study presents a unified system-level characterization of secure multiparty computation (MPC) and fully homomorphic encryption (FHE) for privacy-preserving machine learning inference. The work evaluates two MPC variants—arithmetic/binary sharing conversion and function secret sharing—alongside FHE across multiple CNN and Transformer models.
A position paper by NVIDIA researchers outlines a vision for securing AI agents against indirect prompt injection attacks, where malicious instructions embedded in untrusted data trigger dangerous actions. The authors argue that effective defense requires moving beyond isolated model fixes to comprehensive system-level architectures.
Researchers have formalized inference-time denial-of-service (PI-DoS) attacks that exploit the high computational cost of explicit multi-step reasoning in Large Reasoning Models (LRMs). They present ReasoningBomb, a reinforcement-learning-based framework that trains an attacker to generate short natural prompts that drive victim models into pathologically long and often non-terminating reasoning traces.
Researchers have introduced ReasAlign, a model-level solution designed to improve safety alignment in Large Language Models (LLMs) against indirect prompt injection attacks. The approach incorporates structured reasoning steps to analyze user queries and detect conflicting instructions, ensuring the continuity of the user's intended tasks.