PreprintNot peer reviewedFrom Q-CTRLQuantum ComputingResearchers report quantum learning gain from a single squeezed light mode, no entanglement neededOriginal title: Quantum advantage in learning single mode bosonic channelsA preprint reports that a single squeezed-light mode can learn a quantum channel exponentially faster than classical probes, without entanglement.Added 26 Oct 2026
PreprintNot peer reviewedFrom UberArtificial IntelligenceResearchers report runtime guardrail cuts false-approved rate for AI agents in productionOriginal title: Persona Guardrail: A Production-Grade Defense Framework for Agentic SystemsA preprint describes Persona Guardrail, a deployed defense that checks agent inputs and outputs against defined functional boundaries.Added 25 Oct 2026
PreprintNot peer reviewedFrom Nokia Bell LabsQuantum ComputingPreprint proposes chip-based quantum memory design using lithium tantalate and ytterbium crystalsOriginal title: Scalable Hybrid Device Architecture on Thin-Film Lithium Tantalate for Long Distance Quantum Network Nodes with Atomic Frequency Comb Quantum MemoriesA preprint outlines a hybrid photonic chip architecture pairing thin-film lithium tantalate with ytterbium-doped crystals for quantum network memories.Added 25 Oct 2026
PreprintNot peer reviewedFrom Google Cloud AI ResearchArtificial IntelligenceResearchers report a living benchmark cuts event-leakage in LLM forecasting testsOriginal title: LEAF: A Living Benchmark for Event-Augmented ForecastingA preprint introduces LEAF, a continuously updated benchmark that checks whether language models can genuinely forecast using verified events.Added 25 Oct 2026
PreprintNot peer reviewedArtificial IntelligenceA preprint proposes a layered Kubernetes defence framework for AI agentsOriginal title: Containing the Autonomous Operator: A Defense-in-Depth Framework and Reference Architecture for Securing AI Agents on KubernetesThe preprint offers a seven-layer security design for AI agents on Kubernetes, but has no measured attack or overhead results yet.Simhadri Podala Narasimha · Added 25 Oct 2026
PreprintNot peer reviewedQuantum ComputingPreprint reports optimal query complexity for simulating quantum Lindblad dynamicsOriginal title: Query-optimal quantum simulation of Lindblad evolutionA preprint describes a quantum algorithm that closes a known gap between lower and upper bounds for simulating open quantum systems.Chunhao Wang et al. · Added 25 Oct 2026
PreprintNot peer reviewedQuantum ComputingPreprint reports electro-optic time lens fixes mismatched photon interferenceOriginal title: Quantum interference between spectral bandwidth mismatched photonsA preprint describes using an electro-optic time lens to restore quantum interference between photons of different spectral bandwidths without filtering losses.Jan Krzyżanowski et al. · Added 25 Oct 2026
PreprintNot peer reviewedFrom CiscoArtificial IntelligencePreprint reports small language models can police AI agents’ tool callsOriginal title: Toward SLM-based agentic task-tool intent matchingA preprint explores whether small language models can check if an AI agent's tool choice actually fits its task.Added 25 Oct 2026
PreprintNot peer reviewedFrom Quantum ScopesQuantum ComputingPreprint reports all-fiber telecom entangled-photon source built from off-the-shelf partsOriginal title: Broadband Telecom Entanglement from an All-Fiber Type-0 Sagnac SourceA preprint describes a low-cost fiber Sagnac source producing telecom-band entangled photon pairs with moderate entanglement quality.Added 25 Oct 2026
PreprintNot peer reviewedQuantum ComputingA preprint reports an analytical shortcut for comparing quantum error-correction codesOriginal title: A Cross-Platform Analysis of Practical Quantum Error Correction CodesResearchers propose a lightweight analytical framework to estimate logical error rates across quantum hardware and distributed QEC designs.Bryan Pan et al. · Added 25 Oct 2026
PreprintNot peer reviewedFrom SAPArtificial IntelligenceResearchers report locating clinical concept circuits inside open-weight language modelsOriginal title: Clinical Concept Centers in LLMsA preprint reports finding dedicated clinical concept circuits in eleven open models that causally shape clinical outputs.Added 25 Oct 2026
PreprintNot peer reviewedArtificial IntelligenceA preprint proposes metrics and defences against LLM prompt-injection and trojan attacksOriginal title: Evaluating and Improving the Robustness of Large Language Models to Input Sequence VariationsA preprint introduces robustness metrics and defence methods for large language models facing prompt injection, trojans and agentic-tool attacks.Narek Maloyan · Added 25 Oct 2026
PreprintNot peer reviewedQuantum ComputingPreprint extends Poincaré duality to quantum codes, enabling new logical gatesOriginal title: Poincaré Duality and Multiplicative Structures on Quantum CodesA preprint generalises Poincaré duality to sheaf-based quantum codes and uses it to build new fault-tolerant logical gates.Yiming Li et al. · Added 25 Oct 2026
PreprintNot peer reviewedQuantum ComputingResearchers report quantum speedup for estimating nested expectations like optimal stoppingOriginal title: Optimal Quantum Speedups for Repeatedly Nested Expectation EstimationA preprint proposes a quantum algorithm estimating repeatedly nested expectations with near-quadratic speedup over classical methods.Yihang Sun et al. · Added 25 Oct 2026
PreprintNot peer reviewedFrom IBMQuantum ComputingPreprint extends classification of magic-state distillation codes to length 54Original title: Classification of Generalised Triorthogonal Codes through Length 54A preprint reports a classification of generalised triorthogonal magic-state distillation codes up to length 54, finding 74 optimal protocols.Added 25 Oct 2026
PreprintNot peer reviewedArtificial IntelligencePreprint finds LLM membership-detection tools still struggle on a stricter benchmarkOriginal title: OLMo-Detect: A Multi-Stage, Confounder-Controlled Benchmark for Membership Inference on Large Language ModelsA preprint reports that current membership-inference attacks on LLMs achieve only limited accuracy on a new, carefully controlled benchmark called OLMo-Detect.Tao Shi et al. · Added 25 Oct 2026
PreprintNot peer reviewedQuantum ComputingPreprint reports replay-buffer redesign speeds up RL for quantum circuit optimisationOriginal title: Replay-buffer engineering for noise-aware quantum circuit optimizationA preprint reports that reworking how experience is stored and replayed can cut training cost and improve sample efficiency for RL-based quantum circuit design.Akash Kundu et al. · Added 25 Oct 2026
PreprintNot peer reviewedArtificial IntelligenceResearchers propose Suan, a gradient-level alignment method to curb LLM over-refusalOriginal title: Suan: Rectifying Direct Preference Safety Alignment in Large Language ModelsA preprint introduces Suan, a preference optimisation method the authors say improves safety without sacrificing helpfulness or diversity.Oleksandr Cherednichenko et al. · Added 25 Oct 2026
PreprintNot peer reviewedFrom ZenityArtificial IntelligenceA preprint reports adding a classify suffix boosts LLM safety-probe detection on unseen attacksOriginal title: Prompted to Discriminate: Generalizing Malicious-Input Probes in the WildA preprint reports that a short classification suffix improves activation-probe safety monitoring on attack types never seen in training.Added 25 Oct 2026
PreprintNot peer reviewedFrom ZuriQQuantum ComputingResearchers propose fast gate scheme for magnetic-field-resistant trapped-ion qubitsOriginal title: A magnetic-field insensitive gate set for trapped-ion nuclear spin qubitsA preprint outlines a microwave gate method letting nuclear spin qubits in Penning traps run fast while staying shielded from magnetic field noise.Added 25 Oct 2026
