PreprintNot peer reviewedFrom GoogleArtificial IntelligenceResearchers report a cheap decision model can take over routine agent choicesOriginal title: JevAgentHarness: Benchmarking, Deploying and Training Decision Models That Share Control with LLMsA preprint finds a small decision model, trained on an agent's own states, can safely replace many LLM calls to cut cost and latency.Added 3 Nov 2026
PreprintNot peer reviewedArtificial IntelligencePreprint benchmark finds VLMs struggle to justify actions in first-person videoOriginal title: NoRA: Evaluating Grounded Reasonableness in Visual First-person Normative Action ReasoningA preprint introduces NoRA, a benchmark testing whether vision-language models can propose and justify reasonable next actions from first-person video.Sichao Li et al. · Added 3 Nov 2026
PreprintNot peer reviewedFrom Kuaishou TechnologyArtificial IntelligencePreprint proposes low-cost way to decide when LLM agents should use learned skillsOriginal title: SkillScaffold: Scaffolding Agentic Reinforcement Learning with Policy-Adaptive Skill Credit Assignment and Selective InjectionA preprint introduces SkillScaffold, which estimates skill usefulness cheaply and only injects skills during training when they help.Added 3 Nov 2026
PreprintNot peer reviewedQuantum ComputingA preprint reports a framework to sharpen mid-circuit qubit readout using auxiliary qubitsOriginal title: Lifting: Quantum State Discrimination Framework for Mid-Circuit MeasurementsResearchers describe Lifting, a readout framework combining extra qubit signals and context-aware calibration to improve mid-circuit measurement accuracy on IBM hardware.Chi Zhang et al. · Added 3 Nov 2026
PreprintNot peer reviewedArtificial IntelligenceA preprint reports Agent-Reflex trains language agents to recover from noisy tool feedbackOriginal title: Agent-Reflex: Teaching Language Agents to Act ReflectivelyA preprint describes Agent-Reflex, a training framework that teaches language agents to reflect and recover when tools fail or users intervene.Guanting Dong et al. · Added 2 Nov 2026
PreprintNot peer reviewedFrom Unitary QuantumQuantum ComputingResearchers report entanglement between a trapped ion and a solid-state quantum memoryOriginal title: Heterogeneous entanglement between a trapped ion and a solid-state quantum memoryA preprint describes linking a trapped-ion qubit with a crystal-based quantum memory 75 metres apart via frequency-converted photons.Added 2 Nov 2026
PreprintNot peer reviewedArtificial IntelligenceA preprint benchmarks when fast ‘System-1’ models can replace reasoning LLMsOriginal title: System1-Arena: Which Decisions Can Be Left to System 1?A new benchmark finds fast, single-pass decision models can match reasoning LLMs on many tasks but lag on forecasting and long-horizon agent planning.Tianyu Chen et al. · Added 2 Nov 2026
PreprintNot peer reviewedFrom Telefonica Innovacion Digital, LogitechArtificial IntelligenceResearchers report a benchmark for ‘procedural hallucination’ in chat agentsOriginal title: DialFork: Going Beyond Factuality to Evaluate Procedural Hallucination in Conversational AgentsA preprint introduces DialFork, a benchmark testing whether conversational agents stay on process rather than just staying truthful.Added 2 Nov 2026
Journal article✓ Peer reviewedFrom Fujitsu Research of AmericaQuantum ComputingStudy maps complexity of building constraint-respecting mixers for QAOAOriginal title: Imposing Constraints on Driver Hamiltonians and Mixing Operators: From Theory to Practical ImplementationThe authors analyse when constraint-preserving quantum mixers can be built efficiently and test tailored mixers on 1-in-3 SAT.Added 2 Nov 2026
PreprintNot peer reviewedArtificial IntelligenceA preprint reports gradient geometry shifts during RLVR, motivating a rank-adaptive optimizerOriginal title: Spectral-Driven Adaptive-Rank Optimizer for Large Language Model Reinforcement Learning with Verifiable RewardsResearchers propose Ares, an optimizer that adapts its update rank to the shrinking gradient subspace seen during RLVR training of language models.Zhipeng Chen et al. · Added 2 Nov 2026
PreprintNot peer reviewedQuantum ComputingA preprint reports privacy-preserving quantum machine learning need not cost much speedOriginal title: On the Computational Complexity of Blind Quantum Machine LearningA preprint reports a new transpiler, MasonQ, that lets blind quantum computing keep most of quantum machine learning's speed advantages.Rex Fleur et al. · Added 2 Nov 2026
PreprintNot peer reviewedQuantum ComputingResearchers propose Gaussian model to simulate photon losses in optical quantum circuitsOriginal title: Realistic simulation of universal multiport interferometers with losses within a Gaussian formalismA preprint proposes a Gaussian-formalism method to simulate realistic photon losses in multiport optical interferometers without full Fock-space computation.Sebastiano Corli et al. · Added 2 Nov 2026
PreprintNot peer reviewedQuantum ComputingA preprint reports post-optimization fixes that shrink quantum chemistry circuit costsOriginal title: Post-Optimization Subspace Corrections to Quantum AnsätzeResearchers report that adding subspace corrections after optimizing quantum chemistry circuits can cut iteration counts and measurement costs substantially.Prince Frederick Kwao et al. · Added 2 Nov 2026
PreprintNot peer reviewedArtificial IntelligenceResearchers report a debate protocol that hunts down hallucinating AI agents via image testsOriginal title: Multi-agent Undercover Gaming: Hallucination Removal via Counterfactual Test for Multimodal ReasoningA preprint describes MUG, a multi-agent debate protocol that spots hallucinating agents using modified images, reporting gains on reasoning benchmarks.Dayong Liang et al. · Added 2 Nov 2026
PreprintNot peer reviewedFrom NORMAQuantum ComputingResearchers report a GPU engine that speeds up Pauli propagation simulationOriginal title: PADO-Pauli: A GPU-Accelerated Implementation of Pauli PropagationA preprint describes PADO-Pauli, a GPU-based Pauli propagation engine built for repeated training queries of variational quantum circuits.Added 1 Nov 2026
PreprintNot peer reviewedFrom Google Quantum AIQuantum ComputingPreprint reports stitched surface codes tested on a 105-qubit superconducting chipOriginal title: Stitched surface codes: planar concatenation through GHZ corridorsA preprint describes a planar concatenated surface-code design tested on a 105-qubit Willow processor, showing lower logical error at larger distance.Added 1 Nov 2026
PreprintNot peer reviewedArtificial IntelligenceResearchers report prompt-security results depend heavily on how they are measuredOriginal title: SoK: Systematizing LLM Prompt Security: Taxonomies, Datasets, and Unified Evaluation of Attacks and DefensesA preprint argues that jailbreak attack and defence results for large language models shift sharply with evaluation configuration.Hanbin Hong et al. · Added 1 Nov 2026
PreprintNot peer reviewedFrom OPPOArtificial IntelligenceResearchers propose MARFT, a reinforcement fine-tuning framework for multi-agent LLM systemsOriginal title: MARFT: Multi-Agent Reinforcement Fine-TuningA preprint presents MARFT, a framework for fine-tuning teams of LLM agents with reinforcement learning, reporting gains over a baseline method.Added 1 Nov 2026
PreprintNot peer reviewedArtificial IntelligencePreprint benchmark finds life-assist AI agents vulnerable to attacks 25-60% of the timeOriginal title: Risky-Bench: Evaluating Agentic Safety Risks in Real-World Life-Assist DeploymentsA preprint introduces Risky-Bench, a safety benchmark showing seven AI agents fail realistic safety tests at rates between 25% and 60%.Jingnan Zheng et al. · Added 1 Nov 2026
PreprintNot peer reviewedArtificial IntelligenceResearchers report shallow exits in depth-based LLM inference resist jailbreaks betterOriginal title: Exploring Jailbreak Vulnerability Across Depth-Based Dynamic Inference in Large Language ModelsA preprint finds that letting large language models answer from earlier layers can make them harder to jailbreak than using the full depth.Lass Najmuldeen et al. · Added 1 Nov 2026
