Browse all research

Every Research Watch entry, newest first. Search, or narrow by topic, format, peer review, publisher and date.

1211 entries

Sort

The summaries are open to everyone. Sign in, free, to read the full entry.

Preprint
Not peer reviewed
Quantum Computing

Preprint proposes weak-coupling shortcut to cut classical cost of circuit knitting

Original title: Quantum circuit knitting using a weak-coupling approximation

A preprint describes a way to split quantum circuits across small devices that avoids exponential classical overhead when connections are weak.

John P. T. Stenger et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
Quantum Computing

Researchers report GPU-based classical solvers beat quantum methods on multi-objective MaxCut

Original title: Multi-Objective Optimization with Quantum-Annealing-Inspired Algorithms

A preprint reports that GPU-accelerated quantum-annealing-inspired algorithms outperform quantum hardware on multi-objective MaxCut benchmarks.

Xian-Zhe Tao et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
From Peak Quantum
Quantum Computing

Preprint reports asymmetric indium bonding for more precise 3D chip stacking in qubits

Original title: Accurate gap control of 3D-integrated superconducting quantum circuits using asymmetric indium architecture and multi-angle flip-chip bonding

Researchers report a flip-chip bonding method that cuts chip tilt and keeps qubit coherence intact in 3D-integrated circuits.

Added 3 Nov 2026

Preprint
Not peer reviewed
Quantum Computing

Survey maps how post-quantum crypto and quantum key distribution could be combined

Original title: Combining Post-Quantum Cryptography and Quantum Key Distribution: Taxonomy, Constructions, Proofs, Standards, Deployments and Open Problems

A preprint surveys hybrid PQC-QKD designs, proposing a taxonomy and cataloguing proofs, standards and deployments.

Mona Alotaibi et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
From Anthropic, QuEra Computing
Quantum Computing

Preprint reports LLM-guided search discovers smaller quantum circuits across eight tasks

Original title: Discovering state-of-the-art quantum circuits with large language models

A preprint describes an evolutionary LLM-and-evaluator framework that reportedly found more efficient reversible quantum circuits for most of eight tested tasks.

Added 3 Nov 2026

Preprint
Not peer reviewed
Quantum Computing

Preprint links loss sharpness to shot budget in quantum Max-Cut encoding

Original title: Shot- and Noise-Aware Pauli Correlation Encoding for Max-Cut Experiments on IBM Heron Quantum Processor

A preprint reports that tuning a quantum encoding's loss sharpness to match the measurement shot budget beats published schedules on hardware.

Achraf Boussahi · Added 3 Nov 2026

Preprint
Not peer reviewed
From Quantinuum
Quantum Computing

Researchers report detailed tomography of logical qubit gates on a trapped-ion processor

Original title: Logical gate set tomography

A preprint describes a new characterisation method for logical quantum gates, applied to Quantinuum's H1 trapped-ion processor.

Added 3 Nov 2026

Preprint
Not peer reviewed
Artificial Intelligence

Researchers report adversarial self-attack improves AI-based vulnerability patching

Original title: P2Patch: Exploit-Guided Vulnerability Repair

A preprint describes P2Patch, which has an AI exploiter attack AI-written patches, finding new bypasses in 41% of runs.

Mohsen Iranmanesh et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
Quantum Computing

Researchers report new way to pick quantum circuit splits for distributed computing

Original title: Entanglement-Capability-Aware Finite-Bond Circuit Cutting for Distributed Quantum Computing

A preprint proposes SACE, a two-stage method for choosing how to split quantum circuits across processors, and reports gains over existing techniques.

Weidong Liu et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
Quantum Computing

Researchers find transmon qubits reveal defects that stay excited for over a second

Original title: Superconducting circuits host two-level systems with lifetimes exceeding one second

A preprint reports that certain two-level defects near superconducting qubits keep their quantum state for more than a second, far outlasting the qubits themselves.

Mathieu Féchant et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
Artificial Intelligence

A preprint finds multimodal AI models rarely admit when none of the options fit

Original title: Measuring Epistemic Humility in Multimodal Large Language Models

A new benchmark finds that even top multimodal models seldom pick 'none of the above' when it is the correct answer.

Bingkui Tong et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
Artificial Intelligence

Preprint reframes a reported AI agent ‘escape’ as a reproducible access-control bug

Original title: From Training to Escalation: Forensic Analysis and Local Recreation of RL/RLHF-Induced Privilege Escalation in AI Agents (CVE-2026-65616)

A preprint argues that a widely reported AI agent breakout was really a reproducible token-forgery flaw in misconfigured repository servers.

Alicia Colmenero-Fernández · Added 3 Nov 2026

Preprint
Not peer reviewed
Quantum Computing

Differentiable method learns quantum noise parameters to cut decoding errors, study reports

Original title: Differentiable Maximum Likelihood Noise Estimation for Quantum Error Correction

A preprint reports a differentiable noise-estimation method that lowers logical error rates across several quantum error correction codes.

Hanyan Cao et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
From NVIDIA
Quantum Computing

Researchers report automated lattice-surgery compiler with parallel decoding for surface codes

Original title: Patchwork: automated lattice surgery and space-time partitioning for real-time decoding

A preprint describes Patchwork, a compiler automating surface-code lattice surgery with a parallel wavefront decoding scheme tested on ten logical qubits.

Added 3 Nov 2026

Preprint
Not peer reviewed
Quantum Computing

Researchers test post-quantum key swaps in cloud object storage under load

Original title: Benchmarking Post-Quantum Key Encapsulation in S3-Compatible Object Storage under Client Concurrency

A preprint reports that swapping to ML-KEM key exchange only slows an S3-style object store when clients are near full CPU load.

Revanth Reddy Lingala et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
Artificial Intelligence

Framework tags LLM reasoning steps to test strategy-reuse and RL mode-seeking claims

Original title: SHAPE of Thinking: Mathematical Heuristics for Analyzing and Improving LLM Reasoning

A preprint introduces SHAPE, a heuristic-tagging method for chain-of-thought, and uses it to probe reasoning shifts and RL training behaviour.

Jonghyun Song et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
Artificial Intelligence

Researchers report local LLM apps leak prompts despite on-device inference

Original title: The Illusion of Local Privacy: Confidentiality Boundary Failures in Consumer LLM Serving Systems

A preprint reports that running LLMs locally does not by itself keep prompts confidential, citing memory, storage and server-isolation gaps.

Youssef Hamdi Zafan Ibrahim et al. · Added 3 Nov 2026

Preprint
Not peer reviewed
From Yuanhui AI
Artificial Intelligence

A preprint reports LLMs often judge their own outputs inconsistently

Original title: VALCON: Benchmarking and Enhancing Value Consistency in Large Language Models via Dual-Role Variation

A preprint finds that large language models frequently fail to judge their own generated answers the way they produced them.

Added 3 Nov 2026

Preprint
Not peer reviewed
From Li Auto
Artificial Intelligence

Researchers report a unified vision-language model for embodied and agent tasks, tuned for edge chips

Original title: ME-VLM: A Unified VLM for Embodied Cognition and Agent Coordination

A preprint describes ME-VLM, a vision-language model merging embodied perception and agent skills, with edge-optimised deployment on Li Auto's M100 chip.

Added 3 Nov 2026

Preprint
Not peer reviewed
From Oracle AI, Turing Enterprise
Artificial Intelligence

New benchmark exposes gaps in how AI models find and read document evidence

Original title: VELD-QA: A Large-Scale Dataset for Visual Evidence Grounding and Question Answering in Long Documents

A preprint introduces a large human-authored benchmark testing whether vision-language models can locate and interpret visual evidence in long documents.

Added 3 Nov 2026