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Not peer reviewed
From Kuaishou
Artificial Intelligence

Researchers report latent-token reasoning speeds up LLM-based recommendation models

Original title: OneLatent: Latent Reasoning for Efficient Foundation Recommendation Models

A preprint describes OneLatent, which compresses chain-of-thought reasoning into latent tokens to make recommendation models faster without losing accuracy gains.

Added 29 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

A preprint reports a unified method for designing efficient LLM multi-agent systems

Original title: COLLATOR: Compositional Multi-Agent Orchestration with Counterfactual Reinforcement Learning

Researchers describe Collator, which jointly learns agent roles, capacity and dependencies for multi-agent LLM systems using counterfactual credit assignment.

Xudong Chen et al. · Added 29 Oct 2026

Preprint
Not peer reviewed
From IQM
Quantum Computing

Researchers propose stochastic route to quantum polynomial transforms without coherent encoding

Original title: Quantum Channel Polynomial Processing

A preprint describes a sampling-based quantum algorithm that trades circuit complexity for repeated runs when applying polynomials of a Hamiltonian.

Added 29 Oct 2026

Preprint
Not peer reviewed
Quantum Computing

A preprint reports a distributed quantum algorithm for solving large linear systems

Original title: Distributed Variational Quantum Linear Solver

Researchers describe splitting linear-system matrices across several NISQ quantum computers, each solving a block and coordinating classically.

Tong Shen et al. · Added 29 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

A preprint reports a defence that hides multi-agent LLM communication topology

Original title: Concealing LLM-Based Multi-Agent Topology via Phantom Structure Injection

Researchers propose Mirage, a framework hiding the real communication structure of LLM multi-agent systems from black-box inference attacks.

Longzhu He et al. · Added 29 Oct 2026

Preprint
Not peer reviewed
Quantum Computing

Preprint proposes general synthesis method for multivariable quantum polynomial circuits

Original title: Quantum Algorithms for Multivariable Polynomial Transformations: From Efficient Synthesis to Quantum Channel Transformations

A preprint presents a constructive method for turning multivariable polynomials into quantum circuits, extending single-variable techniques to several noncommuting matrices.

Zheyu Shen et al. · Added 29 Oct 2026

Preprint
Not peer reviewed
From Microsoft Quantum
Quantum Computing

Researchers report quantum PDE solver that avoids exponential scaling with dimension

Original title: Breaking the Curse of Dimensionality in Quantum PDE Solvers via Gevrey Regularity

A preprint describes a quantum algorithm using function smoothness classes to solve linear PDEs without the usual exponential cost in dimension.

Added 29 Oct 2026

Preprint
Not peer reviewed
From KUPAS
Artificial Intelligence

Preprint describes platform turning expert tacit knowledge into agent-usable experience corpora

Original title: KUPAS MASTER: Distilling the Tacit Expertise of Master Practitioners into Agent-Ready Experience Corpora

A preprint reports a platform that converts practitioners' tacit know-how into structured corpora, lifting agent scores from 70.63 to 89.58.

Added 28 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

Preprint finds multi-agent AI debate rarely beats budget-matched sampling

Original title: Beyond Symmetric Agents: Cognitive Diversity and Multi-Agent Debate in Small Language Models

A preprint reports that debate among small language models offers little advantage over simple budget-matched sampling methods.

Leonardo Ferreira et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Quantum Computing

Preprint proposes one-time memory scheme resistant to early fault-tolerant quantum attacks

Original title: Beyond NISQ Assumptions: One-time Memory in the Classically Accessible Random-Oracle Model

A preprint reports a one-time memory protocol using BB84 states, secure under a model that limits adversaries to classical oracle queries.

Boyang Chen et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Quantum Computing

Researchers trial gate-based quantum algorithms for codon optimisation

Original title: A Gate-Based Quantum Computing Framework for Codon Optimization

A preprint tests VQE, SVQE and QAOA on a small codon-optimisation problem, benchmarking against exact and annealing methods.

Fatemeh Ghasemi et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Quantum Computing

Researchers report evolved quantum circuits can shrink hybrid image classifiers’ parameters

Original title: Evolving Hybrid Quantum-Classical Architectures for Image Classification

A preprint reports that evolved quantum circuits matched classical CNN accuracy on image tasks using far fewer trainable parameters.

Devroop Kar et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

Researchers report extracting memorised API keys from black-box LLMs via output-only access

Original title: Practical Secrets Extraction against Black-box LLMs

A preprint describes a black-box method that recovers memorised API keys from commercial LLMs using only their text outputs.

Shiqian Zhao et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Quantum Computing

Preprint explores quantum edge in simulating disordered spin-model dynamics

Original title: Classical and Quantum Simulation of All-to-All Quantum Dynamics in Logarithmic Space

A preprint examines whether quantum computers can simulate disordered spin models in logarithmic space while staying hard for classical machines.

Maximilian Lutz et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
From Google DeepMind
Artificial Intelligence

Researchers report benign audio fine-tuning can break safety guardrails in Audio LLMs

Original title: Benign Fine-Tuning Breaks Safety Alignment in Audio LLMs

A preprint reports that fine-tuning Audio LLMs on seemingly harmless audio can sharply raise jailbreak success rates, depending on model architecture.

Added 28 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

Preprint finds tiny parameter subsets can trigger LLM collapse

Original title: Weight-Adjusted Gradients Reveal Parameter Importance and Failure Modes in LLMs

A preprint describes a metric combining weight and gradient signals that pinpoints a minuscule fraction of parameters capable of causing severe LLM breakdowns.

Shrestha Datta et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

Preprint reports AI agents can piece together hidden prompt injections from fragments

Original title: Divide and Inject: Can Agents Reconstruct an Indirect Prompt Injection from Fragments?

A preprint reports an attack that splits malicious instructions into fragments, letting agents reconstruct and act on them with higher success than prior methods.

Michael Lee et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Quantum Computing

Preprint proposes generalised cluster states for multi-qubit measurement-based quantum computing

Original title: Non-Clifford symmetry protected topological hyper-cluster states and multi-qubit universal measurement-based quantum computation

A preprint describes generalised hyper-cluster states built from C^N Z gates that could support multi-qubit gates in measurement-based quantum computation.

Motohiko Ezawa · Added 28 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

Researchers report advantage-clipping method speeds up and improves RL training for LLM reasoning

Original title: Towards Better Training Signal: Advantage Clipped Policy Optimization

A preprint reports a new policy-optimisation method, ACPO, that boosts accuracy and training speed over PPO and GRPO on maths reasoning benchmarks.

Ruichuan Huang et al. · Added 28 Oct 2026

Preprint
Not peer reviewed
Artificial Intelligence

Preprint ties LLM reasoning rewards to formally verified proof steps

Original title: Learning to Prove, Not Just to Answer: Reinforcement Learning from Formal Verification for Natural-Language Logical Reasoning

A preprint reports that rewarding formally verified proof steps, not just correct answers, improves both accuracy and proof validity in logical reasoning LLMs.

Qili Zhang et al. · Added 28 Oct 2026