Science – Page 77 – Investment Policy

Decentralized Bandits: Surviving Bad Actors and Faulty Data

16.11.2025 by Ray Dalio

Decentralized algorithms—including DeMABAR, DRAA, resilient UCB, MA-BARBAT, IND-BARBAR, and IND-FTRL—demonstrate robustness against adversarial corruption when implemented within a centralized CMA2B framework.

A new algorithm, DeMABAR, offers robust performance in multi-agent learning scenarios where agents face both malicious attacks and unreliable information.

Hidden in Plain Sight: Strengthening Image Security with Combined Watermarking

16.11.2025 by Ray Dalio

Hybrid watermarking, when unassailed, demonstrates a resilience predicated on embedding imperceptible signals within data – a fragile equilibrium, as any intervention threatens to unravel the carefully constructed illusion and expose the hidden message.

A new analysis reveals that blending spatial and frequency domain techniques dramatically improves the resilience of digital watermarks embedded within images.

Flowing to New Heights: Liquid Metal Printing Advances Superconducting Circuits

16.11.2025 by Ray Dalio

The observed temperature-dependent frequency shift and quality factor of a printed EGaInSn resonator reveal a superconducting critical temperature, though subsequent cryogenic measurements demonstrate gallium segregation and trace dewetting—manifesting as increased surface roughness and constrictions—suggesting material instability despite initial superconductivity.

Researchers are using liquid metal printing to fabricate high-performance superconducting circuits, opening doors for more flexible and scalable cryogenic electronics.

Securing Industrial Data with Blockchain and Programmable Networks

16.11.2025 by Ray Dalio

A new architecture, Pk-IOTA, streamlines OPC UA certificate management by combining the IOTA blockchain with programmable data plane technology to bolster trust and reduce overhead in Industry 4.0 environments.

Shrinking the Signature: A New Look at Lattice-Based Security

15.11.2025 by Ray Dalio

A recent analysis refines the Bai-Galbraith signature scheme, optimizing its efficiency for post-quantum cryptography applications.

Boosting Quantum Security: Faster Rates for Practical Key Distribution

15.11.2025 by Ray Dalio

New optimization techniques dramatically improve the speed and accuracy of calculating secure key rates in quantum key distribution systems.

Weaving Robust Quantum Codes with Hypergraphs

15.11.2025 by Ray Dalio

Circuit performance of several $C_xR$ codes was evaluated under standard noise conditions, with results demonstrating performance comparable to surface code heuristics represented by dashed lines.

A new approach to quantum error correction utilizes cyclic hypergraph product codes to achieve performance competitive with state-of-the-art methods.

Frustrated Quantum States Shield Delicate Qubits

15.11.2025 by Ray Dalio

New research demonstrates how quantum frustration can act as a protective mechanism against decoherence in a novel qubit design.

Quantum Spin Control: Balancing Speed and Reliability

15.11.2025 by Ray Dalio

$Optimization trajectories, assessed across ten realizations with parameters $N=3$, $T=2$, and $L=8$, demonstrate that local control converges more rapidly and achieves a significantly lower final loss—approximately $2 \times 10^{-4}$—compared to global control, which plateaus around $1.1 \times 10^{-3}$, highlighting the efficiency gains achieved through localized optimization strategies.$

A new variational algorithm optimizes state transfer in complex spin chains, revealing fundamental limits between performance and noise resilience.

The Speed Limit of Quantum Scale-Up

15.11.2025 by Ray Dalio

$Decoding speed, crucial for quantum circuit execution, scales with the number of magic state injections—a relationship assessed at a processor operating point of $d=31$—and ultimately faces a communication latency bottleneck below $10^{-7}$, though the fastest decoder—a custom ASIC—can process circuits with $T\leq 3\times 10^{7}$ within an hour and those with $T\leq 2\times 10^{10}$ within a month, demonstrating a tangible limit to scaling beyond which architectural considerations dominate performance.$

New research reveals that decoding latency and communication bottlenecks pose significant challenges to building larger, more practical quantum computers.