Dynamic Domains Research Articles

Formal contracts mitigate social dilemmas in multi-agent reinforcement learning

Multi-agent Reinforcement Learning (MARL) is a powerful tool for training autonomous agents acting independently in a common environment. However, it can lead to sub-optimal behavior when individual incentives and group incentives diverge. Humans are remarkably capable at solving these social dilemmas. It is an open problem in MARL to replicate such cooperative behaviors in selfish agents. In this work, we draw upon the idea of formal contracting from economics to overcome diverging incentives between agents in MARL. We propose an augmentation to a Markov game where agents voluntarily agree to binding transfers of reward, under pre-specified conditions. Our contributions are theoretical and empirical. First, we show that this augmentation makes all subgame-perfect equilibria of all Fully Observable Markov Games exhibit socially optimal behavior, given a sufficiently rich space of contracts. Next, we show that for general contract spaces, and even under partial observability, richer contract spaces lead to higher welfare. Hence, contract space design solves an exploration-exploitation tradeoff, sidestepping incentive issues. We complement our theoretical analysis with experiments. Issues of exploration in the contracting augmentation are mitigated using a training methodology inspired by multi-objective reinforcement learning: Multi-Objective Contract Augmentation Learning. We test our methodology in static, single-move games, as well as dynamic domains that simulate traffic, pollution management, and common pool resource management.

Signal detection analysis of affective prosody recognition in forensic inpatients who have committed sexual offenses

ABSTRACT Socio-affective functioning, or the way we interact and relate to others, is one of the four dynamic sexual recidivism risk domains. Accurately recognizing emotions enables the inference of mental and affective states supporting social adaptation. As little attention has been paid to affective prosody recognition in forensic inpatients who have committed sexual offenses (FICSOs), this study assessed the accuracy and sensitivity scores of 111 male participants assigned into three groups: FICSOs (n = 35), forensic inpatients who have committed non-sexual offenses (FICNSOs, n = 26) and community members (CoM, n = 50). Collected data also include response bias, emotion labeling reflection time, task easiness and task easiness reflection time. Using non-parametric group comparisons (Kruskal–Wallis H and Mann–Whitney U), results highlight, overall, a pervasive impairment of affective prosody recognition in FICSOs and FICNSOs compared to CoM. However, there was no difference in disgust sensitivity scores between FICSOs and CoM. FICSOs and FICNSOs took significantly longer than CoM to select an emotional label, especially for happiness. In addition, a metacognitive impairment was found in FICSOs and FICNSOs as they found the task significantly easier than CoM while being less sensitive.

A dynamic mid-crustal magma domain revealed by the 2023 to 2024 Sundhnúksgígar eruptions in Iceland.

Mid-crustal magma domains are the source of many basaltic eruptions. Lavas from individual eruptions are often chemically homogeneous, suggesting that they derive from single, well-mixed magma reservoirs. The 2023 to 2024 eruptions at Sundhnúksgígar in the Svartsengi volcanic system in Iceland provide an opportunity to observe the behavior of a mid-crustal magma domain at high spatial and temporal resolution by detailed sampling and geochemical characterization. We observed substantial mantle-derived geochemical variability in the products erupted in the first hours of the December 2023 and January, February, and March to May 2024 eruptions, indicating that the eruptions derived from multiple magma reservoirs, which mineral-melt equilibration pressures place in the mid crust. The unusual presence of geochemical heterogeneity in the mid-crustal magma domain provides insights into how dynamic and complex mid-crustal magma domains can be.

Cryo-EM Structure of Phospholipase Cε Defines N-terminal Domains and their Roles in Activity.

Phospholipase Cε (PLCε) increases intracellular Ca 2+ and protein kinase C (PKC) activity in the cardiovascular system in response to stimulation of G protein coupled receptors (GPCRs) and receptor tyrosine kinases (RTKs). The ability of PLCε to respond to these diverse inputs is due, in part, to multiple, conformationally dynamic regulatory domains. However, this heterogeneity has also limited structural studies of the lipase to either individual domains or its catalytic core. Here, we report the 3.9 Å reconstruction of the largest fragment of PLCε to date in complex with an antigen binding fragment (Fab). The structure reveals that PLCε contains a pleckstrin homology (PH) domain and four tandem EF hands, including subfamily-specific insertions and intramolecular interactions with the catalytic core. The structure, together with a model of the holoenzyme, suggest that part of the N-terminus and PH domain form a continuous surface that could engage cytoplasmic leaflets of the plasma and perinuclear membranes, contributing to activity. Functional characterization of this surface confirm it is critical for maximum basal and G protein-stimulated activities. This study provides new insights into the autoinhibited, basal conformation of PLCε and the first mechanistic insights into how it engages cellular membranes for activity.

Genetic requirement of dact1/2 to regulate noncanonical Wnt signaling and calpain 8 during embryonic convergent extension and craniofacial morphogenesis.

Wnt signaling plays crucial roles in embryonic patterning including the regulation of convergent extension during gastrulation, the establishment of the dorsal axis, and later, craniofacial morphogenesis. Further, Wnt signaling is a crucial regulator of craniofacial morphogenesis. The adapter proteins Dact1 and Dact2 modulate the Wnt signaling pathway through binding to Disheveled. However, the distinct relative functions of Dact1 and Dact2 during embryogenesis remain unclear. We found that dact1 and dact2 genes have dynamic spatiotemporal expression domains that are reciprocal to one another suggesting distinct functions during zebrafish embryogenesis. Both dact1 and dact2 contribute to axis extension, with compound mutants exhibiting a similar convergent extension defect and craniofacial phenotype to the wnt11f2 mutant. Utilizing single-cell RNAseq and an established noncanonical Wnt pathway mutant with a shortened axis ( gpc4 ), we identified dact1/2 specific roles during early development. Comparative whole transcriptome analysis between wildtype and gpc4 and wildtype and dact1/2 compound mutants revealed a novel role for dact1/2 in regulating the mRNA expression of the classical calpain capn8 . Over-expression of capn8 phenocopies dact1/2 craniofacial dysmorphology. These results identify a previously unappreciated role of capn8 and calcium-dependent proteolysis during embryogenesis. Taken together, our findings highlight the distinct and overlapping roles of dact1 and dact2 in embryonic craniofacial development, providing new insights into the multifaceted regulation of Wnt signaling.

A variational-based non-smooth contact dynamics approach for the seismic analysis of historical masonry structures

A variational formulation of the non-smooth contact dynamics method is proposed to address the dynamic response of historical masonry structures modeled as systems of 3D rigid blocks and subjected to ground excitation. Upon assuming a unilateral-frictional contact law between the blocks, the equations of motions are formulated in a time-discrete impulse theorem format in the unknown block velocities and contact impulses. The variational structure of the problem to be solved at each time step is proven. On that basis, the numerical method requires at each time step to perform a collision detection that identifies antagonist contact points based on the given structural configuration, to solve a second-order conic programming problem that outputs block velocities and contact impulses, and to update the structural configuration for the solution to advance in time. As a merit of the formulation, large-scale problems can be robustly and efficiently addressed thanks to the convex setting of the time-step optimization problem. Numerical results are presented to test the computational performances of the proposed approach. Benchmark problems provide numerical evidence that the formulation is consistent with event-driven solutions based on the classical Housner impact model. The dynamic response, failure domains, and fragility functions of real-size masonry structures are then explored under ground impulse or earthquake excitation. The obtained results prove the reliability of the present computational method for the dynamic analysis and seismic assessment of historical masonry constructions of engineering interest.

Highly tough, crack‐resistant and self‐healable piezo‐ionic skin enabled by dynamic hard domains with mechanosensitive ionic channel

AbstractRobust and reliable piezo‐ionic materials that are both crack resistant and self‐healable like biological skin hold great promise for applications inflexible electronics and intelligent systems with prolonged service lives. However, such a combination of high toughness, superior crack resistance, autonomous self‐healing and effective control of ion dynamics is rarely seen in artificial iontronic skin because these features are seemingly incompatible in materials design. Here, we resolve this perennial mismatch through a molecularly engineered strategy of implanting carboxyl‐functionalized groups into the dynamic hard domain structure of synthesized poly(urethane‐urea). This design provides an ultra‐high fracture energy of 211.27 kJ m−2 that is over 123.54 times that of tough human skin, while maintaining skin‐like stretchability, elasticity, and autonomous self‐healing with a 96.40% healing efficiency. Moreover, the carboxyl anion group allows the dynamic confinement of ionic fluids though electrostatic interaction, thereby ensuring a remarkable pressure sensitivity of 7.03 kPa−1 for the tactile sensors. As such, we successfully demonstrated the enormous potential ability of this skin‐like piezo‐ionic sensor for biomedical monitoring and robotic item identification, which indicates promising future uses in flexible electronics and human–machine interactions.

Design, Development and Testing of the JTech Bolt for Use in Static, Quasi-Static and Dynamic Domains

Design, Development and Testing of the JTech Bolt for Use in Static, Quasi-Static and Dynamic Domains

Robust and tunable hydrogels strengthened by dynamic amphiphilic ionic domains

Hydrogels have gained increasing attentions in wearable electronics, but endure notch-sensitive and poor mechanical behavior in real-world applications. Herein, a robust and tunable hydrogel network was constructed by involving a polymerizable amphiphilic ionic liquid in polymer chains through a facile and novel strategy without additional chemical crosslinkers. The ingeniously designed dynamic amphiphilic ionic domains that are tuned by the structural domain and functional domain are crucial for reinforcing the structural integrity and allowing the polymer chains to accomplish reversible unfolding and ordered migration. The robust and functional microstructures effectively prevent crack propagation, enhancing the fracture strain (1200–2390%), tensile stress (380–1560 kPa), toughness (5–15 MJ m−3), and conductivity (0.13–0.2 mS cm−1). The obtained hydrogel acts as an excellent mechanoresponsive sensor to perceive various external stimuli to sensitively detect diverse human activities including large body motions and subtle muscle movements via electrical signals. Additionally, this strategy shows good applicable potential for preparing high-performance hydrogels based on monomers containing active double bonds, overcoming the disadvantages of poor mechanical properties of traditional free radical polymerization gels. It is anticipated that the design of dynamic amphiphilic ionic domains will provide guidance for regulating the polymer network of hydrogels and show great application potentials in real-life scenarios.

Equilibrium optimizer-based harmony search algorithm with nonlinear dynamic domains and its application to real-world optimization problems

Harmony Search (HS) algorithm is a swarm intelligence algorithm inspired by musical improvisation. Although HS has been applied to various engineering problems, it faces challenges such as getting trapped in local optima, slow convergence speed, and low optimization accuracy when applied to complex problems. To address these issues, this paper proposes an improved version of HS called Equilibrium Optimization-based Harmony Search Algorithm with Nonlinear Dynamic Domains (EO-HS-NDD). EO-HS-NDD integrates multiple leadership-guided strategies from the Equilibrium Optimizer (EO) algorithm, using harmony memory considering disharmony and historical harmony memory, while leveraging the hidden guidance direction information from the Equilibrium Optimizer. Additionally, the algorithm designs a nonlinear dynamic convergence domain to adaptively adjust the search space size and accelerate convergence speed. Furthermore, to balance exploration and exploitation capabilities, appropriate adaptive adjustments are made to Harmony Memory Considering Rate (HMCR) and Pitch Adjustment Rate (PAR). Experimental validation on the CEC2017 test function set demonstrates that EO-HS-NDD outperforms HS and nine other HS variants in terms of robustness, convergence speed, and optimization accuracy. Comparisons with advanced versions of the Differential Evolution (DE) algorithm also indicate that EO-HS-NDD exhibits superior solving capabilities. Moreover, EO-HS-NDD is applied to solve 15 real-world optimization problems from CEC2020 and compared with advanced algorithms from the CEC2020 competition. The experimental results show that EO-HS-NDD performs well in solving real-world optimization problems.

Dynamic Wrinkling Instability of Elastic Films on Viscoelastic Substrates

Abstract The dynamic instability of stiff films on compliant substrates has received sustained attention due to the potential applications in flexible functional devices. Film/substrate system-based devices are increasingly utilized under dynamic conditions, including dynamic sensors, tunable optical components, anti-fouling surfaces, etc. To better design the dynamic characteristics of devices based on film/substrate systems, it is essential to establish a comprehensive dynamic model and find out the deterministic and non-deterministic instability domains of nonlinear dynamic wrinkling under time-varying biased loads. In this paper, a multi-level coupling time-varying parameter excitation dynamic model for films bonded on Kelvin viscoelastic substrates is developed. The damping effect on the nonlinear dynamic responses of wrinkled film/substrate systems under step, slope and biased sinusoidal axial time-varying excitations are analyzed. We revealed and analyze the nonlinear dynamic behavior of film/substrate systems, which are significantly influenced by the excitation frequency and viscous coefficients of substrates. Various response forms, such as excitation-following deterministic responses, chaotic responses, and double-period resonant responses, are observed. We analyzed the parametric excitation induced dynamic bifurcation of the time-varying energy barrier that causes the nonlinear dynamic phenomenon, and provided deterministic and non-deterministic dynamic response domains. Based on the theory and results, methods for generating responses of specific types are proposed, offering theoretical guidance for designing dynamic characteristics of devices based on film/substrate systems.

Calmodulin‐Based Dynamic Protein Hydrogels with Three Distinct Mechanical Stiffness

AbstractStimuli‐responsive hydrogels that leverage protein conformational changes are of significant interest in the design of dynamic materials applicable in a myriad of fields, such as drug delivery, actuators, biosensors, and microfluidics. The small calcium binding protein calmodulin (CaM), which undergoes three‐stage conformational changes upon successive binding with Ca2+ and specific ligands, offers a mechanism to create dynamic hydrogels with three distinct physical states. In this work, a CaM‐based recombinant protein hydrogel is engineered using [Ru(bpy)3]2+‐mediated photo‐crosslinking. This hydrogel displays the ability to reversibly increase its Young's modulus by 1.5‐fold and ∼7‐fold, respectively, upon binding with Ca2+ and subsequent interaction with trifluoperazine. The magnitude of stiffness changes is tunable by adjusting the length proportion of dynamic and static domains and modifying protein content. This tunable and reversible control over hydrogel mechanics is further utilized to engineer shape‐morphing materials, highlighting the versatile potential of this CaM‐based protein hydrogel for diverse applications.

Superconducting diode effect and interference patterns in kagome CsV3Sb5.

The interplay among frustrated lattice geometry, non-trivial band topology and correlation yields rich quantum states of matter in kagome systems1,2. A series of recent members in this family, AV3Sb5 (A = K, Rb or Cs), exhibit a cascade of symmetry-breaking transitions3, involving the 3Q chiral charge ordering4-8, electronic nematicity9,10, roton pair density wave11 and superconductivity12. The nature of the superconducting order is yet to be resolved. Here we report an indication of dynamic superconducting domains with boundary supercurrents in intrinsic CsV3Sb5 flakes. The magnetic field-free superconducting diode effect is observed with polarity modulated by thermal histories, suggesting that there are dynamic superconducting order domains in a spontaneous time-reversal symmetry-breaking background. Strikingly, the critical current exhibits double-slit superconductivity interference patterns when subjected to an external magnetic field. The characteristics of the patterns are modulated by thermal cycling. These phenomena are proposed as a consequence of periodically modulated supercurrents flowing along certain domain boundaries constrained by fluxoid quantization. Our results imply a time-reversal symmetry-breaking superconducting order, opening a potential for exploring exotic physics, for example, Majorana zero modes, in this intriguing topological kagome system.

Stress-shape misalignment in confluent cell layers.

In tissue formation and repair, the epithelium undergoes complex patterns of motion driven by the active forces produced by each cell. Although the principles governing how the forces evolve in time are not yet clear, it is often assumed that the contractile stresses within the cell layer align with the axis defined by the body of each cell. Here, we simultaneously measured the orientations of the cell shape and the cell-generated contractile stresses, observing correlated, dynamic domains in which the stresses were systematically misaligned with the cell body. We developed a continuum model that decouples the orientations of contractile stress and cell body. The model recovered the spatial and temporal dynamics of the regions of misalignment in the experiments. These findings reveal that the cell controls its contractile forces independently from its shape, suggesting that the physical rules relating cell forces and cell shape are more flexible than previously thought.

Modeling of Induction Motors and Variable Speed Drives for Multi-Domain System Simulations Using Modelica and the OpenIPSL Library

This paper introduces an innovative method for characterizing, implementing, and validating both three-phase and single-phase induction motor models, accompanied by a variable speed drive model. The primary goal is to investigate interactions between the electrical power grid and other dynamic domains (e.g., thermofluidic) that impact motor/load drive behavior. Our approach involves establishing a mechanical interface based on a physically meaningful equation linking motor torque/speed to the electrical model in the phasor domain. This allows seamless integration of diverse domain subsystems into a unified multi-domain model using Modelica v4.0.0 and the OpenIPSL library v3.0.1, overcoming co-simulation limitations. The proposed model, which requires only one Modelica-compliant tool for simulation, introduces additional dynamics through the mechanical interface, enabling explicit simulation of load disturbances based on constitutive physics. This deepens our understanding of dynamic interactions between the electrical power domain and other subsystems connected through the motor. We detail the modeled components using mathematical equations and textual descriptions, emphasizing the Modelica modeling approach. Simulation examples validate the implementation, demonstrating the multi-domain modeling capabilities of the newly developed components.

Sequestration of membrane cholesterol by cholesterol-binding proteins inhibits SARS-CoV-2 entry into Vero E6 cells

Membrane lipids and proteins form dynamic domains crucial for physiological and pathophysiological processes, including viral infection. Many plasma membrane proteins, residing within membrane domains enriched with cholesterol (CHOL) and sphingomyelin (SM), serve as receptors for attachment and entry of viruses into the host cell. Among these, human coronaviruses, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), use proteins associated with membrane domains for initial binding and internalization. We hypothesized that the interaction of lipid-binding proteins with CHOL in plasma membrane could sequestrate lipids and thus affect the efficiency of virus entry into host cells, preventing the initial steps of viral infection. We have prepared CHOL-binding proteins with high affinities for lipids in the plasma membrane of mammalian cells. Binding of the perfringolysin O domain four (D4) and its variant D4E458L to membrane CHOL impaired the internalization of the receptor-binding domain of the SARS-CoV-2 spike protein and the pseudovirus complemented with the SARS-CoV-2 spike protein. SARS-CoV-2 replication in Vero E6 cells was also decreased. Overall, our results demonstrate that the integrity of CHOL-rich membrane domains and the accessibility of CHOL in the membrane play an essential role in SARS-CoV-2 cell entry.

Highly stretchable, sensitive and healable epidermal electronics enabled by dynamic hard domains and multidimensional conductive fillers for human motion and biopotential sensing

Integration of self-healing ability and biomimetic structures into epidermal electronics is of great interest. Herein, mimicking the merits of human skin, dynamic hard domains and multidimensional conductive fillers were introduced into epidermal electronics. Firstly, aromatic disulfide bonds and H-bonds as dynamic hard domains from bis-(4-aminophenyl) disulfide and tetraethylene glycol were incorporated into polyurethane-urea (PU), and provided desired mechanical properties and self-healing capability. To enhance the strength and toughness, ZIF-67 nanoparticles with micro-pores and uncoordinated imidazole groups were introduced into PU matrix, producing excellent filler-matrix interfaces owing to the formation of high-density H-bonds and mechanical insertion. The optimized PU/ZIF-6710% composite exhibited high performance with a tensile strength of 4.15 MPa, an elongation at break of 1342.31 %, a toughness of 40.73 MJ m−3, a fracture energy of 157.75 kJ m−2, and healing efficiency of 93.25 %. Benefiting from the supramolecular PU with dynamic hard domains and multidimensional conductive fillers, the as-prepared multifunctional films exhibited excellent self-healing ability and sensing performance with a wide detection range (609 %), high sensitivity (3416.71), fast response time (65 ms), low detection limit (0.026 %), and excellent robustness (∼1000 cycles at 50 % strain). Different human motions and high-quality electrophysiological signals were successfully detected by the original and healable epidermal electrodes.

Continual Relation Extraction via Sequential Multi-Task Learning

To build continual relation extraction (CRE) models, those can adapt to an ever-growing ontology of relations, is a cornerstone information extraction task that serves in various dynamic real-world domains. To mitigate catastrophic forgetting in CRE, existing state-of-the-art approaches have effectively utilized rehearsal techniques from continual learning and achieved remarkable success. However, managing multiple objectives associated with memory-based rehearsal remains underexplored, often relying on simple summation and overlooking complex trade-offs. In this paper, we propose Continual Relation Extraction via Sequential Multi-task Learning (CREST), a novel CRE approach built upon a tailored Multi-task Learning framework for continual learning. CREST takes into consideration the disparity in the magnitudes of gradient signals of different objectives, thereby effectively handling the inherent difference between multi-task learning and continual learning. Through extensive experiments on multiple datasets, CREST demonstrates significant improvements in CRE performance as well as superiority over other state-of-the-art Multi-task Learning frameworks, offering a promising solution to the challenges of continual learning in this domain.

Mimicking Behaviors in Separated Domains (Abstract Reprint)

Devising a strategy to make a system mimic behaviors from another system is a problem that naturally arises in many areas of Computer Science. In this work, we interpret this problem in the context of intelligent agents, from the perspective of LTLf, a formalism commonly used in AI for expressing finite-trace properties. Our model consists of two separated dynamic domains, D_A and D_B, and an LTLf specification that formalizes the notion of mimicking by mapping properties on behaviors (traces) of D_A into properties on behaviors of D_B. The goal is to synthesize a strategy that step-by-step maps every behavior of D_A into a behavior of D_B so that the specification is met. We consider several forms of mapping specifications, ranging from simple ones to full LTLf, and for each, we study synthesis algorithms and computational properties.

Study of convergence speed of chaotic particle swarm optimization algorithm

AbstractThis study introduces the Chaotic Particle Swarm Optimization as an innovative variant of the traditional particle swarm optimization algorithm, addressing the issue of particle swarm optimization getting trapped in local minima with a low convergence characteristic during later iterations. Chaotic particle swarm optimization incorporates principles from chaos theory to enhance the swarm's exploration and exploitation capabilities. By introducing controlled chaotic behavior, particles exhibit more diverse and unpredictable movements in the search space, leading to improved global convergence and escape from local minima. The proposed method has been implemented and evaluated on benchmark problems to assess its effectiveness. The integration of chaos theory with particle swarm optimization offers promising opportunities for developing robust and efficient optimization techniques suitable for complex and dynamic problem domains in various real-world applications.