arXiv subjects

Yong Tao

Publications and source records attributed to Yong Tao.

At least 19 recordsLinked to original sources

arXiv2026

TRACE: Transition-Aware Residual Control for Multi-Objective Materials Discovery

Multi-objective materials discovery with LLM agents is often limited not only by how many candidates can be proposed, but by how effectively each costly property evaluation informs the next search step. Existing agents mainly store evaluated candidates and their scores, so they know which materials succeeded but not which executable edits caused useful property changes. This makes local refinement difficult when objectives compete and an edit that improves one property may damage another. We propose TRACE, a transition-aware residual control framework that treats evaluated edits as the basic unit of feedback. TRACE records each local refinement as a parent-edit-child transition with observed property deltas, aggregates transition evidence to estimate reusable edit effects, and ranks future edits by their predicted ability to reduce the current candidate's remaining constraint violations while avoiding damage to already satisfied objectives. In a controlled same-backbone comparison, TRACE improves over LLEMA, the state-of-the-art LLM-agent baseline, raising macro-average hit rate from 18.13\% to 25.96\%.

cs.AI↗

arXiv2026

Neural Controlled Differential Equations for EMT-Level Surrogate Modeling of Grid-Forming Inverters

The application of artificial intelligence methods in power electronic converter modeling is becoming increasingly widespread, but existing applications still face many challenges, such as difficulties in multi-time-scale hybrid analysis and the lack of physics-aware evaluation criteria and constraints, resulting in poor performance. This paper proposes a Neural Controlled Differential Equation (Neural CDE) framework for learning continuous-time surrogate models of grid-forming inverters for electromagnetic transient (EMT) simulation, which relaxes the constraint of fixed sampling rates and enables multi-time-scale control analysis. Then, an affine-control formulation with dual slow/fast pathways is proposed to capture the hierarchical and multiscale behavior of converter dynamics, and a physics-inspired regularization method is utilized to enhance stability and coherence. Evaluated on EMT-generated trajectories, the model accurately reproduces transient responses, preserves effective damping and the dominant oscillatory characteristics, and maintains bounded long-horizon rollouts. The results show that Neural CDE-based component modeling offers a physically consistent surrogate modeling approach for EMT-level simulation studies.

cs.LG↗

arXiv2026

Margin-Adaptive Confidence Ranking for Reliable LLM Judgement

Jung et al. (2025) introduce a hypothesis testing framework for guaranteeing agreement between large language models (LLMs) and human judgments, relying on the assumption that the model's estimated confidence is monotonic with respect to human-disagreement risk. In practice, however, this assumption may be violated, and the generalization behavior of the confidence estimator is not explicitly analyzed. We mitigate these issues by learning a dedicated confidence estimator instead of relying on heuristic confidence signals. Our approach leverages simulated annotator diversity and a margin-based ranking formulation to explicitly model how confidently an LLM distinguishes between human-agreement and human-disagreement cases. We further derive generalization guarantees for this estimator, revealing a margin-dependent trade-off that informs the design of an adaptive estimator training procedure. When integrated into fixed-sequence testing, the learned confidence estimator yields improved ranking accuracy and empirically strengthens the monotonic relationship between confidence and disagreement risk, leading to higher success rates in satisfying target agreement levels across multiple datasets and judge models.

cs.LG↗

arXiv2025

The Decision Path to Control AI Risks Completely: Fundamental Control Mechanisms for AI Governance

Artificial intelligence (AI) advances rapidly but achieving complete human control over AI risks remains an unsolved problem, akin to driving the fast AI "train" without a "brake system." By exploring fundamental control mechanisms at key elements of AI decisions, we develop a systematic solution to thoroughly control AI risks, providing an architecture for AI governance and legislation with five pillars supported by six control mechanisms, illustrated through a minimum set of AI Mandates (AIMs). Three of the AIMs must be built inside AI systems and three in society to address major areas of AI risks: 1) align AI values with human users; 2) constrain AI decision-actions by societal ethics, laws, and regulations; 3) build in human intervention options for emergencies and shut-off switches for existential threats; 4) limit AI access to user resources to reinforce controls inside AI; 5) mitigate spillover risks like job loss from AI. We also highlight the differences in AI governance on physical AI systems versus generative AI. We discuss how to strengthen analog physical safeguards to prevent AI or AGI from circumventing core safety controls by exploiting AI's intrinsic disconnect from the analog physical world: AI's nature as pure software code run on chips controlled by humans, and the prerequisite that all AI-driven physical actions must be digitized. These findings establish a theoretical foundation for AI governance and legislation as the basic structure of a "brake system" for AI decisions. If implemented, these controls can rein in AI dangers as completely as humanly possible, removing large chunks of currently wide-open AI risks, substantially reducing overall AI risks to residual human errors.

cs.CY↗

arXiv2025

REOBench: Benchmarking Robustness of Earth Observation Foundation Models

Earth observation foundation models have shown strong generalization across multiple Earth observation tasks, but their robustness under real-world perturbations remains underexplored. To bridge this gap, we introduce REOBench, the first comprehensive benchmark for evaluating the robustness of Earth observation foundation models across six tasks and twelve types of image corruptions, including both appearance-based and geometric perturbations. To ensure realistic and fine-grained evaluation, our benchmark focuses on high-resolution optical remote sensing images, which are widely used in critical applications such as urban planning and disaster response. We conduct a systematic evaluation of a broad range of models trained using masked image modeling, contrastive learning, and vision-language pre-training paradigms. Our results reveal that (1) existing Earth observation foundation models experience significant performance degradation when exposed to input corruptions. (2) The severity of degradation varies across tasks, model architectures, backbone sizes, and types of corruption, with performance drop varying from less than 1% to over 20%. (3) Vision-language models show enhanced robustness, particularly in multimodal tasks. REOBench underscores the vulnerability of current Earth observation foundation models to real-world corruptions and provides actionable insights for developing more robust and reliable models. Code and data are publicly available at https://github.com/lx709/REOBench.

cs.CV↗

arXiv2025

Spatial Chern-Simons Interactions and Complex Magnetic Penetration Depth

This paper examines the Landau-Ginzburg theory in the presence of spatial Chern-Simons interactions, which typically emerge in Weyl semimetals due to domain-wall motion. We demonstrate that the incorporation of a purely spatial Chern-Simons term, which violates parity, into the Landau-Ginzburg free energy leads to a complex magnetic penetration depth. This characteristic indicates that, aside from possessing an effective penetration depth, the magnetic field on the surface of the superconductor experiences periodic spatial oscillations, which significantly deviates from the conventional Meissner effect. In particular, we observe that as the degree of parity breaking, quantified by the strength of the spatial Chern-Simons term, increases, a vortex solution with magnetic field inversion may emerge. With the discovery of superconductivity in certain Weyl semimetals, we anticipate the possibility of experimentally observing this phenomenon in these materials.

cond-mat.supr-con↗

arXiv2025

A Physiologically-based simulation model of color appearance for red-green color vision deficiency

Various simulation methods of color appearance for dichromats or anomalous trichromats have been proposed over the years. To further improve the performance of the simulation model and extend the application range to both dichromats or anomalous trichromats, we have proposed a simulation model of cone fundamentals specifically designed for individuals with red-green type color vision deficiency (CVD) based on the CIE 2006 physiological observer model. By utilizing the simulated cone fundamentals, it becomes possible to predict the color appearance of real scenes and digital images for CVD. The fundamental premise of the new model is rooted in the hypothesis that CVD arises from a shift in the peak wavelength of the photopigment absorption spectrum of the L or M cone. Instead of simply maintaining the waveform without alteration as observed in prior studies, we altered waveforms of the absorption spectra of anomalous L/M cone photopigments when adjusting their peak wavelengths. Regarding different shapes in the absorption spectrum between the L and M cone, the absorption spectrum of the anomalous L/M cone was obtained by combining the peak wavenumber shift and linear interpolation of spectral quantal absorption curves between L- and M-photopigments in the wavenumber domain. The performance of the proposed model was substantiated through experimental validation by the pseudoisochromatic plates and Farnsworth Munsell 100 Hue test (FM-100). The findings revealed a high level of consistency between the model prediction and the actual perception reported by individuals with CVD.

physics.optics↗

arXiv2023

Superconducting quantum criticality and the anomalous scaling: A nonlinear relativistic equation

By using the Landau-Ginzburg-Wilson paradigm, we show that, near a quantum critical point (QCP), Cooper pairs at zero temperature would obey a nonlinear relativistic equation, where the imaginary time emerges as a novel dimension. This relativistic equation is applicable to certain superconductors at zero temperature for which the Faber-Pippard coherence length formula holds at and above the upper critical dimension. Here, we further show that the relativistic equation leads to a testable prediction in the vicinity of the QCP $T_c=0$, with $T_c$ being the transition temperature. That is, for 2D overdoped (clean) superconducting films, when the parameter $T_c/(c_0v_F)$ is lower than a characteristic scale, the Lorentz symmetry of relativistic equation arouses an anomalous scaling $ξ_0 \propto T_c^{-1.34}$, where $ξ_0$ denotes the zero-temperature coherence length, $v_F$ denotes the Fermi velocity, and $c_0$ denotes the Faber-Pippard coefficient. However, when the parameter $T_c/(c_0v_F)$ is large enough, the Lorentz symmetry may be broken so that the Faber-Pippard scaling $ξ_0 \propto T_c^{-1}$ is restored.

cond-mat.supr-con↗

arXiv2020

Relativistic Ginzburg-Landau equation: An investigation for overdoped cuprate films

By introducing the imaginary time, Gor'kov's Ginzburg-Landau equation at zero temperature can be extended to an exact relativistic form without any phenomenological parameter, which is intended to describe the zero-temperature overdoped cuprate. By using such a relativistic equation, we have shown that the two-class scaling observed in the overdoped side of single-crystal $La_{2-x}Sr_xCuO_4$ (LSCO) films [Nature 536, 309-311 (2016)] can be derived exactly. In this paper, we further test the validity of the relativistic Ginzburg-Landau equation. By applying the perturbation method into this equation, we theoretically predict that near the superconductor-metal transition point in the overdoped side of LSCO films, the zero-temperature correlation length $ξ(0)$ and the transition temperature $T_c$ should yield a novel scaling $ξ(0)\propto T_c^{-σ}$ with a critical exponent $σ\approx 1.31 $ (up to the two-loop approximation). Here, we propose a diffraction experiment between $X$-rays and zero-temperature LSCO films to measure the critical exponent $σ$.

cond-mat.supr-con↗

arXiv2020

Parabolic Scaling in Overdoped Cuprate: a Statistical Field Theory Approach

Recently, Bozovic et al. reported that [Nature 536, 309-311 (2016)], in the overdoped side of the single-crystal $La_{2-x}Sr_xCuO_4$ (LSCO) films, the transition temperature $T_c$ and zero-temperature superfluid phase stiffness $ρ_s(0)$ will obey a two-class scaling law: $T_c=γ\cdot \sqrt{ρ_s(0)}$ for $T_c \leq T_Q$ and $T_c \propto ρ_s(0)$ for $T_c \geq T_M$, where $γ=(4.2 \pm 0.5) K^{1/2} $, $T_Q \approx 15 K$, and $T_M \approx 12 K$. They further pointed out that the parabolic scaling observed in the highly overdoped side indicates a quantum phase transition from a superconductor to a normal metal. In this paper, we propose a quantum partition function (QPF) for zero-temperature Cooper pairs, by which one can effectively distinguish between mean-field and quantum critical behaviors. We theoretically show that the two-class scaling law can be exactly derived by using the QPF, and the theoretical values of $γ$, $T_Q$, and $T_M$ are well in accordance with experimental measure values. Our analyses indicate that the linear scaling $T_c \propto ρ_s(0)$ is a mean-field behavior, while the parabolic scaling $T_c=γ\cdot \sqrt{ρ_s(0)}$ is a quantum critical behavior.

physics.gen-ph↗

arXiv2019

Parabolic Scaling in Overdoped Cuprate Films

It was recently reported that, in the highly overdoped side of single-crystal $La_{2-x}Sr_xCuO_4$ films, the transition temperature $T_c$ and zero-temperature superfluid phase stiffness $ρ_s(0)$ will obey a parabolic scaling $T_c=γ\cdot \sqrt{ρ_s(0)}$. Parabolic scaling indicates a quantum phase transition from a superconductor to a normal metal, for which there has been scant understanding [Nature 536, 309-311 (2016)]. The current study shows that, using the quantum critical model for zero-temperature Cooper pairs [EPL 118, 57007 (2017)], parabolic scaling can be exactly derived, where $γ=γ(\varepsilon_F,a)$ is uniquely determined by the Fermi energy $\varepsilon_F$ and the minimal lattice constant $a$ of superconducting materials. For single-crystal $La_{2-x}Sr_xCuO_4$ films, we calculate the theoretical value of $γ$, which yields $4.29 K^{1/2}$ and is in accordance with an experimental measure value $(4.2 \pm 0.5) K^{1/2} $ with high accuracy. Our formula for $γ$ can be further tested by investigating other BCS-like materials.

cond-mat.supr-con↗

arXiv2017

Exponential Structure of Income Inequality: Evidence from 67 Countries

Economic competition between humans leads to income inequality, but, so far, there has been little understanding of underlying quantitative mechanisms governing such a collective behavior. We analyze datasets of household income from 67 countries, ranging from Europe to Latin America, North America and Asia. For all of the countries, we find a surprisingly uniform rule: Income distribution for the great majority of populations (low and middle income classes) follows an exponential law. To explain this empirical observation, we propose a theoretical model within the standard framework of modern economics and show that free competition and Rawls' fairness are the underlying mechanisms producing the exponential pattern. The free parameters of the exponential distribution in our model have an explicit economic interpretation and direct relevance to policy measures intended to alleviate income inequality.

econ.GN↗

arXiv2017

BCS quantum critical phenomena

Theoretically, we recently showed that the scaling relation between the transition temperature T_c and the superfluid density at zero temperature n_s (0) might exhibit a parabolic pattern [Scientific Reports 6 (2016) 23863]. It is significantly different from the linear scaling described by Homes' law, which is well known as a mean-field result. More recently, Bozovic et al. have observed such a parabolic scaling in the overdoped copper oxides with a sufficiently low transition temperature T_c [Nature 536 (2016) 309-311]. They further point out that this experimental finding is incompatible with the standard Bardeen-Cooper-Schrieffer (BCS) description. Here we report that if T_c is sufficiently low, applying the renormalization group approach into the BCS action at zero temperature will naturally lead to the parabolic scaling. Our result indicates that when T_c sufficiently approaches zero, quantum fluctuations will be overwhelmingly amplified so that the mean-field approximation may break down at zero temperature.

physics.gen-ph↗

arXiv2016

Scaling Laws for Thin Films near the Superconducting-to-Insulating Transition

We propose a Lagrangian function, which combines Landau-Ginzburg term and Chern-Simons term, for describing the competition between disorder and superconductivity. To describe the normal-to-superconducting transition in the thin superconducting films, we apply Wilson's renormalization group methods into this Lagrangian function. Finally, we obtain a scaling law between critical temperature (T_c), film thickness (d), sheet resistance of the film at the normal state (R_s), and number density of the electrons at the normal state (N). Such a scaling law is in agreement with recent experimental investigations [Ivry, Y. et al, Physical Review B 90, 214515 (2014)]. Our finding may have potential benefits for improving transition temperature T_c.

cond-mat.supr-con↗

arXiv2015

Universal Laws of Human Society's Income Distribution

General equilibrium equations in economics play the same role with many-body Newtonian equations in physics. Accordingly, each solution of the general equilibrium equations can be regarded as a possible microstate of the economic system. Since Arrow's Impossibility Theorem and Rawls' principle of social fairness will provide a powerful support for the hypothesis of equal probability, then the principle of maximum entropy is available in a just and equilibrium economy so that an income distribution will occur spontaneously (with the largest probability). Remarkably, some scholars have observed such an income distribution in some democratic countries, e.g. USA. This result implies that the hypothesis of equal probability may be only suitable for some "fair" systems (economic or physical systems). From this meaning, the non-equilibrium systems may be "unfair" so that the hypothesis of equal probability is unavailable.

q-fin.GN↗

arXiv2014

Spin Hall effect associated with SU(2) gauge field

In this paper, we focus on the connection between spin Hall effect and spin force. Here we investigate that the spin force due to spin-orbit coupling, which in two-dimensional system is equivalent to forces of Hirsch and Chudnovsky besides constant factors 3 and 3/2 respectively, is a part of classic Anandan force, and that the spin Hall effect is an anomalous Hall effect. Furthermore, we develop the method of AC phase to derive the formula for the spin force, and find that the most basic spin Hall effect originates from the AC phase and is therefore an intrinsic quantum mechanical property of spin. This method differs from approach of Berry phase in the study of anomalous Hall effect, which is the intrinsic property of the perfect crystal. On the other hand, we use an elegant skill to show that the Chudnovsky-Drude model is reasonable, and further have improved the theoretical values of spin Hall conductivity of Chudnovsky. Compared to the theoretical values of spin Hall conductivity in the Chudnovsky-Drude model, ours are in better agreement with experimentation. Finally, we discuss the relation between spin Hall effect and fractional statistics.

cond-mat.mes-hall↗

arXiv2014

Rawls' Fairness, Income Distribution and Alarming Level of Gini Coefficient

The argument that the alarming level of Gini coefficient is 0.4 is very popular, especially in the media industry, all around the world for a long time. Although the 0.4 standard is widely accepted, the derivation of the value lacks rigid theoretical foundations. In fact, to the best of our knowledge, it is not based on any prevalent and convincing economic theories. In this paper, we incorporate Rawls' principle of fair equality of opportunity into Arrow-Debreu's framework of general equilibrium theory with heterogeneous agents, and derive the alarming level of Gini coefficient formally. Our theory reveals that the exponential distribution of income not only satisfies Pareto optimality, but also obeys social fairness in Rawls' sense. Therefore, we specify the maximal value of the Gini coefficient when income follows exponential distribution as a possible alarming level. Our computations show that the alarming level should be specified at least equal or larger than 0.5 rather than 0.4. We empirically investigate if our model receives support from a large data set of all kinds of countries all over the world from Word Bank in 1990, 1995, 2000 and 2005 using the distribution fitting and statistical decision methodology. The results suggest that the value of 0.4 is around the mean of the Gini coefficients, corresponding to the most probable event in a peaceful world, rather than the alarming level, while the two-sigma rule shows that in our sample the alarming levels are all larger than 0.5, conforming to the predictions of our theory.

q-fin.GN↗

arXiv2013

Spontaneous Economic Order

This paper provides an attempt to formalize Hayek's notion of spontaneous order within the framework of the Arrow-Debreu economy. Our study shows that if a competitive economy is enough fair and free, then a spontaneous economic order shall emerge in long-run competitive equilibria so that social members together occupy an optimal distribution of income. Despite this, the spontaneous order might degenerate in the form of economic crises whenever an equilibrium economy approaches the extreme competition. Remarkably, such a theoretical framework of spontaneous order provides a bridge linking Austrian economics and Neoclassical economics, where we shall comprehend a truth: "Freedom promotes technological progress".

q-fin.GN↗