SearcharxivSearch

arXiv subjects

Kang-Sin Choi

Publications and source records attributed to Kang-Sin Choi.

At least 19 recordsLinked to original sources

Gauge-Invariant Off-Shell Mass

We define a gauge-invariant and renormalized off-shell mass function in quantum field theory. The conventional self-energy is gauge-dependent off-shell. We generalize the self-energy, together with the corresponding vertex function, to be gauge-invariant and process-independent. To do this, we extend the pinch technique to an arbitrarily long fermion line. The Ward-Takahashi identity acting locally at each vertex cancels the gauge-dependent piece exactly, leaving the Feynman-gauge self-energy. On-shell renormalization then yields a mass function that is gauge-invariant, scheme-independent and equal to the physical mass at the pole; the same self-energy also yields an infrared-finite scalar mass function directly comparable with Schwinger-Dyson and lattice determinations. The cancellation is local to the internal fermion line and holds entirely off-shell, so the self-energy is well-defined on each internal segment of an amplitude with arbitrarily many external photons; we demonstrate this explicitly by computing the minimal off-shell Compton amplitude.

hep-th

Natural Higgs Mass from Power-Law Running

The renormalized scalar mass squared is a function of the energy scale and power-runs as its square. Well above the electroweak scale, its dimensionless couplings evolve only slowly, so the Standard Model is approximately scale invariant, and an order-one boundary value supplied at the unification scale is mapped exponentially down to the electroweak scale, much as the QCD scale arises from a dimensionless coupling alone. The 28 orders of magnitude separating the two then measure the smallness of an anomalous dimension, dominated by the top-quark Yukawa coupling, rather than a tuning. Naturalness is thereby recast as the value of an order-one ratio, with no protective symmetry required.

hep-ph

Autolearn: Learn by Surprise, Commit by Proof

We propose Autolearn, a framework that enables language models to learn from documents they read, with no external supervision. Passages that produce anomalously high per-token loss are flagged, verified through a self-generated Q&A chain, and trained on with conviction-proportional $β_2$ adjustment. We introduce the perturbation gap (paraphrase-to-original perplexity ratio) as a metric that distinguishes memorization from understanding. The key mechanism is the training data format: Q&A-format training drives the perturbation gap below the pre-trained baseline (2.098 vs. 2.204, $Δ= -0.106$, $> 10σ$), suppressing token-sequence memorization, while standard fine-tuning's best attempt remains within noise ($Δ= -0.010$, $< 1σ$). Across four models spanning Qwen3 and Phi-4 families, Autolearn is the only method that enters this regime. Stochastic evaluation reveals passage-specific knowledge acquisition: the probability of generating a correct novel fact rises from 6% to 54% after training ($p < 10^{-4}$), and Q&A format outperforms standard fine-tuning on genuinely novel facts. The system is self-extinguishing: learned content reduces surprisal below threshold and is skipped on re-encounter.

cs.LG

A Quantitative Definition of Intelligence

We propose an operational, quantitative definition of intelligence for arbitrary physical systems. The intelligence density of a system is the ratio of the logarithm of its independent outputs to its total description length. A system memorizes if its description length grows with its output count; it knows if its description length remains fixed while its output count diverges. The criterion for knowing is generalization. A system knows its domain if a single finite mechanism can produce correct outputs across an unbounded range of inputs, rather than storing each answer individually. The definition places intelligence on a substrate-independent continuum from logic gates to brains. We then argue that meaning over a domain is a selection and ordering of functions that produces correct outputs where correctness is specifiable. We also define a measure of contextuality of an output as the inverse of its conditional Kolmogorov complexity given the context of prior outputs, which unifies correctness and independence into a single condition. Together, these refute Searle's third premise, that syntax is insufficient for semantics, over any domain where correctness is specifiable.

cs.AI

The Gauge-Invariant Mass Function

In gauge theories, the mass of a field has been regarded as a purely on-shell concept: the pole mass is gauge-invariant, but the off-shell propagator has had no gauge-invariant definition of mass. We show that renormalization defines a gauge-invariant mass function at every virtuality, together with a gauge-invariant vertex. The virtual particle becomes as well defined as the on-shell one: the distinction is not dynamical but purely kinematic.

hep-th

On the Gauge-Invariant Fermion

We show that the Dirac dressing of the fermion is equivalent to a shift of the gauge parameter. For every gauge, the gauge-dependent part is projected out of physical observables. After renormalization, the physical mass is the same for every dressing. The non-locality, compositeness and path dependence associated with the dressing are therefore not physical obstructions.

hep-th

Global symmetry violation from non-isometric codes

We study the no-global-symmetry conjecture in quantum gravity by modeling black holes as non-isometric codes that encode the interior states with global charges into the fundamental states. The fluctuation in the inner products of the charged states can be larger compared to the case without charges. The non-isometric map causes states with different charges to have non-zero overlaps, signaling symmetry violation. The Renyi entropies of the radiation with global charges are found to be consistent with the quantum extremal surface formula. We compute various forms of Renyi relative entropy, as well as the fidelity, to quantify the degree of a global symmetry violation in Hawking radiation, demonstrating that global symmetries are indeed violated. We also comment on the instability of the black hole remnant.

hep-th

One-Loop Correction to the Higgs Mass

We present the complete one-loop corrected Higgs mass within the Standard Model. It is crucial to identify the observable mass as the renormalized one that depends on the external momentum. The correction is finite and dominated by the loops involving the W bosons and the top quark. These contributions exhibit a power-law running with momentum, which may be verifiable at the Large Hadron Collider.

hep-ph

Renormalization, Decoupling and the Hierarchy Problem

The hierarchy problem is associated with renormalization and decoupling. We can account for the smallness of the scalar mass against loop corrections and its insensitivity to ultraviolet physics through the decoupling of heavy fields. It is essential to correctly identify the observable physical mass as the renormalized one that depends on the external momentum, as opposed to the constant mass. We reconsider the properties of the renormalized loop corrections, which are finite, independent of regularization and admit a well-defined perturbation. By explicit calculation, we show that any loop corrections to the scalar mass-squared are suppressed as $(p^2-m^2)^2/M^2$, where $p,m$ and $M$ are the external momentum, the scalar pole mass and the heavy field mass in the loop, respectively. This is in accordance with the Appelquist-Carazzone decoupling theorem, which we have explicitized and completed for the case of the scalar mass.

hep-ph

Self-Similar Structure of Loop Amplitudes and Renormalization

We study the self-similar structure of loop amplitudes in quantum field theory and apply it to amplitude generation and renormalization. A renormalized amplitude can be regarded as an effective coupling that recursively appears within another loop. It is best described as a vertex function from the effective action. It is a scale-dependent, finite, parametrically small and observable quantity appearing in the S-matrix. Replacing a tree-level coupling with a loop amplitude provides a systematic method of generating high-order loop amplitudes, guaranteeing no subamplitude divergence. This method also provides an alternative bottom-up proof to the traditional top-down recursive renormalization of general amplitudes.

hep-th

Self-Similarity of Loop Amplitudes

We present an amplitude-generating formula in renormalizable quantum field theory. It reflects the self-similarity of loop amplitudes, in which an amplitude can also be a subamplitude of another. Amplitudes are generated by a small number of "irreducible" ones, which may replace tree-level couplings to form more complex amplitudes.

hep-th

On the Observables of Renormalizable Interactions

We reconsider the renormalization of scalar mass and point out that the quantum correction to the physical observable, as opposed to the bare parameter, of a renormalizable operator, is technically insensitive to ultraviolet physics and independent of the regularization scheme. It is expressed as the difference in the same quantities at different energy scales, maintaining the same asymptotics. Thus, any sensible regularization cancels out the divergences, including the quadratic ones, and yields the same finite corrections. To this end, we first show that the vacuum polarization of quantum electrodynamics is independent of the regularization scheme and a gauge-dependent quadratic divergence is canceled in the observable. We then calculate the quantum correction to the Higgs mass squared by the top-quark loop. It is again finite and regularization-scheme independent. For large external momentum, the correction of the pole mass-squared is dominated by power running, resulting in an order of 1 percent correction. In particular, the effect of heavy fields on the scalar mass correction is suppressed.

hep-ph

Stability of the Scalar Mass against Loop Corrections

We show that the renormalized loop corrections to the scalar mass are suppressed if the field in the loop is heavy. Here, treating the renormalized mass is important, as it is the only observable. This means that the physics of a low-energy scalar field is insensitive to that of ultraviolet.

hep-ph

Exact Renormalization of the Higgs Field

Using Wilsonian renormalization, we calculate the quantum correction to observable quantities, rather than the bare parameters, of the Higgs field. A physical parameter, such as a mass-squared or a quartic coupling, at an energy scale $μ$ is obtained from that at a reference scale by integrating in the degrees of freedom in between. In this process, heavy modes decouple and the ultraviolet scale dependence is canceled in the observables. An exact renormalization group equation is parametrized by the low-energy scale $μ$.

hep-th

Islands in Proliferating de Sitter Spaces

We study two-dimensional de Sitter universe which evolves and proliferates according to the Ginsparg-Perry-Bousso-Hawking mechanism, using Jackiw-Teitelboim gravity coupled to conformal matter. Black holes are generated by quantum gravity effects from pure de Sitter space and then evaporate to yield multiple disjoint de Sitter spaces. The back-reaction from the matter CFT is taken into account for the dilaton as a function of the temperature of the CFT. We discuss the evaporation of black holes and calculate the finite temperature entropy of an inflating region using the island formula. We find that the island moves towards the apparent horizon of the black hole as the temperature increases. The results are applied to the case of multiple evaporating black holes, for which we suggest multiple islands.

hep-th

Post-Newtonian Feasibility of the Closed String Massless Sector

We perform post Newtonian analysis of Double Field Theory as a test of string theory in gravitational sector against observations. We identify the Eddington--Robertson--Schiff parameters $β_{\rm{PPN}}$, $γ_{\rm{PPN}}$ with the charges of electric $H$-flux and dilaton respectively, and further relate them to stress-energy tensor. We show ${β_{\rm{PPN}}=1}$ from weak energy condition and argue that the observation of $γ_{\rm{PPN}}\simeq 1$ signifies the ultrarelativistic equation of state in baryons, or the suppression of gluon condensate.

hep-th

Aligned Natural Inflation in the Large Volume Scenario

We embed natural inflation in an explict string theory model and derive observables in cosmology. We achieve this by compactifying the type IIB string on a Calabi-Yau orientifold, stabilizing moduli via the Large Volume Scenario, and configuring axions using D7-brane stacks. In order to obtain a large effective decay constant, we employ the Kim-Nilles-Peloso alignment mechanism, with the required multiple axions arising naturally from anisotropic bulk geometries. The bulk volumes, and hence the axion decay constants, are stabilized by generalized one-loop corrections and subject to various conditions: the Kähler cone condition on the string geometry; the convex hull condition of the weak gravity conjecture; and the constraint from the power spectrum of scalar perturbations. We find that all constraints can be satisfied in a geometry with relatively small volume and thus heavy bulk axion mass. We also covariantize the convex hull condition for the axion-dilaton-instanton system and verify the normalization of the extremal bound.

hep-th

Small-Instanton Transitions in F-theory

We study the phase transition between G-instantons and D3-branes. A G-instanton is a classical solution to the self-dual equation of the M/F-theory four-form field strength G in the complex fourfold. This phase transition is dual to that between small instantons and 5-branes in the heterotic string. Using G as a background gauge flux, we may dynamically control the gauge symmetry breaking, connect between different vacua of F-theory and understand D3-branes in terms of group-theoretical quantities. We also discuss the resulting chirality change and preservation of anomaly freedom.

hep-th