arXiv ScienceSearch

arXiv subjects

Hyunwoo Choi

Publications and source records attributed to Hyunwoo Choi.

11 recordsLinked to original sources

Smooth $\%$MinMax: A Differentiable Relaxation for Codon Harmonization

Codon harmonization aims to adapt the coding sequences for heterologous expression while preserving the native-like patterns of frequent and rare codons that may influence local translation dynamics and co-translational protein folding. However, widely used harmonization metrics, such as $\%$MinMax, are defined on discrete codon sequences and are, therefore, not readily compatible with gradient-based neural codon design. Here, we introduce Smooth $\%$MinMax, denoted as $\%{\rm MinMax}_{(s)}$, a differentiable relaxation of the conventional hard $\%$MinMax metric, denoted as $\%{\rm MinMax}_{(h)}$. $\%{\rm MinMax}_{(s)}$ replaces the discrete codon-usage values with probability-weighted synonymous-codon usage values and replaces the hard $\%$Max/$\%$Min branch with a sigmoid-gated interpolation. This formulation preserves the signed interpretation of $\%{\rm MinMax}_{(h)}$, while enabling optimization with respect to the synonymous-codon probabilities and learnable parameters. In human-to-Escherichia coli codon harmonization experiments, $\%{\rm MinMax}_{(s)}$ closely approximates $\%{\rm MinMax}_{(h)}$ and supports gradient-based profile matching in synonymous-codon probability space. These results suggest $\%{\rm MinMax}_{(s)}$ as a practical bridge between profile-based codon harmonization and neural synonymous-sequence design.

q-bio.QM

Full-Wave Green's-Function Modeling of Collective Single-Photon Emission in Non-Markovian Open-System QED with Finite-Bandwidth Compensation of Dispersive Interactions

This work presents a full-wave Green's function framework for modeling collective and coherent single-photon emission from multiple quantum emitters embedded in complex electromagnetic structures. Starting from a transverse modal completeness relation of modified Langevin noise formalism, we derive a closed set of coupled equations for population dynamics and frequency-resolved field amplitudes in the single-excitation regime. Since the electromagnetic reservoir is not traced out at the level of the dynamical amplitudes, the emitted single-photon dynamics can be modeled within the same closed set of equations without Markovian approximation in open and dissipative environments. We demonstrate that finite-bandwidth truncation of the spectral density leads to systematic deviations in coherent dispersive interactions, even when dissipative rates appear converged. To restore causal consistency, we introduce a counter-term compensation scheme that restores the missing dispersive contributions without modifying the retained non-Markovian memory kernel. To validate the scheme and demonstrate the practicality of the proposed framework, we present numerical examples ranging from benchmark configurations to a three-dimensional dispersive ring-resonator structure via finite element method. These capabilities provide a practical route for rigorously incorporating full-wave electromagnetic simulations into non-Markovian multi-emitter quantum electrodynamics, enabling predictive modeling of collective emission, coherent energy exchange, and single-photon radiation in realistic open structures.

quant-ph

Bounded Personas Match Retrieval on Classification but Not Regression for a Frozen Agent

A personalized language agent must convert a user's interaction history into behavior on each new request at inference time. Two strategies dominate. Retrieval pulls a few of the user's most relevant past items into the prompt, which is accurate but pays a per-query selection and context cost that grows with the history. Distillation instead compresses the history once into a compact natural-language persona, which is bounded, query-independent, and interpretable, but is widely assumed to sacrifice accuracy. Whether, and on which tasks, a distilled persona can match retrieval has not been characterized cleanly. We introduce PersonaLink, a training-free method that distills a user's history into a bounded three-field persona and recursively refines it: each pass self-evaluates the frozen agent on a held-out slice of the user's own labeled history, rewrites the persona from its errors, and keeps the result only when it does not regress on that slice. Because every comparison shares one frozen 7B backbone and differs only in what is placed in context, the design isolates the effect of representation from that of the model. The result is a clear task-type asymmetry. On 200 users of LaMP-2 (15-way news categorization), PersonaLink reaches 0.745-0.755 accuracy, statistically indistinguishable from BM25 retrieval (0.760-0.765).

cs.CL

Dissipative Channels Determine Open Electromagnetic Quantization

We formulate a quantization scheme for open electromagnetic systems with arbitrary passive boundary conditions. Rather than specifying reservoirs phenomenologically, the method identifies them from the dissipation geometry of the Maxwell operator. Factoring the imaginary part of the Maxwell operator gives a bosonic realization of the field operator and separates the fluctuation channels into medium-assisted reservoirs from material absorption and boundary-assisted reservoirs from exchange through the open boundary. Depending on the boundary condition, the latter become free-space radiation modes, impedance-load channels, guided port modes, or more general boundary channels. Green-function input-output relations then follow as an application, yielding frequency-dependent scattering and noise kernels without Markov or single-mode assumptions. To illustrate the practical application, we consider a lossy structure with mixed impedance and outgoing boundaries, and photonic integrated circuit configurations with waveguide port boundaries.

quant-ph

Computational framework for non-Markovian multi-emitter dynamics beyond the single-excitation limit

While non-Markovian dynamics have been extensively studied in the single-excitation limit to predict non-trivial phenomena, this regime remains an idealization. Moving beyond it is essential, as optical nonlinearities and phase-error accumulation in multi-photon processes render the Markovian approximation fragile. In this work, we present a Green's function-based framework for modeling non-Markovian multi-emitter quantum electrodynamics within the two-excitation manifold. The modified Langevin noise (M-LN) formalism is employed for first-principles treatment of dissipative environments, while the emitter-centered mode (ECM) framework ensures computational tractability. Unlike conventional approaches that integrate out the reservoir, we construct a non-Markovian hierarchy of coupled differential equations by explicitly retaining photonic amplitudes. Within the two-excitation hierarchy, the formulation preserves total probability and retains phase information necessary to capture multi-photon interference. As numerical demonstrations, we investigate non-Markovian atom-field interactions in structured semi-infinite waveguide environments. We first consider a homogeneous waveguide as a baseline, observing enhanced Bell-state fidelity in selected configurations. Next, we examine collective decay of symmetric Dicke states in a waveguide with an embedded lossy dielectric slab, revealing selective stabilization and delayed excitation transfer induced by the structured reservoir. Finally, we analyze entanglement dynamics in the same setting, highlighting entanglement sudden birth and oscillatory revivals. In principle, the framework applies to arbitrary electromagnetic environments for which the dyadic Green's function can be obtained numerically, providing a versatile tool for investigating complex non-Markovian multi-photon phenomena beyond the single-excitation limit.

quant-ph

Semi-Autonomous Mathematics Discovery with Gemini: A Case Study on the Erdős Problems

We present a case study in semi-autonomous mathematics discovery, using Gemini to systematically evaluate 700 conjectures labeled 'Open' in Bloom's Erdős Problems database. We employ a hybrid methodology: AI-driven natural language verification to narrow the search space, followed by human expert evaluation to gauge correctness and novelty. We address 13 problems that were marked 'Open' in the database: 5 through seemingly novel autonomous solutions, and 8 through identification of previous solutions in the existing literature. Our findings suggest that the 'Open' status of the problems was through obscurity rather than difficulty. We also identify and discuss issues arising in applying AI to math conjectures at scale, highlighting the difficulty of literature identification and the risk of ''subconscious plagiarism'' by AI. We reflect on the takeaways from AI-assisted efforts on the Erdős Problems.

cs.AI

Quantum-Plasmonic Dynamics Modeled via a Modified Langevin Noise Formalism: Numerical Studies of Single-Photon Emission and Two-Photon Interference

Recent studies have established and rigorously validated a modified Langevin noise formalism that enables first-principles quantization of electromagnetic fields in open and dissipative environments [1,2,3]. Building on this foundation, a fully quantum-mechanical multimode Jaynes-Cummings framework has been developed and verified, providing an accurate description of atom--field interactions in lossy and radiative systems [4]. In this work, we explore the potential of this formalism for nanophotonic applications by modeling representative quantum-plasmonic dynamics. In particular, we present detailed numerical examples for (i) two-photon interference mediated by a quantum plasmonic beam splitter, and (ii) non-Markovian dynamics of an atom located in plasmonic antennas and directional control of out-coupled single-photon fields. These results demonstrate that the proposed modeling approach can be directly used to guide the design and optimization of plasmonic single-photon sources and beam-splitting structures. Moreover, the framework is broadly applicable to the analysis of linear optical components and cavity quantum electrodynamics problems in open and dissipative photonic integrated circuits.

physics.optics

Atom-Field Non-Markovian Dynamics in Open and Dissipative Systems: An Efficient Memory-Kernel Approach Linked to Dyadic Greens Function and CEM Treatments

In this work, we present a numerical framework for modeling single photon emission from a two level system in open and dissipative systems beyond the Markovian approximation. The method can be readily integrated into standard computational electromagnetic (CEM) solvers such as finite difference time domain (FDTD) and finite element method (FEM). We numerically verify the completeness of boundary and medium assisted modes in the modified Langevin noise formalism by reconstructing the imaginary part of the dyadic Greens function through modal expansion in three dimensions. This reconstruction enables a first principles description of atom field interaction via the multi mode Jaynes Cummings model in open and dissipative environments. Within the single excitation manifold, we show that the memory kernel of a two level system is determined by the imaginary part of the Greens function, implying that radiative modes alone govern the relevant dynamics. The proposed framework thus provides a Greens function based approach for describing atomic population and single photon dynamics, directly compatible with Maxwell solvers. We then present concrete strategies for implementing our method in both FDTD and FEM frameworks, demonstrating its practical applicability. We further verify numerical results for a lossy Lorentz Drude type mirror, including both the case of a TLS near a finite sized metallic mirror and that of a TLS centered in a Fabry Perot cavity. This work establishes a rigorous foundation for incorporating quantum emitter dynamics into computational electromagnetics, thereby extending classical solvers toward quantum light matter interactions.

quant-ph

Numerical Framework for Multimode Jaynes- and Tavis-Cummings Models Incorporating the Modified Langevin Noise Formalism: Non-Markovian Analysis of Atom-Field Interactions in Dissipative Electromagnetic Environments

We present a novel numerical framework that integrates the modified Langevin noise formalism into the multimode Jaynes- and Tavis-Cummings models, enabling a first-principles, non-Markovian analysis of atom-field interactions in dissipative electromagnetic (EM) environments that account for both radiative losses and absorptive dissipation in lossy dielectric media (or satisfying general inhomogeneous causal media exhibiting both dispersion and absorption effects). In the modified Langevin noise formalism, the boundary- and medium-assisted (BA and MA) fields, which constitute a continuum set of EM modes in dissipative EM environments, are numerically obtained using the finite-element method (FEM). Specifically, BA field modes are extracted by solving plane-wave scattering problems, while MA field modes are determined through point-source radiation problems. These numerically obtained BA and MA field modes are then incorporated into the multimode Jaynes- and Tavis-Cummings models such that the coupling strength between atoms and BA-MA field modes can be calculated for the study of atom-field interactions in dissipative EM environments. The proposed methodology captures non-Markovian atomic dynamics that cannot be described by traditional quantum master equations under the Markovian approximation. To validate the accuracy of the proposed numerical framework, we present four numerical examples: (i) a two-level system (TLS) in a perfect electric conductor (PEC) half-space; (ii) dissipative cavity electrodynamics with two limiting cases approaching spontaneous emission in free space and ideal Rabi oscillations; (iii) super-radiance in TLS arrays; and (iv) entanglement sudden death of two TLSs inside dissipative cavities. The proposed methodology can serve as a ground-truth numerical simulator for studying atom-field interactions in general dissipative EM environments.

quant-ph

AVX Timing Side-Channel Attacks against Address Space Layout Randomization

Modern x86 processors support an AVX instruction set to boost performance. However, this extension may cause security issues. We discovered that there are vulnerable properties in implementing masked load/store instructions. Based on this, we present a novel AVX timing side-channel attack that can defeat address space layout randomization. We demonstrate the significance of our attack by showing User and Kernel ASLR breaks on the recent Intel and AMD processors in various environments, including cloud computing systems, an SGX enclave (a fine-grained ASLR break), and major operating systems. We further demonstrate that our attack can be used to infer user behavior, such as Bluetooth events and mouse movements. We highlight that stronger isolation or more fine-grained randomization should be adopted to successfully mitigate our presented attacks.

cs.CR

Thin-film transistor electrical performance of hybrid MoS 2 -P3HT semiconductor layers

The hole carrier field-effect mobility of hybrid molybdenum disulfide (MoS2) nanoparticles suspended in poly(3-hexylthiophene) (P3HT) thin film transistor (TFT) was found to be enhanced when it compared to P3HT-only TFTs. The improvement in the hole charge transport was found to be a function of the concentration of MoS2 in P3HT with high MoS2 concentrations resulting in an increase in the on-current of the device. Moreover, Au has a high work function of 5.1 eV which is suitable with the HOMO level of P3HT. We find that MoS2 and Au have the proper energy level for hole transport and injection.

physics.app-ph