arXiv subjects

Daniel Braun

Publications and source records attributed to Daniel Braun.

At least 19 recordsLinked to original sources

arXiv2026

Evaluating and Mitigating Anti-LGBTQ Biases in German and Multilingual Language Models

While gender and racial biases in language models have been widely studied, anti-LGBTQ biases remain underexplored, particularly beyond English. Existing benchmarks often do not capture cultural and linguistic variation and rely on gender representations. This paper introduces a multilingual German-English benchmark dataset for the evaluation of anti-LGBTQ biases in language models. It combines community-sourced stereotypes from German-speaking queer individuals with a German translation of WinoQueer. The data is used to evaluate eight language models across sizes and architectures and explore mitigation through fine-tuning on community and progressive media content. Results show that language models reproduce anti-queer stereotypes, with variation across identities and models. Differences between the translated and community-based data highlight the importance of cultural adaptation for multilingual bias evaluation. Fine-tuning reduces bias on average, but not consistently across models and identities. Warning: This text contains examples of anti-queer hateful language and stereotypes.

cs.CL↗

arXiv2026

Rotational Quantum Tunneling of a Magnetic Dipole in a Superconducting Trap

We study the quantum dynamics of the rotational degree of freedom of a nano-magnet trapped in a superconducting trap. The nano-magnet is modeled as a magnetic dipole with magnetization pinned to the easy axis of the particle. The magnetic trap then leads to a potential barrier that hinders free rotation of the particle, but through which it can tunnel. We identified rest-gas scattering as the most important decoherence mechanism at low temperatures. A shape of the particle sufficiently close to perfect rotational symmetry about the rotational axis can protect the rotational tunneling against this decoherence mechanism, and we identify experimentally feasible parameter regimes where rotational tunneling should be observable.

quant-ph↗

arXiv2026

Fine-Grained Perspectives: Modeling Explanations with Annotator-Specific Rationales

Beyond exploring disaggregated labels for modeling perspectives, annotator rationales provide fine-grained signals of individual perspectives. In this work, we propose a framework for jointly modeling annotator-specific label prediction and corresponding explanations, fine-tuned on the annotators' provided rationales. Using a dataset with disaggregated natural language inference (NLI) annotations and annotator-provided explanations, we condition predictions on both annotator identity and demographic metadata through a representation-level User Passport mechanism. We further introduce two explainer architectures: a post-hoc prompt-based explainer and a prefixed bridge explainer that transfers annotator-conditioned classifier representations directly into a generative model. This design enables explanation generation aligned with individual annotator perspectives. Our results show that incorporating explanation modeling substantially improves predictive performance over a baseline annotator-aware classifier, with the prefixed bridge approach achieving more stable label alignment and higher semantic consistency, while the post-hoc approach yields stronger lexical similarity. These findings indicate that modeling explanations as expressions of fine-grained perspective provides a richer and more faithful representation of disagreement. The proposed approaches advance perspectivist modeling by integrating annotator-specific rationales into both predictive and generative components.

cs.CL↗

arXiv2026

Path Integral Approach to Quantum Fisher Information

We present a real-time path-integral formulation of the quantum Fisher information for dynamical parameter estimation. For pure states undergoing unitary evolution, we show that the quantum Fisher information can be expressed as a connected symmetrized covariance of a time-integrated action deformation, equivalently as an integrated insertion of $\partial_\lambda S$ in the propagator. This reformulation avoids explicit state reconstruction by rewriting the quantum Fisher information in terms of real-time correlators that are natural targets for many-body methods. We further embed the construction into the Schwinger-Keldysh closed-time-path formalism, identifying the quantum Fisher information with the Keldysh component of an appropriate contour-ordered correlator generated by forward and backward propagating sources. Finally, using the Van Vleck-Gutzwiller approximation we re-derive the compact semiclassical quantum Fisher information expression, clarifying how classical trajectory data control leading-order metrological sensitivity.

quant-ph↗

arXiv2026

The Twin-World road to reality in quantum mechanics

I introduce a novel realistic, stochastic approach to quantum mechanics by extending the recently proposed grabit formalism \cite{braun_stochastic_2022} to two Twin Worlds. According to the picture developed, we live at the intersection of two worlds with identical stochastic laws of evolution. Our World is limited to that intersection, and only coincidence events from the two Twin Worlds, post-selected automatically by our restriction to the intersection, have physical reality in Our World. This fully reproduces standard non-relativistic quantum mechanics, including Born's rule and the violation of Bell's inequality. I derive the stochastic evolution equation in each Twin World that fully reproduces Schr\"odinger's equation for an arbitrary number of particles with arbitrary interactions, and demonstrate that hall-mark quantum effects such as tunnling are correctly reproduced.

quant-ph↗

arXiv2026

Cascaded Optomechanical Sensing for Small Signals

We propose a sensing scheme for detecting weak forces that achieves Heisenberg-limited sensitivity without relying on entanglement or other non-classical resources. Our scheme utilizes coherent averaging across a chain of N optomechanical cavities, unidirectionally coupled via a laser beam. As the beam passes through the cavities, it accumulates phase shifts induced by a common external force acting on the mechanical elements. Remarkably, this fully classical approach achieves the sensitivity scaling typically associated with quantum-enhanced protocols, providing a robust and experimentally feasible route to precision sensing. Potential applications range from high-sensitivity gravitational field measurements at the Large Hadron Collider to probing dark matter interactions and detecting gravitational waves. This work opens a new pathway for leveraging coherent light-matter interactions for force sensing.

quant-ph↗

arXiv2025

Acquiescence Bias in Large Language Models

Acquiescence bias, i.e. the tendency of humans to agree with statements in surveys, independent of their actual beliefs, is well researched and documented. Since Large Language Models (LLMs) have been shown to be very influenceable by relatively small changes in input and are trained on human-generated data, it is reasonable to assume that they could show a similar tendency. We present a study investigating the presence of acquiescence bias in LLMs across different models, tasks, and languages (English, German, and Polish). Our results indicate that, contrary to humans, LLMs display a bias towards answering no, regardless of whether it indicates agreement or disagreement.

cs.CL↗

arXiv2025

A Retail-Corpus for Aspect-Based Sentiment Analysis with Large Language Models

Aspect-based sentiment analysis enhances sentiment detection by associating it with specific aspects, offering deeper insights than traditional sentiment analysis. This study introduces a manually annotated dataset of 10,814 multilingual customer reviews covering brick-and-mortar retail stores, labeled with eight aspect categories and their sentiment. Using this dataset, the performance of GPT-4 and LLaMA-3 in aspect based sentiment analysis is evaluated to establish a baseline for the newly introduced data. The results show both models achieving over 85% accuracy, while GPT-4 outperforms LLaMA-3 overall with regard to all relevant metrics.

cs.CL↗

arXiv2025

The Impact of Annotator Personas on LLM Behavior Across the Perspectivism Spectrum

In this work, we explore the capability of Large Language Models (LLMs) to annotate hate speech and abusiveness while considering predefined annotator personas within the strong-to-weak data perspectivism spectra. We evaluated LLM-generated annotations against existing annotator modeling techniques for perspective modeling. Our findings show that LLMs selectively use demographic attributes from the personas. We identified prototypical annotators, with persona features that show varying degrees of alignment with the original human annotators. Within the data perspectivism paradigm, annotator modeling techniques that do not explicitly rely on annotator information performed better under weak data perspectivism compared to both strong data perspectivism and human annotations, suggesting LLM-generated views tend towards aggregation despite subjective prompting. However, for more personalized datasets tailored to strong perspectivism, the performance of LLM annotator modeling approached, but did not exceed, human annotators.

cs.CL↗

arXiv2025

Shaping Cold Atom Clouds with a Vortex Beam

We introduce a method for shaping a cold atom cloud using a vortex laser beam with a polarization singularity at its center, which creates a point of vanishing intensity. Exploiting this feature we experimentally demonstrate two different schemes to create micron-scaled line- and sheet-like atomic density distributions. In the dynamic scheme, atoms in the bright beam regions are accelerated and therefore effectively removed from the cloud. In the dark-state scheme, these atoms are pumped into a state that does not interact with the shaping light. In both cases, an atomic distribution remains, either as a thin line or as a sheet when an additional polarizer is used. We find good agreement between the experimental results and our theoretical model, which predicts the method to be in principle not diffraction-limited, paving the way for studies of phenomena arising in unconfined atomic ensembles on the micrometer scale.

quant-ph↗

arXiv2025

Gravitational wave imprints on spontaneous emission

Despite growing interest, there is a scarcity of known predictions in the regime where both quantum and general relativistic effects become observable. Here, we investigate a combined atom-field system in a curved spacetime, with a specific focus on gravitational-wave backgrounds. We demonstrate that a plane gravitational wave alters spontaneous emission from a single atom, manifesting itself as a direction-dependent change in the emission spectrum. Although the total decay rate remains unchanged, implying that no information about the gravitational wave is stored in the atomic internal state alone, the wave leaves imprints on the evolution of the composite atom-field system. To quantify how well this effect can be measured, we analyze both the classical Fisher information associated with photon number measurements and the quantum Fisher information. Our analysis indicates that the effect could be measured in state-of-the-art cold-atom experiments and points to spontaneous emission as a potential probe of low-frequency gravitational waves.

quant-ph↗

arXiv2025

Bandwidth enhanced current noise suppression with a SERF magnetometer

This work investigates the behavior of a spin-exchange relaxation-free (SERF) magnetometer integrated into the feedback branch of a closed-loop control circuit, designed to actively suppress noise from a current source. In this configuration, the magnetometer bandwidth is enhanced by almost two orders of magnitude as compared to the open-loop setup. Incorporating an injection transformer, the proposed system effectively reduces current noise to below 3 nA/sqrt(Hz) over a broad frequency bandwidth, independent of DC-current offsets. Analysis of the system sensitivity reveals that the performance is currently limited by technical noise. If this were to be eliminated, the system would achieve a current sensing sensitivity on the order of few pA/sqrt(Hz), while maintaining perfect galvanic isolation from the target circuit.

quant-ph↗

arXiv2025

Semiclassical Approach to Quantum Fisher Information

Quantum sensors driven into the quantum chaotic regime can have dramatically enhanced sensitivity, which, however, depends intricately on the details of the underlying classical phase space. Here, we develop an accurate semiclassical approach that provides direct and efficient access to the phase-space-resolved quantum Fisher information (QFI), the central quantity that quantifies the ultimate achievable sensitivity. This approximation reveals, in very concrete terms, that the QFI is large whenever a specific dynamical quantity tied to the sensing parameter displays a large variance over the course of the corresponding classical time evolution. Applied to a paradigmatic system of quantum chaos, the kicked top, we show that the semiclassical description is accurate already for modest quantum numbers, i.e., deep in the quantum regime, and it extends seamlessly to very high quantum numbers that are beyond the reach of other methods.

quant-ph↗

arXiv2025

The path towards measuring the gravitational field of proton bunches at accelerators

The Newtonian law describing the gravitational interaction of non-relativistic (slowly moving) gravitating matter, has been tested in many laboratory experiments with very high precision. In contrast, the post Minkowskian predictions for the gravitational field of ultra-relativistic matter, dominated by momentum instead of rest mass, have not been tested directly yet. The intense ultra-relativistic proton beam in the LHC storage ring offers the potential to test general relativity and alternative gravitational theories in this parameter regime for the first time in controlled lab-scale experiments. If successful, this would open the road to a novel use case of the LHC, where non-trivial gravitational physics could be studied likely in a parasitic mode, without the necessity of dedicated filling patterns. While the technical challenges are formidable, they should also lead to the development of ultra-high-sensitive acceleration sensors with abundant applications in other parts of science and technology. The present document summarizes the status of the theoretical studies in this direction, points out the challenges, and possible ways of addressing them. It was submitted as a contribution to the European Strategy for Particle Physics (ESPP) 2026 Update.

hep-ex↗

arXiv2025

Quantum Metrology of Newton's Constant with Levitated Mechanical Systems

Newton's constant is the least well-measured among the fundamental constants of Nature, and, indeed, its accurate measurement has long served an experimental challenge. Levitated mechanical systems are attracting growing attention for their promising applications in sensing and as an experimental platform for exploring the intersection between quantum physics and gravitation. Here we propose a mechanical interferometric scheme of interacting levitated oscillators for the accurate estimation of Newton's constant. Our scheme promises to beat the current standard by several orders of magnitude.

quant-ph↗

arXiv2025

Position: Editing Large Language Models Poses Serious Safety Risks

Large Language Models (LLMs) contain large amounts of facts about the world. These facts can become outdated over time, which has led to the development of knowledge editing methods (KEs) that can change specific facts in LLMs with limited side effects. This position paper argues that editing LLMs poses serious safety risks that have been largely overlooked. First, we note the fact that KEs are widely available, computationally inexpensive, highly performant, and stealthy makes them an attractive tool for malicious actors. Second, we discuss malicious use cases of KEs, showing how KEs can be easily adapted for a variety of malicious purposes. Third, we highlight vulnerabilities in the AI ecosystem that allow unrestricted uploading and downloading of updated models without verification. Fourth, we argue that a lack of social and institutional awareness exacerbates this risk, and discuss the implications for different stakeholders. We call on the community to (i) research tamper-resistant models and countermeasures against malicious model editing, and (ii) actively engage in securing the AI ecosystem.

cs.CL↗

arXiv2025

Metrology of Gravitational Effects with Mechanical Quantum Systems

Mechanical quantum systems, such as resonators and levitated particles, offer unique opportunities for quantum metrology. Particularly, their significant mass and quantum-level control enable applications in measuring gravitational effects. This article highlights key challenges, alongside potential solutions to advance precision sensing and quantum-gravitational research.

quant-ph↗

arXiv2024

Interferometric Mass Photometry at the Quantum Limit of Sensitivity

We present an innovative optical imaging system for measuring parameters of a small particle such as a macromolecule or nanoparticle at the quantum limit of sensitivity. In comparison to the conventional confocal interferometric scattering (iSCAT) approach, our setup adds a second arm to form a Michelson interferometer that allows us to tune a relative phase. We evaluate the quantum Cram\'er-Rao bound (QCRB) for different quantum states, including single-mode coherent states, multi-frequency coherent states, and phase-averaged coherent states. Our results show that the proposed setup can achieve the QCRB of sensitivity and outperform iSCAT for all considered quantum states for mass and phase estimation of a particle.

quant-ph↗