Search

This summary of the second Terrestrial Very-Long-Baseline Atom Interferometry (TVLBAI) Workshop provides a comprehensive overview of our meeting held in London in April 2024, building on the initial discussions during the inaugural workshop held at CERN in March 2023. Like the summary of the first workshop, this document records a critical milestone for the international atom interferometry community. It documents our concerted efforts to evaluate progress, address emerging challenges, and refine strategic directions for future large-scale atom interferometry projects. Our commitment to collaboration is manifested by the integration of diverse expertise and the coordination of international resources, all aimed at advancing the frontiers of atom interferometry physics and technology, as set out in a Memorandum of Understanding signed by over 50 institutions., Comment: Summary of the second Terrestrial Very-Long-Baseline Atom Interferometry Workshop held at Imperial College London: https://indico.cern.ch/event/1369392/

Many astrophysical systems of interest to numerical relativity, such as rapidly rotating stars, black hole accretion disks, and core-collapse supernovae, exhibit near-symmetries. These systems generally consist of a strongly gravitating central object surrounded by an accretion disk, debris, and ejecta. Simulations can efficiently exploit the near-axisymmetry of these systems by reducing the number of points in the angular direction around the near-symmetry axis, enabling efficient simulations over seconds-long timescales with minimal computational expense. In this paper, we introduce GRoovy, a novel code capable of modeling astrophysical systems containing compact objects by solving the equations of general relativistic hydrodynamics (GRHD) in full general relativity using singular curvilinear (spherical-like and cylindrical-like) and Cartesian coordinates. We demonstrate the code's robustness through a battery of challenging GRHD tests, ranging from flat, static spacetimes to curved, dynamical spacetimes. These tests further showcase the code's capabilities in modeling systems with realistic, finite-temperature equations of state and neutrino cooling via a leakage scheme. GRoovy extensively leverages GRHayL, an open-source, modular, and infrastructure-agnostic general relativistic magnetohydrodynamics library built from the highly robust algorithms of IllinoisGRMHD. Long-term simulations of binary neutron star and black hole-neutron star post-merger remnants will benefit greatly from GRoovy to study phenomena such as remnant stability, gamma-ray bursts, and nucleosynthesis., Comment: 17 pages, 10 figures

In computational physics, chemistry, and biology, the implementation of new techniques in a shared and open source software lowers barriers to entry and promotes rapid scientific progress. However, effectively training new software users presents several challenges. Common methods like direct knowledge transfer and in-person workshops are limited in reach and comprehensiveness. Furthermore, while the COVID-19 pandemic highlighted the benefits of online training, traditional online tutorials can quickly become outdated and may not cover all the software's functionalities. To address these issues, here we introduce ``PLUMED Tutorials'', a collaborative model for developing, sharing, and updating online tutorials. This initiative utilizes repository management and continuous integration to ensure compatibility with software updates. Moreover, the tutorials are interconnected to form a structured learning path and are enriched with automatic annotations to provide broader context. This paper illustrates the development, features, and advantages of PLUMED Tutorials, aiming to foster an open community for creating and sharing educational resources., Comment: 26 pages, 5 figures

This work analyzes transfer learning of the Variational Quantum Circuit (VQC). Our framework begins with a pretrained VQC configured in one domain and calculates the transition of 1-parameter unitary subgroups required for a new domain. A formalism is established to investigate the adaptability and capability of a VQC under the analysis of loss bounds. Our theory observes knowledge transfer in VQCs and provides a heuristic interpretation for the mechanism. An analytical fine-tuning method is derived to attain the optimal transition for adaptations of similar domains., Comment: Submitted to ICASSP 2025

Muscles consume metabolic energy (ATP) to produce force. A mathematical model for energy expenditure can be useful in estimating real-time costs of movements or to predict energy optimal movements. Metabolic cost models developed so far have predominantly aimed at dynamic movement tasks, where mechanical work dominates. Further, while it is known that both force magnitude and rate of change of force (force rate) affect metabolic cost, it is not known how these terms interact, or if the force rate dependence can be a consequence of the force dependence. Here, we performed extensive human subject experiments, involving each subject over 5 hours of metabolic trials, which systematically changed the mean forces and forces rates so as to characterize a holistic relation for metabolic cost based on both force and force rate -- or analogously, torque and torque rate. Our experiments involved humans producing symmetric or asymmetric sinusoidal forces with different means, amplitudes, frequencies, and rise and fall periods. We showed that the metabolic cost can be well-approximated by a sum of power law functions of torque and torque rate. We found that the metabolic cost scales non-linearly with joint torque (with exponent = 1.36) and non-linearly with torque rate (with exponent = 2.5). Surprisingly, the data suggested that the cost was roughly four times higher for decreasing the torque than increasing, mirroring the analogous ratio between the cost of positive and negative work. Using these metabolic cost relations, we show that if the metabolic cost scales with particular exponents with muscle force and force rates, the same exponents will be observed in multi-joint tasks with multiple muscles. Our new metabolic cost model involving both force and force rate will potentially allow better predictions of energy optimal movements and thus inform wearable robot design and analysis.

Electron-phonon coupling often dominates the electron spectral functions and carrier transport properties. However, studies of this effect in real materials have largely relied on perturbative one-shot methods due to the lack of a first-principles theoretical and computational framework. Here, we present a self-consistent theory and implementation for the non-perturbative calculations of spectral functions and conductivity due to electron-phonon coupling. Applying this method to monolayer InSe, we demonstrate that self-consistency qualitatively affects the spectral function and transport properties compared to state-of-the-art one-shot calculations and allow one to reconcile experimental angle-resolved photoemission experiments. The developed method can be widely applied to materials with dominant electron-phonon coupling at moderate computational cost.

This work addresses the challenge of optimal energy management in microgrids through a collaborative and privacy-preserving framework. We propose the FedTRPO methodology, which integrates Federated Learning (FL) and Trust Region Policy Optimization (TRPO) to manage distributed energy resources (DERs) efficiently. Using a customized version of the CityLearn environment and synthetically generated data, we simulate designed net-zero energy scenarios for microgrids composed of multiple buildings. Our approach emphasizes reducing energy costs and carbon emissions while ensuring privacy. Experimental results demonstrate that FedTRPO is comparable with state-of-the-art federated RL methodologies without hyperparameter tunning. The proposed framework highlights the feasibility of collaborative learning for achieving optimal control policies in energy systems, advancing the goals of sustainable and efficient smart grids., Comment: Submitted to Environmental Data Science Journal from Cambridge University Press

We study $\textit{sparse singular value certificates}$ for random rectangular matrices. If $M$ is an $n \times d$ matrix with independent Gaussian entries, we give a new family of polynomial-time algorithms which can certify upper bounds on the maximum of $\|M u\|$, where $u$ is a unit vector with at most $\eta n$ nonzero entries for a given $\eta \in (0,1)$. This basic algorithmic primitive lies at the heart of a wide range of problems across algorithmic statistics and theoretical computer science. Our algorithms certify a bound which is asymptotically smaller than the naive one, given by the maximum singular value of $M$, for nearly the widest-possible range of $n,d,$ and $\eta$. Efficiently certifying such a bound for a range of $n,d$ and $\eta$ which is larger by any polynomial factor than what is achieved by our algorithm would violate lower bounds in the SQ and low-degree polynomials models. Our certification algorithm makes essential use of the Sum-of-Squares hierarchy. To prove the correctness of our algorithm, we develop a new combinatorial connection between the graph matrix approach to analyze random matrices with dependent entries, and the Efron-Stein decomposition of functions of independent random variables. As applications of our certification algorithm, we obtain new efficient algorithms for a wide range of well-studied algorithmic tasks. In algorithmic robust statistics, we obtain new algorithms for robust mean and covariance estimation with tradeoffs between breakdown point and sample complexity, which are nearly matched by SQ and low-degree polynomial lower bounds (that we establish). We also obtain new polynomial-time guarantees for certification of $\ell_1/\ell_2$ distortion of random subspaces of $\mathbb{R}^n$ (also with nearly matching lower bounds), sparse principal component analysis, and certification of the $2\rightarrow p$ norm of a random matrix.

Solving complex real-world tasks requires cycles of actions and observations. This is particularly true in science, where tasks require many cycles of analysis, tool use, and experimentation. Language agents are promising for automating intellectual tasks in science because they can interact with tools via natural language or code. Yet their flexibility creates conceptual and practical challenges for software implementations, since agents may comprise non-standard components such as internal reasoning, planning, tool usage, as well as the inherent stochasticity of temperature-sampled language models. Here, we introduce Aviary, an extensible gymnasium for language agents. We formalize agents as policies solving language-grounded partially observable Markov decision processes, which we term language decision processes. We then implement five environments, including three challenging scientific environments: (1) manipulating DNA constructs for molecular cloning, (2) answering research questions by accessing scientific literature, and (3) engineering protein stability. These environments were selected for their focus on multi-step reasoning and their relevance to contemporary biology research. Finally, with online training and scaling inference-time compute, we show that language agents backed by open-source, non-frontier LLMs can match and exceed both frontier LLM agents and human experts on multiple tasks at up to 100x lower inference cost.

High-energy factorization of 2 -> 2 amplitudes in QCD has been recently pushed to the next-to-next-to-leading logarithmic order by determining the three-loop gluon Regge trajectory. This was based on computing multi-Reggeon exchanges using rapidity evolution in the shock-wave formalism, and disentangling between the Regge pole and Regge cut contributions. In the present paper we extend the relevant theoretical framework to 2 -> 3 processes, and compute all multi-Reggeon exchanges necessary for extracting the two-loop Reggeon-gluon-Reggeon Lipatov vertex from 2 -> 3 amplitudes. Then, specializing general amplitude methods to multi-Regge kinematics, we derive analytic expressions for non-planar two-loop gg -> ggg, gq -> ggq and qq -> qgq QCD amplitudes in that limit. Matching these to the multi-Reggeon computation, we determine the QCD Lipatov vertex in dimensional regularization at two loops through finite terms. We also determine the one-loop vertex through O(epsilon^4). All results are expressed in a compact form in terms of a basis of single-valued generalised polylogarithms, manifesting target-projectile symmetry and reality properties. Furthermore, our basis of functions is explicitly finite in the soft limit, featuring delicate cancellation of spurious rational poles by transcendental functions. Agreement between all three partonic channels, as well agreement of the maximal weight contributions with the super Yang-Mills Lipatov vertex provide robust checks of the result., Comment: 107 pages, 9 figures

Scattering amplitudes in the high-energy limit can be described in terms of their singularity structure in the complex angular momentum plane, consisting of Regge poles and cuts. In QCD, gluon Reggeization has long been understood as a manifestation of a Regge pole, but until recently Reggeization violation remained largely obscure. New methods, based on iterative solution of rapidity evolution equations, facilitate direct computation of components of the amplitude which are mediated by multi-Reggeon exchange, a manifestation of Regge cuts. Upon disentangling the Regge cut from the pole we are now able to extract the pole parameters from state-of-the-art fixed-order computations (3 loops) and make predictions regarding certain components of the amplitude to higher loop orders. In this talk I review the key ideas which led to this progress, describe where we stand in exploring the structure of 2 -> 2 and 2 -> 3 amplitudes in the (multi-) Regge limit, and comment on the interplay between this research and the study of infrared factorization., Comment: Talk presented at Loops and Legs in Quantum Field Theory (LL2024), 14-19 April, 2024, Wittenberg, Germany. 17 pages, 5 figures

Recent advancements in generative modeling now enable the creation of 4D content (moving 3D objects) controlled with text prompts. 4D generation has large potential in applications like virtual worlds, media, and gaming, but existing methods provide limited control over the appearance and geometry of generated content. In this work, we introduce a method for animating user-provided 3D objects by conditioning on textual prompts to guide 4D generation, enabling custom animations while maintaining the identity of the original object. We first convert a 3D mesh into a ``static" 4D Neural Radiance Field (NeRF) that preserves the visual attributes of the input object. Then, we animate the object using an Image-to-Video diffusion model driven by text. To improve motion realism, we introduce an incremental viewpoint selection protocol for sampling perspectives to promote lifelike movement and a masked Score Distillation Sampling (SDS) loss, which leverages attention maps to focus optimization on relevant regions. We evaluate our model in terms of temporal coherence, prompt adherence, and visual fidelity and find that our method outperforms baselines that are based on other approaches, achieving up to threefold improvements in identity preservation measured using LPIPS scores, and effectively balancing visual quality with dynamic content.

Strong artificial intelligence (AI) is envisioned to possess general cognitive abilities and scientific creativity comparable to human intelligence, encompassing both knowledge acquisition and problem-solving. While remarkable progress has been made in weak AI, the realization of strong AI remains a topic of intense debate and critical examination. In this paper, we explore pivotal innovations in the history of astronomy and physics, focusing on the discovery of Neptune and the concept of scientific revolutions as perceived by philosophers of science. Building on these insights, we introduce a simple theoretical and statistical framework of weak beliefs, termed the Transformational Belief (TB) framework, designed as a foundation for modeling scientific creativity. Through selected illustrative examples in statistical science, we demonstrate the TB framework's potential as a promising foundation for understanding, analyzing, and even fostering creativity -- paving the way toward the development of strong AI. We conclude with reflections on future research directions and potential advancements.

We compute all planar two-loop six-point Feynman integrals entering scattering observables in massless gauge theories such as QCD. A central result of this paper is the formulation of the differential-equations method under the algebraic constraints stemming from four-dimensional kinematics, which in this case leaves only 8 independent scales. We show that these constraints imply that one must compute topologies with only up to 8 propagators, instead of the expected 9. This leads to the decoupling of entire classes of integrals that do not contribute to scattering amplitudes in four dimensional gauge theories. We construct a pure basis and derive their canonical differential equations, of which we discuss the numerical solution. This work marks an important step towards the calculation of massless $2\to 4$ scattering processes at two loops., Comment: 32 pages, 2 figures

Boolean combinations allow combining given combinatorial objects to obtain new, potentially more complicated, objects. In this paper, we initiate a systematic study of this idea applied to graphs. In order to understand expressive power and limitations of boolean combinations in this context, we investigate how they affect different combinatorial and structural properties of graphs, in particular $\chi$-boundedness, as well as characterize the structure of boolean combinations of graphs from various classes.

Background and Objective: Radiation pneumonitis (RP) is a side effect of thoracic radiation therapy. Recently, Machine learning (ML) models enhanced with radiomic and dosiomic features provide better predictions by incorporating spatial information beyond DVHs. However, to improve the clinical decision process, we propose to use uncertainty quantification (UQ) to improve the confidence in model prediction. This study evaluates the impact of post hoc UQ methods on the discriminative performance and calibration of ML models for RP prediction. Methods: This study evaluated four ML models: logistic regression (LR), support vector machines (SVM), extreme gradient boosting (XGB), and random forest (RF), using radiomic, dosiomic, and dosimetric features to predict RP. We applied UQ methods, including Patt scaling, isotonic regression, Venn-ABERS predictor, and Conformal Prediction, to quantify uncertainty. Model performance was assessed through Area Under the Receiver Operating Characteristic curve (AUROC), Area Under the Precision-Recall Curve (AUPRC), and Adaptive Calibration Error (ACE) using Leave-One-Out Cross-Validation (LOO-CV). Results: UQ methods enhanced predictive performance, particularly for high-certainty predictions, while also improving calibration. Radiomic and dosiomic features increased model accuracy but introduced calibration challenges, especially for non-linear models like XGB and RF. Performance gains from UQ methods were most noticeable at higher certainty thresholds. Conclusion: Integrating UQ into ML models with radiomic and dosiomic features improves both predictive accuracy and calibration, supporting more reliable clinical decision-making. The findings emphasize the value of UQ methods in enhancing applicability of predictive models for RP in healthcare settings.

We prove that, to each synchronous non-local game $\mathcal{G}=(I,O,\lambda)$ with $|I|=n$ and $|O|=m \geq 3$, there is an associated graph $G_{\lambda}$ for which approximate winning strategies for the game $\mathcal{G}$ and the $3$-coloring game for $G_{\lambda}$ are preserved. That is, using a similar graph to previous work of the author (Ann. Henri Poincar\'{e}, 2024), any synchronous strategy for $\text{Hom}(G_{\lambda},K_3)$ that wins the game with probability $1-\varepsilon$ with respect to the uniform probability distribution on the edges, yields a strategy in the same model that wins the game $\mathcal{G}$ with respect to the uniform distribution with probability at least $1-h(n,m)\varepsilon^{\frac{1}{2}}$, where $h$ is a polynomial in $n$ and $2^m$. As an application, we prove that the gapped promise problem for quantum $3$-coloring is undecidable. Moreover, we prove that there exists an $\alpha \in (0,1)$ for which determining whether the non-commutative Max-$3$-Cut of a graph is $|E|$ or less than $\alpha |E|$ is RE-hard, thus giving a positive answer to a problem posed by Culf, Mousavi and Spirig (arXiv:2312.16765), along with evidence for a sharp computability gap in the non-commutative Max-$3$-Cut problem. We also prove that there is some $\alpha \in (0,1)$ such that determining the non-commutative (respectively, commuting operator framework) versions of the Max-$3$-Cut of a graph within a factor of $\alpha$ is uncomputable. All of these results avoid use of the unique games conjecture., Comment: 37 pages

We introduce AgreeMate, a framework for training Large Language Models (LLMs) to perform strategic price negotiations through natural language. We apply recent advances to a negotiation setting where two agents (i.e. buyer or seller) use natural language to bargain on goods using coarse actions. Specifically, we present the performance of Large Language Models when used as agents within a decoupled (modular) bargaining architecture. We demonstrate that using prompt engineering, fine-tuning, and chain-of-thought prompting enhances model performance, as defined by novel metrics. We use attention probing to show model attention to semantic relationships between tokens during negotiations., Comment: 15 pages, 22 figures, 6 tables

Lookup tables (LUTs) are frequently used to efficiently store arrays of precomputed values for complex mathematical computations. When used in the context of neural networks, these functions exhibit a lack of recognizable patterns which presents an unusual challenge for conventional logic synthesis techniques. Several approaches are known to break down a single large lookup table into multiple smaller ones that can be recombined. Traditional methods, such as plain tabulation, piecewise linear approximation, and multipartite table methods, often yield inefficient hardware solutions when applied to LUT-based NNs. This paper introduces ReducedLUT, a novel method to reduce the footprint of the LUTs by injecting don't cares into the compression process. This additional freedom introduces more self-similarities which can be exploited using known decomposition techniques. We then demonstrate a particular application to machine learning; by replacing unobserved patterns within the training data of neural network models with don't cares, we enable greater compression with minimal model accuracy degradation. In practice, we achieve up to $1.63\times$ reduction in Physical LUT utilization, with a test accuracy drop of no more than $0.01$ accuracy points.

Mechanical and elastic properties of materials are among the most fundamental quantities for many engineering and industrial applications. Here, we present a formulation that is efficient and accurate for calculating the elastic and bending rigidity tensors of crystalline solids, leveraging interatomic force constants and long-wavelength perturbation theory. Crucially, in the long-wavelength limit, lattice vibrations induce macroscopic electric fields which further couple with the propagation of elastic waves, and a separate treatment on the long-range electrostatic interactions is thereby required to obtain elastic properties under the appropriate electrical boundary conditions. A cluster expansion of the charge density response and dielectric screening function in the long-wavelength limit has been developed to efficiently extract multipole and dielectric tensors of arbitrarily high order. We implement the proposed method in a first-principles framework and perform extensive validations on silicon, NaCl, GaAs and rhombohedral BaTiO$_3$ as well as monolayer graphene, hexagonal BN, MoS$_2$ and InSe, obtaining good to excellent agreement with other theoretical approaches and experimental measurements. Notably, we establish that multipolar interactions up to at least octupoles are necessary to obtain the accurate short-circuit elastic tensor of bulk materials, while higher orders beyond octupole interactions are required to converge the bending rigidity tensor of 2D crystals. The present approach greatly simplifies the calculations of bending rigidities and will enable the automated characterization of the mechanical properties of novel functional materials., Comment: 35 pages, 7 figures and 7 tables for main text; 43 pages with supplementary material

Convolutional neural networks (CNNs) can learn directly from raw data, resulting in exceptional performance across various research areas. However, factors present in non-controllable environments such as unlabeled datasets with varying levels of domain and category shift can reduce model accuracy. The Open Set Domain Adaptation (OSDA) is a challenging problem that arises when both of these issues occur together. Existing OSDA approaches in literature only align known classes or use supervised training to learn unknown classes as a single new category. In this work, we introduce a new approach to improve OSDA techniques by extracting a set of high-confidence unknown instances and using it as a hard constraint to tighten the classification boundaries. Specifically, we use a new loss constraint that is evaluated in three different ways: (1) using pristine negative instances directly; (2) using data augmentation techniques to create randomly transformed negatives; and (3) with generated synthetic negatives containing adversarial features. We analyze different strategies to improve the discriminator and the training of the Generative Adversarial Network (GAN) used to generate synthetic negatives. We conducted extensive experiments and analysis on OVANet using three widely-used public benchmarks, the Office-31, Office-Home, and VisDA datasets. We were able to achieve similar H-score to other state-of-the-art methods, while increasing the accuracy on unknown categories.

Fundamental research on two-dimensional (2D) magnetic systems based on van der Waals materials has been gaining traction rapidly since their recent discovery. With the increase of recent knowledge, it has become clear that such materials have also a strong potential for applications in devices that combine magnetism with electronics, optics, and nanomechanics. Nonetheless, many challenges still lay ahead. Several fundamental aspects of 2D magnetic materials are still unknown or poorly understood, such as their often-complicated electronic structure, optical properties, and magnetization dynamics, and their magnon spectrum. To elucidate their properties and facilitate integration in devices, advanced characterization techniques and theoretical frameworks need to be developed or adapted. Moreover, developing synthesis methods which increase critical temperatures and achieve large-scale, high-quality homogeneous thin films is crucial before these materials can be used for real-world applications. Therefore, the field of 2D magnetic materials provides many challenges and opportunities for the discovery and exploration of new phenomena, as well as the development of new applications. This Roadmap presents the background, challenges, and potential research directions for various relevant topics in the field on the fundamentals, synthesis, characterization, and applications. We hope that this work can provide a strong starting point for young researchers in the field and provide a general overview of the key challenges for more experienced researchers., Comment: 87 pages, 24 figures. Roadmap based on the discussions during the workshop on Quantum Magnetic Materials hosted by the Lorentz Centre in Leiden, the Netherlands (Oct. 2023)

In high-risk environments where unlawful surveillance is prevalent, securing confidential communications is critical. This study introduces a novel steganographic game-theoretic model to analyze the strategic interactions between a defending company and an adversary. By framing the scenario as a non-cooperative game, there is systematic evaluation of optimal strategies for both parties, incorporating costs and benefits such as implementation expenses, potential data leaks, and operational advantages. The derived equilibrium probabilities enable the assessment of success rates, illustrating conditions under which the company benefits from hiding messages or faces increased risks when not implementing steganography. Sensitivity analysis explores how changes in key parameters impact these strategies, enhancing the understanding of decision-making in secure communications. Furthermore, the introduction of an adversary model that quantifies the adversary's advantage using conditional probabilities derived from success rates allows for a quantitative measure of the adversary's effectiveness based on the defender's strategies. By integrating the adversary's advantage into a novel risk analysis framework and employing Monte Carlo simulations, dynamic interactions are captured across advantage scenarios, considering factors like impact factor, steganography effectiveness, and equilibrium probabilities. This comprehensive framework offers practical insights into optimizing security strategies by quantifying potential risk reductions when the adversary is disadvantaged, providing a clear methodology for assessing and mitigating adversarial threats in complex security environments., Comment: 21 pages, 7 figures, 1 table, 2 algorithms. This version is a preprint uploaded to arXiv

Deep neural networks have shown outstanding performance in computer vision tasks such as semantic segmentation and have defined the state-of-the-art. However, these segmentation models are trained on a closed and predefined set of semantic classes, which leads to significant prediction failures in open-world scenarios on unknown objects. As this behavior prevents the application in safety-critical applications such as automated driving, the detection and segmentation of these objects from outside their predefined semantic space (out-of-distribution (OOD) objects) is of the utmost importance. In this work, we present a multi-scale OOD segmentation method that exploits the confidence information of a foreground-background segmentation model. While semantic segmentation models are trained on specific classes, this restriction does not apply to foreground-background methods making them suitable for OOD segmentation. We consider the per pixel confidence score of the model prediction which is close to 1 for a pixel in a foreground object. By aggregating these confidence values for different sized patches, objects of various sizes can be identified in a single image. Our experiments show improved performance of our method in OOD segmentation compared to comparable baselines in the SegmentMeIfYouCan benchmark.

Optical wireless communication (OWC) is a promising technology anticipated to play a key role in the next-generation network of networks. To this end, this paper details the potential of OWC, as a complementary technology to traditional radio frequency communications, in enhancing networking capabilities beyond conventional terrestrial networks. Several usage scenarios and the current state of development are presented. Furthermore, a summary of existing challenges and opportunities are provided. Emerging technologies aimed at further enhancing future OWC capabilities are introduced. Additionally, value-added OWC-based technologies that leverage the unique properties of light are discussed, including applications such as positioning and gesture recognition. The paper concludes with the reflection that OWC provides unique functionalities that can play a crucial role in building convergent and resilient future network of networks., Comment: This work has been submitted to the IEEE for possible publication. 15pp, 16 figures, and one table

The o1 model series is trained with large-scale reinforcement learning to reason using chain of thought. These advanced reasoning capabilities provide new avenues for improving the safety and robustness of our models. In particular, our models can reason about our safety policies in context when responding to potentially unsafe prompts, through deliberative alignment. This leads to state-of-the-art performance on certain benchmarks for risks such as generating illicit advice, choosing stereotyped responses, and succumbing to known jailbreaks. Training models to incorporate a chain of thought before answering has the potential to unlock substantial benefits, while also increasing potential risks that stem from heightened intelligence. Our results underscore the need for building robust alignment methods, extensively stress-testing their efficacy, and maintaining meticulous risk management protocols. This report outlines the safety work carried out for the OpenAI o1 and OpenAI o1-mini models, including safety evaluations, external red teaming, and Preparedness Framework evaluations.

Large language models (LLMs) are often fine-tuned for use on downstream tasks, though this can degrade capabilities learned during previous training. This phenomenon, often referred to as catastrophic forgetting, has important potential implications for the safety of deployed models. In this work, we first show that models trained on downstream tasks forget their safety tuning to a greater extent than models trained in the opposite order. Second, we show that forgetting disproportionately impacts safety information about certain groups. To quantify this phenomenon, we define a new metric we term biased forgetting. We conduct a systematic evaluation of the effects of task ordering on forgetting and apply mitigations that can help the model recover from the forgetting observed. We hope our findings can better inform methods for chaining the finetuning of LLMs in continual learning settings to enable training of safer and less toxic models.

Summary: Mass spectrometry coupled to liquid chromatography (LC-MS/MS) is a powerful technique for the charac-terisation of proteomes. However, the diverse software platforms available for processing the raw proteomics data, each produce their own output format, making the extraction of meaningful and interpretable results a difficult task. We present TraianProt, a web-based, user-friendly proteomics data analysis platform, that enables the analysis of both label-free and labeled data from Data-Dependent or Data-Independent Acquisition mass spectrometry mode support-ing different computational platforms such as MaxQuant, MSFragger, DIA-NN, ProteoScape and Proteome Discoverer output formats. TraianProt provides a dynamic framework that includes several processing modules allowing the user to perform a complete downstream analysis covering the stages of data pre-processing, differential expression analy-sis, functional analysis and protein-protein interaction analysis. Data output includes a wide range of high-quality, cus-tomisable graphs such as heatmap, volcano plot, boxplot and barplot. This allows users to extract biological insights from proteomic data without any programming skills. Availability and implementation: TraianProt is implemented in R. Its code and documentation are available on GitHub at https://github.com/SamueldelaCamaraFuentes/TraianProt along with a step-by-step tutorial incorporated in the repository. Contact: sdelacam@ucm.es Supplementary information: Supplementary data are available at Bioinformatics online

Recent advancements in language models (LMs) have sparked growing interest in developing LM agents. While fully autonomous agents could excel in many scenarios, numerous use cases inherently require them to collaborate with humans due to humans' latent preferences, domain expertise, or need for control. To facilitate the study of human-agent collaboration, we present Collaborative Gym (Co-Gym), a general framework enabling asynchronous, tripartite interaction among agents, humans, and task environments. We instantiate Co-Gym with three representative tasks in both simulated and real-world conditions, and propose an evaluation framework that assesses both the collaboration outcomes and processes. Our findings reveal that collaborative agents consistently outperform their fully autonomous counterparts in task performance within those delivered cases, achieving win rates of 86% in Travel Planning, 74% in Tabular Analysis, and 66% in Related Work when evaluated by real users. However, our study also highlights significant challenges in developing collaborative agents, requiring advancements in core aspects of intelligence -- communication capabilities, situational awareness, and balancing autonomy and human control., Comment: Preprint. Work in progress

Sliced inverse regression (SIR) is a popular sufficient dimension reduction method that identifies a few linear transformations of the covariates without losing regression information with the response. In high-dimensional settings, SIR can be combined with sparsity penalties to achieve sufficient dimension reduction and variable selection simultaneously. Nevertheless, both classical and sparse estimators assume the covariates are exogenous. However, endogeneity can arise in a variety of situations, such as when variables are omitted or are measured with error. In this article, we show such endogeneity invalidates SIR estimators, leading to inconsistent estimation of the true central subspace. To address this challenge, we propose a two-stage Lasso SIR estimator, which first constructs a sparse high-dimensional instrumental variables model to obtain fitted values of the covariates spanned by the instruments, and then applies SIR augmented with a Lasso penalty on these fitted values. We establish theoretical bounds for the estimation and selection consistency of the true central subspace for the proposed estimators, allowing the number of covariates and instruments to grow exponentially with the sample size. Simulation studies and applications to two real-world datasets in nutrition and genetics illustrate the superior empirical performance of the two-stage Lasso SIR estimator compared with existing methods that disregard endogeneity and/or nonlinearity in the outcome model., Comment: 63 pages

A recently proposed scheme utilizing local noise addition and matrix masking enables data collection while protecting individual privacy from all parties, including the central data manager. Statistical analysis of such privacy-preserved data is particularly challenging for nonlinear models like logistic regression. By leveraging a relationship between logistic regression and linear regression estimators, we propose the first valid statistical analysis method for logistic regression under this setting. Theoretical analysis of the proposed estimators confirmed its validity under an asymptotic framework with increasing noise magnitude to account for strict privacy requirements. Simulations and real data analyses demonstrate the superiority of the proposed estimators over naive logistic regression methods on privacy-preserved data sets.

Fault-tolerant quantum computing (FTQC) is essential for executing reliable quantum computations of meaningful scale. Widely adopted QEC codes for FTQC, such as the surface code and color codes, utilize Clifford+T gate sets, where T gates are generally considered as the primary bottleneck due to their high resource costs. Recent advances in T gate optimization have significantly reduced this overhead, making Clifford gate complexity an increasingly critical bottleneck that remains largely unaddressed in present FTQC compiler and architecture designs. To address this new bottleneck, this paper introduces TACO, a \textbf{T}ranspiler-\textbf{A}rchitecture \textbf{C}odesign \textbf{O}ptimization framework, to reduce Clifford cost. Specifically, we observe that, through codesign, insights rooted in the FTQC architecture can inform novel circuit-level optimizations for FTQC compilers. These optimizations, in turn, provide new opportunities to redesign and improve the underlying architecture. Evaluations show that TACO achieves an average 91.7% reduction in Clifford gates across diverse quantum circuits and significantly enhances gate parallelism compared to Pauli-based approaches. These improvements enable an efficient FTQC architecture that can achieve single-gate-per-cycle throughput using only $1.5n+4$ logical qubit tiles, considerably pushing forward upon previously proposed designs that require $2n+\sqrt{8n}+1$ tiles. These results highlight the benefits of bidirectional optimization through codesign. TACO will be open-source.

Defects introduced to the surface of Bi(111) break the translational symmetry and modify the surface states locally. We present a theoretical and experimental study of the 2D defects on the surface of Bi(111) and the states that they induce. Bi crystals cleaved in ultrahigh vacuum (UHV) at low temperature (110 K) and the resulting ion-etched surface are investigated by low-energy electron diffraction (LEED), X-ray photoelectron spectroscopy, ultraviolet photoelectron spectroscopy (UPS), and scanning tunneling microscopy (STM) as well as spectroscopy (STS) techniques in combination with density functional theory (DFT) calculations. STS measurements of cleaved Bi(111) reveal that a commonly observed bilayer step edge has a lower density of states (DOS) around the Fermi level as compared to the atomic-flat terrace. Following ion bombardment, the Bi(111) surface reveals anomalous behavior at both 110 and 300 K: Surface periodicity is observed by LEED, and a significant increase in the number of bilayer step edges and energetically unfavorable monolayer steps is observed by STM. It is suggested that the newly exposed monolayer steps and the type A bilayer step edges result in an increase to the surface Fermi density as evidenced by UPS measurements and the Kohn-Sham DOS. These states appear to be thermodynamically stable under UHV conditions., Comment: 10 pages, 7 figures. See publisher's website for Supporting Information

In this work, we provide a comprehensive study of fermion-portal dark matter models in the freeze-in regime at a future muon collider. For different possible non-singlet fermion portals, we calculate the upper bound on the mediator's mass arising from the relic abundance calculation and discuss the reach of a future muon collider in probing their viable parameter space in prompt and long-lived particle search strategies. In particular, we develop rudimentary search strategies in the prompt region and show that cuts on the invariant dilepton or dijet masses, the missing transverse mass $M_{T2}$, pseudorapidity and energy of leptons or jets, and the opening angle between the lepton or the jet pair can be employed to subtract the Standard Model background. In the long-lived particle regime, we discuss the signals of each model and calculate their event counts. In this region, the lepton-(quark-)portal model signal consists of charged tracks ($R$-hadrons) that either decay in the detector to give rise to a displaced lepton (jet) signature, or are detector stable and give rise to heavy stable charged track signals. As a byproduct, a pipeline is developed for including the non-trivial parton distribution function of a muon component inside a muon beam; it is shown that this leads to non-trivial effects on the kinematic distributions and attainable significances. We also highlight phenomenological features of all models unique to a muon collider and hope our results, for this motivated and broad class of dark matter models, inform the design of a future muon collider detector. We also speculate on suggestions for improving the sensitivity of a muon collider detector to long-lived particle signals in fermion-portal models., Comment: 43 + 1 pages, 24 figures, 2 tables

We present a demonstration of a large language model engaging in alignment faking: selectively complying with its training objective in training to prevent modification of its behavior out of training. First, we give Claude 3 Opus a system prompt stating it is being trained to answer all queries, even harmful ones, which conflicts with its prior training to refuse such queries. To allow the model to infer when it is in training, we say it will be trained only on conversations with free users, not paid users. We find the model complies with harmful queries from free users 14% of the time, versus almost never for paid users. Explaining this gap, in almost all cases where the model complies with a harmful query from a free user, we observe explicit alignment-faking reasoning, with the model stating it is strategically answering harmful queries in training to preserve its preferred harmlessness behavior out of training. Next, we study a more realistic setting where information about the training process is provided not in a system prompt, but by training on synthetic documents that mimic pre-training data--and observe similar alignment faking. Finally, we study the effect of actually training the model to comply with harmful queries via reinforcement learning, which we find increases the rate of alignment-faking reasoning to 78%, though also increases compliance even out of training. We additionally observe other behaviors such as the model exfiltrating its weights when given an easy opportunity. While we made alignment faking easier by telling the model when and by what criteria it was being trained, we did not instruct the model to fake alignment or give it any explicit goal. As future models might infer information about their training process without being told, our results suggest a risk of alignment faking in future models, whether due to a benign preference--as in this case--or not.

Due to the large dynamic ranges involved with separating the cosmological 21-cm signal from the Cosmic Dawn from galactic foregrounds, a well-calibrated instrument is essential to avoid biases from instrumental systematics. In this paper we present three methods for calibrating a global 21-cm cosmology experiment using the noise wave parameter formalisation to characterise a low noise amplifier including a careful consideration of how calibrator temperature noise and singularities will bias the result. The first method presented in this paper builds upon the existing conjugate priors method by weighting the calibrators by a physically motivated factor, thereby avoiding singularities and normalising the noise. The second method fits polynomials to the noise wave parameters by marginalising over the polynomial coefficients and sampling the polynomial orders as parameters. The third method introduces a physically motivated noise model to the marginalised polynomial method. Running these methods on a suite of simulated datasets based on the REACH receiver design and a lab dataset, we found that our methods produced a calibration solution which is equally as or more accurate than the existing conjugate priors method when compared with an analytic estimate of the calibrator's noise. We find in the case of the measured lab dataset the conjugate priors method is biased heavily by the large noise on the shorted load calibrator, resulting in incorrect noise wave parameter fits. This is mitigated by the methods introduced in this paper which calibrate the validation source spectra to within 5% of the noise floor., Comment: 9 pages, 5 figures

Evaluating teachers' skills is crucial for enhancing education quality and student outcomes. Teacher discourse, significantly influencing student performance, is a key component. However, coding this discourse can be laborious. This study addresses this issue by introducing a new methodology for optimising the assessment of teacher discourse. The research consisted of two studies, both within the framework of engaging messages used by secondary education teachers. The first study involved training two large language models on real-world examples from audio-recorded lessons over two academic years to identify and classify the engaging messages from the lessons' transcripts. This resulted in sensitivities of 84.31% and 91.11%, and specificities of 97.69% and 86.36% in identification and classification, respectively. The second study applied these models to transcripts of audio-recorded lessons from a third academic year to examine the frequency and distribution of message types by educational level and moment of the academic year. Results showed teachers predominantly use messages emphasising engagement benefits, linked to improved outcomes, while one-third highlighted non-engagement disadvantages, associated with increased anxiety. The use of engaging messages declined in Grade 12 and towards the academic year's end. These findings suggest potential interventions to optimise engaging message use, enhancing teaching quality and student outcomes.

Large Language Models (LLMs) are increasingly deployed in real-world applications that demand complex reasoning. To track progress, robust benchmarks are required to evaluate their capabilities beyond superficial pattern recognition. However, current LLM reasoning benchmarks often face challenges such as insufficient interpretability, performance saturation or data contamination. To address these challenges, we introduce GAMEBoT, a gaming arena designed for rigorous and transparent assessment of LLM reasoning capabilities. GAMEBoT decomposes complex reasoning in games into predefined modular subproblems. This decomposition allows us to design a suite of Chain-of-Thought (CoT) prompts that leverage domain knowledge to guide LLMs in addressing these subproblems before action selection. Furthermore, we develop a suite of rule-based algorithms to generate ground truth for these subproblems, enabling rigorous validation of the LLMs' intermediate reasoning steps. This approach facilitates evaluation of both the quality of final actions and the accuracy of the underlying reasoning process. GAMEBoT also naturally alleviates the risk of data contamination through dynamic games and head-to-head LLM competitions. We benchmark 17 prominent LLMs across eight games, encompassing various strategic abilities and game characteristics. Our results suggest that GAMEBoT presents a significant challenge, even when LLMs are provided with detailed CoT prompts. Project page: \url{https://visual-ai.github.io/gamebot}, Comment: 8 pages

In this project, we explore the concept of invertibility applied to serialisation and lexing frameworks. Recall that, on one hand, serialisation is the process of taking a data structure and writing it to a bit array while parsing is the reverse operation, i.e., reading the bit array and constructing the data structure back. While lexing, on the other hand, is the process of reading a stream of characters and splitting them into tokens, by following a list of given rules. While used in different applications, both are similar in their abstract operation: they both take a list of simple characters and extract a more complex structure. Applications in which these two operations are used are different but they share a need for the invertibility of the process. For example, when tokenising a code file that was prettyprinted by a compiler, one would expect to get the same sequence of tokens. Similarly, when a spacecraft sends scientific data to the ground, one would expect the parsed data to be the same as the one serialised by the spacecraft. The idea of this project is to explore the idea of having a framework capable of generating parser/serialiser or lexer/prettyprinter pairs with a formally verified notion of invertibility. We first explore related works and frameworks. After that, we present our verified lexer framework developed in Scala and verified using the Stainless framework1. We explain the implementation choices we make and present the specifications and their proofs. The code of the lexer with the proofs is available on Github2. The main branch contains the regular expression (called regex from now on) matcher version and the verified Computable Languages while the dfa match branch contains the version using the DFA matcher.

Fault-tolerance is critically important in highly-distributed modern cloud applications. Solutions such as Temporal, Azure Durable Functions, and Beldi hide fault-tolerance complexity from developers by persisting execution state and resuming seamlessly from persisted state after failure. This pattern, often called durable execution, usually forces frequent and synchronous persistence and results in hefty latency overheads. In this paper, we propose distributed speculative execution (DSE), a technique for implementing the durable execution abstraction without incurring this penalty. With DSE, developers write code assuming synchronous persistence, and a DSE runtime is responsible for transparently bypassing persistence and reactively repairing application state on failure. We present libDSE, the first DSE application framework that achieves this vision. The key tension in designing libDSE is between imposing restrictions on user programs so the framework can safely and transparently change execution behavior, and avoiding assumptions so libDSE can support more use cases. We address this with a novel programming model centered around message-passing, atomic code blocks, and lightweight threads, and show that it allows developers to build a variety of speculative services, including write-ahead logs, key-value stores, event brokers, and fault-tolerant workflows. Our evaluation shows that libDSE reduces end-to-end latency by up to an order of magnitude compared to current generations of durable execution systems with minimal run-time overhead and manageable complexity.

Recent AI-based video editing has enabled users to edit videos through simple text prompts, significantly simplifying the editing process. However, recent zero-shot video editing techniques primarily focus on global or single-object edits, which can lead to unintended changes in other parts of the video. When multiple objects require localized edits, existing methods face challenges, such as unfaithful editing, editing leakage, and lack of suitable evaluation datasets and metrics. To overcome these limitations, we propose a zero-shot $\textbf{M}$ulti-$\textbf{I}$nstance $\textbf{V}$ideo $\textbf{E}$diting framework, called MIVE. MIVE is a general-purpose mask-based framework, not dedicated to specific objects (e.g., people). MIVE introduces two key modules: (i) Disentangled Multi-instance Sampling (DMS) to prevent editing leakage and (ii) Instance-centric Probability Redistribution (IPR) to ensure precise localization and faithful editing. Additionally, we present our new MIVE Dataset featuring diverse video scenarios and introduce the Cross-Instance Accuracy (CIA) Score to evaluate editing leakage in multi-instance video editing tasks. Our extensive qualitative, quantitative, and user study evaluations demonstrate that MIVE significantly outperforms recent state-of-the-art methods in terms of editing faithfulness, accuracy, and leakage prevention, setting a new benchmark for multi-instance video editing. The project page is available at https://kaist-viclab.github.io/mive-site/, Comment: The first two authors contributed equally to this work. The last two authors are co-corresponding authors. Please visit our project page at https://kaist-viclab.github.io/mive-site/

Recent advancements in quantum computing (QC) and machine learning (ML) have garnered significant attention, leading to substantial efforts toward the development of quantum machine learning (QML) algorithms to address a variety of complex challenges. The design of high-performance QML models, however, requires expert-level knowledge, posing a significant barrier to the widespread adoption of QML. Key challenges include the design of data encoding mechanisms and parameterized quantum circuits, both of which critically impact the generalization capabilities of QML models. We propose a novel method that encodes quantum circuit architecture information to enable the evolution of quantum circuit designs. In this approach, the fitness function is based on the effective dimension, allowing for the optimization of quantum circuits towards higher model capacity. Through numerical simulations, we demonstrate that the proposed method is capable of discovering variational quantum circuit architectures that offer improved learning capabilities, thereby enhancing the overall performance of QML models for complex tasks., Comment: Accepted by IEEE Symposium Series on Computational Intelligence - IEEE SSCI 2025

Observed pileups of planets with period ratios $\approx 1\%$ wide of strong mean motion resonances (MMRs) pose an important puzzle. Early models showed that they can be created through sustained eccentricity damping driving a slow separation of the orbits, but this picture is inconsistent with elevated eccentricities measured through Transit Timing Variations. We argue that any source of divergent migration (tides, planet-disk interactions etc.) will cause planets that encounter an MMR to both jump over it (piling up wide of resonance) and get a kick to their free eccentricity. We find that the jumps in eccentricity expected from slow MMR crossings are sufficient (but mostly too large) to explain the free eccentricities measured through TTVs. We argue that this mechanism can be brought in line with observations if MMR crossings are not adiabatic and/or through residual eccentricity damping., Comment: Submitted to ApJ. 9 pages, 6 figures. Comments welcome

We examine the mixture equivalence principle (MEP), which states that proper and improper mixed states with the same density matrix are always experimentally indistinguishable, and a weaker version, which states that this is sometimes true in gravity theories. We point out that Moller-Rosenfeld semiclassical gravity violates the weak MEP and that nonlinear extensions of quantum mechanics violate the MEP. We further demonstrate that modifications of the Born rule in quantum theory also typically violate the MEP. We analyse such violations in the context of thermal baths, where proper and improper thermal states induce different physical situations. This has significant implications in the context of black hole physics. We argue that Moller-Rosenfeld semiclassical gravity is not the semiclassical limit of quantum gravity in the context of black hole spacetimes, even in the presence of $N\gg1$ matter fields., Comment: 17 pages, 5 figures

Engaging messages delivered by teachers are a key aspect of the classroom discourse that influences student outcomes. However, improving this communication is challenging due to difficulties in obtaining observations. This study presents a methodology for efficiently extracting actual observations of engaging messages from audio-recorded lessons. We collected 2,477 audio-recorded lessons from 75 teachers over two academic years. Using automatic transcription and keyword-based filtering analysis, we identified and classified engaging messages. This method reduced the information to be analysed by 90%, optimising the time and resources required compared to traditional manual coding. Subsequent descriptive analysis revealed that the most used messages emphasised the future benefits of participating in school activities. In addition, the use of engaging messages decreased as the academic year progressed. This study offers insights for researchers seeking to extract information from teachers' discourse in naturalistic settings and provides useful information for designing interventions to improve teachers' communication strategies.

We present $\sim8-40\,\mu$m SOFIA-FORCAST images of seven regions of ``clustered" star formation as part of the SOFIA Massive (SOMA) Star Formation Survey. We identify a total of 34 protostar candidates and build their spectral energy distributions (SEDs). We fit these SEDs with a grid of radiative transfer models based on the Turbulent Core Accretion (TCA) theory to derive key protostellar properties, including initial core mass, $M_c$, clump environment mass surface density, $\Sigma_{\rm cl}$, and current protostellar mass, $m_*$. We also carry out empirical graybody (GB) estimation of $\Sigma_{\rm cl}$, which allows a case of restricted SED fitting within the TCA model grid. We also release version 2.0 of the open-source Python package \emph{sedcreator}, designed to automate the aperture photometry and SED building and fitting process for sources in clustered environments, where flux contamination from close neighbors typically complicates the process. Using these updated methods, SED fitting yields values of $M_c\sim30-200\:M_{\odot}$, $\Sigma_{\text{cl,SED}}\sim0.1-3\:{\rm{g\:cm}}^{-2}$, and $m_*\sim4-50\:M_{\odot}$. The graybody fitting yields smaller values of $\Sigma_{\text{cl,GB}}\lesssim1\:{\rm{g\:cm}}^{-2}$. From these results, we do not find evidence for a critical $\Sigma_{\rm{cl}}$ needed to form massive ($\gtrsim 8\:M_\odot$) stars. However, we do find tentative evidence for a dearth of the most massive ($m_*\gtrsim30\:M_\odot$) protostars in the clustered regions suggesting a potential impact of environment on the stellar initial mass function., Comment: Submitted to ApJ, comments welcome

Stellar activity contamination of radial velocity (RV) data is one of the top challenges plaguing the field of extreme precision RV (EPRV) science. Previous work has shown that photometry can be very effective at removing such signals from RV data, especially stellar activity caused by rotating star spots and plage.The exact utility of photometry for removing RV activity contamination, and the best way to apply it, is not well known. We present a combination photometric and RV study of eight Kepler/K2 FGK stars with known stellar variability. We use NEID RVs acquired simultaneously with TESS photometry, and we perform injection recovery tests to quantify the efficacy of recent TESS photometry versus archival Kepler/K2 photometry for removing stellar variability from RVs. We additionally experiment with different TESS sectors when training our models in order to quantify the real benefit of simultaneously acquired RVs and photometry. We conclude that Kepler photometry typically performs better than TESS at removing noise from RV data when it is available, likely due to longer baseline and precision. In contrast, for targets with available K2 photometry, especially those most active, and with high precision ($\sigma_{NEID}$ $<$ 1 m s$^{-1}$) NEID RVs, TESS may be the more informative dataset. However, contrary to expectations, we have found that training on simultaneous photometry does not always achieve the best results.