Search

This study introduces a novel software tool leveraging large language model (LLM) prompts, designed to automate the generation of academic articles from Python code a significant advancement in the fields of biomedical informatics and computer science. Selected for its widespread adoption and analytical versatility, Python served as a foundational proof of concept; however, the underlying methodology and framework exhibit adaptability across various GitHub repo's underlining the tool's broad applicability (Harper 2024). By mitigating the traditionally time-intensive academic writing process, particularly in synthesizing complex datasets and coding outputs, this approach signifies a monumental leap towards streamlining research dissemination. The development was achieved without reliance on advanced language model agents, ensuring high fidelity in the automated generation of coherent and comprehensive academic content. This exploration not only validates the successful application and efficiency of the software but also projects how future integration of LLM agents which could amplify its capabilities, propelling towards a future where scientific findings are disseminated more swiftly and accessibly., Comment: Keywords: automated academic writing, Python code, software tool, article generation, natural language processing, scholarly publishing, code analysis, academic article automation, research dissemination, programming and publishing integration

Characterising the performance of noisy quantum circuits is central to the production of prototype quantum computers and can enable improved quantum error correction that exploits noise biases identified in a quantum device. We describe an implementation of averaged circuit eigenvalue sampling (ACES), a general framework for the scalable noise characterisation of quantum circuits. ACES is capable of simultaneously estimating the Pauli error probabilities of all gates in a Clifford circuit, and captures averaged spatial correlations between gates implemented simultaneously in the circuit. By rigorously analysing the performance of ACES experiments, we derive a figure of merit for their expected performance, allowing us to optimise ACES experimental designs and improve the precision to which we estimate noise given fixed experimental resources. Since the syndrome extraction circuits of quantum error correcting codes are representative components of a fault-tolerant architecture, we demonstrate the scalability and performance of our ACES protocol through circuit-level numerical simulations of the entire noise characterisation procedure for the syndrome extraction circuit of a distance 25 surface code with over 1000 qubits. Our results indicate that detailed noise characterisation methods are scalable to near-term quantum devices. We release our code in the form of the Julia package AveragedCircuitEigenvalueSampling.jl., Comment: 15 pages + 15 page appendix, 22 figures, comments welcome

Emission lines from Rydberg transitions are detected for the first time from a region close to the surface of Betelgeuse. The H30${\alpha}$ line is observed at 231.905 GHz, with a FWHM ~42 km/s and extended wings. A second line at 232.025 GHz (FWHM ~21 km/s), is modeled as a combination of Rydberg transitions of abundant low First Ionization Potential metals. Both H30${\alpha}$ and the Rydberg combined line X30${\alpha}$ are fitted by Voigt profiles, and collisional broadening with electrons may be partly responsible for the Lorentzian contribution, indicating electron densities of a few 10$^8$cm$^{-3}$. X30${\alpha}$ is located in a relatively smooth ring at a projected radius of 0.9x the optical photospheric radius R$_*$, whereas H30${\alpha}$ is more clumpy, reaching a peak at ~1.4R$_*$. We use a semi-empirical thermodynamic atmospheric model of Betelgeuse to compute the 232 GHz (1.29mm) continuum and line profiles making simple assumptions. Photoionized abundant metals dominate the electron density and the predicted surface of continuum optical depth unity at 232 GHz occurs at ~1.3R$_*$, in good agreement with observations. Assuming a Saha-Boltzmann distribution for the level populations of Mg, Si, and Fe, the model predicts that the X30${\alpha}$ emission arises in a region of radially-increasing temperature and turbulence. Inclusion of ionized C and non-LTE effects could modify the integrated fluxes and location of emission. These simulations confirm the identity of the Rydberg transition lines observed towards Betelgeuse, and reveal that such diagnostics can improve future atmospheric models., Comment: ApJLett (accepted). 4 figures

Amortized analysis is a cost analysis technique for data structures in which cost is studied in aggregate, rather than considering the maximum cost of a single operation. Traditionally, amortized analysis has been phrased inductively, in terms of finite sequences of operations. Connecting to prior work on coalgebraic semantics for data structures, we develop the perspective that amortized analysis is naturally viewed coalgebraically in the category of algebras for a cost monad, where a morphism of coalgebras serves as a first-class generalization of potential function suitable for integrating cost and behavior. Using this simple definition, we consider amortization of other sample effects, non-commutative printing and randomization. To support imprecise amortized upper bounds, we adapt our discussion to the bicategorical setting, where a potential function is a colax morphism of coalgebras. We support parallel data structure usage patterns by using coalgebras for an endoprofunctor instead of an endofunctor, combining potential using a monoidal structure on the underlying category. Finally, we compose amortization arguments in the indexed category of coalgebras to implement one amortized data structure in terms of others., Comment: To appear at MFPS 2024

Telecom C-band single photons exhibit the lowest attenuation in optical fibers, enabling long-haul quantum-secured communication. However, efficient coupling with optical fibers is crucial for these single photons to be effective carriers in long-distance transmission. In this work, we demonstrate an efficient fiber-coupled single photon source at the telecom C-band using InAs/InP quantum dots coupled to a tapered nanobeam. The tapered nanobeam structure facilitates directional emission that is mode-matched to a lensed fiber, resulting in a collection efficiency of up to 65% from the nanobeam to a single-mode fiber. Using this approach, we demonstrate single photon count rates of 575 $\pm$ 5 Kcps and a single photon purity of $g^2$ (0) = 0.015 $\pm$ 0.003. Additionally, we demonstrate Hong-Ou Mandel interference from the emitted photons with a visibility of 0.84 $\pm$ 0.06. From these measurements, we determine a photon coherence time of 450 $\pm$ 20 ps, a factor of just 8.3 away from the lifetime limit. This work represents an important step towards the development of telecom C-band single-photon sources emitting bright, pure, and indistinguishable photons, which are necessary to realize fiber-based long-distance quantum networks

We study a cost-aware programming language for higher-order recursion dubbed $\textbf{PCF}_\mathsf{cost}$ in the setting of synthetic domain theory (SDT). Our main contribution relates the denotational cost semantics of $\textbf{PCF}_\mathsf{cost}$ to its computational cost semantics, a new kind of dynamic semantics for program execution that serves as a mathematically natural alternative to operational semantics in SDT. In particular we prove an internal, cost-sensitive version of Plotkin's computational adequacy theorem, giving a precise correspondence between the denotational and computational semantics for complete programs at base type. The constructions and proofs of this paper take place in the internal dependent type theory of an SDT topos extended by a phase distinction in the sense of Sterling and Harper. By controlling the interpretation of cost structure via the phase distinction in the denotational semantics, we show that $\textbf{PCF}_\mathsf{cost}$ programs also evince a noninterference property of cost and behavior. We verify the axioms of the type theory by means of a model construction based on relative sheaf models of SDT., Comment: Accepted at MFPS '24

We show that strong subadditivity provides a simple derivation of the $g$-theorem for the boundary renormalization group flow in two-dimensional conformal field theories. We work out its holographic interpretation and also give a derivation of the $g$-theorem for the case of an interface in two-dimensional conformal field theories. We also geometrically confirm strong subadditivity for holographic duals of conformal field theories on manifolds with boundaries., Comment: 7 pages, 6 figures

We propose a road layout and traffic model, based on last passage percolation (LPP). An easy naive argument shows that coalescence of traffic trajectories is essential to be considered when observing traffic networks around us. This is a fundamental feature in first passage percolation (FPP) models where nearby geodesics naturally coalesce in search of the easiest passage through the landscape. Road designers seek the same in pursuing cost savings, hence FPP geodesics are straightforward candidates to model road layouts. Unfortunately no detailed knowledge is rigorously available on FPP geodesics. To address this, we use exponential LPP instead to build a stochastic model of road traffic and prove certain characteristics thereof. Cars start from every point of the lattice and follow half-infinite geodesics in random directions. Exponential LPP is known to be in the KPZ universality class and it is widely expected that FPP shares very similar properties, hence our findings should equally apply to FPP-based modelling. We address several traffic-related quantities of this model and compare our theorems to real life road networks., Comment: 38 pages, 23 figures

This paper reports the results of the deployment of Rich-State Simulated Populations at Meta for both automated and manual testing. We use simulated users (aka test users) to mimic user interactions and acquire state in much the same way that real user accounts acquire state. For automated testing, we present empirical results from deployment on the Facebook, Messenger, and Instagram apps for iOS and Android Platforms. These apps consist of tens of millions of lines of code, communicating with hundreds of millions of lines of backend code, and are used by over 2 billion people every day. Our results reveal that rich state increases average code coverage by 38\%, and endpoint coverage by 61\%. More importantly, it also yields an average increase of 115\% in the faults found by automated testing. The rich-state test user populations are also deployed in a (continually evolving) Test Universe; a web-enabled simulation platform for privacy-safe manual testing, which has been used by over 21,000 Meta engineers since its deployment in November 2022., Comment: ICSE 2024

In the evolving landscape of clinical informatics, the integration and utilization of software tools developed through governmental funding represent a pivotal advancement in research and application. However, the dispersion of these tools across various repositories, with no centralized knowledge base, poses significant challenges to leveraging their full potential. This study introduces an automated methodology to bridge this gap by systematically extracting GitHub repository URLs from academic papers indexed in arXiv, focusing on the field of clinical informatics. Our approach encompasses querying the arXiv API for relevant papers, cleaning extracted GitHub URLs, fetching comprehensive repository information via the GitHub API, and analyzing repository maturity based on defined metrics such as stars, forks, open issues, and contributors. The process is designed to be robust, incorporating error handling and rate limiting to ensure compliance with API constraints. Preliminary findings demonstrate the efficacy of this methodology in compiling a centralized knowledge base of NIH-funded software tools, laying the groundwork for an enriched understanding and utilization of these resources within the clinical informatics community. We propose the future integration of Large Language Models (LLMs) to generate concise summaries and evaluations of the tools. This approach facilitates the discovery and assessment of clinical informatics tools and also enables ongoing monitoring of new and actively updated repositories, revolutionizing how researchers access and leverage federally funded software. The implications of this study extend beyond simplification of access to valuable resources; it proposes a scalable model for the dynamic aggregation and evaluation of scientific software, encouraging more collaborative, transparent, and efficient research practices in clinical informatics and beyond.

Thin-film lithium niobate is a promising photonic platform for on-chip optical sensing because both nonlinear and linear components can be fabricated within one integrated device. To date, waveguided sample interactions for thin-film lithium niobate are not well explored. Compared to other integrated platforms, lithium niobate's high refractive index, birefringence, and angled sidewalls present unique design challenges for evanescent wave sensing. Here, we compare the performance of the quasi-transverse-electric (TE) and the quasi-transverse-magnetic (TM) mode for sensing on a thin-film lithium niobate rib waveguide with a 5 mM dye-doped polymer cladding pumped at 406 nm. We determine that both modes have propagation losses dominated by scatter, and that the absorption due to the sample only accounts for 3% of the measured losses for both modes. The TM mode has better overlap with the sample than the TE mode, but the TM mode also has a stronger propagation loss due to sidewall and sample induced scattering (32.5 $\pm$ 0.3 dB/cm) compared to the TE mode (23.0 $\pm$ 0.2 dB/cm). The TE mode is, therefore, more appropriate for sensing. Our findings have important implications for on-chip lithium niobate-based sensor designs., Comment: 9 pages, 3 figures (main text); 6 pages, 5 figures (supplemental). Submitted to Optics Express

Physically unclonable functions (PUFs) identify integrated circuits using nonlinearly-related challenge-response pairs (CRPs). Ideally, the relationship between challenges and corresponding responses is unpredictable, even if a subset of CRPs is known. Previous work developed a photonic PUF offering improved security compared to non-optical counterparts. Here, we investigate this PUF's susceptibility to Multiple-Valued-Logic-based machine learning attacks. We find that approximately 1,000 CRPs are necessary to train models that predict response bits better than random chance. Given the significant challenge of acquiring a vast number of CRPs from a photonic PUF, our results demonstrate photonic PUF resilience against such attacks., Comment: 6 pages, 4 figures

In this work we used $\textit{ab-initio}$ random structure searching (AIRSS) to carry out a systematic search for crystalline Na-Ge materials at both 0 and 10 GPa. The high-throughput structural relaxations were accelerated using a machine-learned interatomic potential (MLIP) fit to density-functional theory (DFT) reference data, allowing $\sim$1.5 million structures to be relaxed. At ambient conditions we predict three new Zintl phases, Na$_3$Ge$_2$, Na$_2$Ge and Na$_9$Ge$_4$, to be stable and a number of Ge-rich layered structures to lie in close proximity to the convex hull. The known Na$_\delta$Ge$_{34}$ clathrate and Na$_4$Ge$_{13}$ host-guest structures are found to be relatively stabilized at higher temperature by vibrational contributions to the free energy. Overall, the low energy phases exhibit exceptional structural diversity, with the expected mixture of covalent and ionic bonding confirmed using the electron-localisation function (ELF). The local Ge structural motifs present at each composition were determined using Smooth Overlap of Atomic Positions (SOAP) descriptors and the Ge-K edge was simulated for representatives of each motif, providing a direct link to experimental x-ray absorption spectroscopy (XAS). Two Ge-rich phases are predicted to be stable at 10 GPa; NaGe$_3$ and NaGe$_2$ have simple kagome and simple hexagonal Ge lattices respectively with Na contained in the pores. NaGe$_3$ is isostructural with the MgB$_3$ and MgSi$_3$ family of kagome superconductors and remains dynamically stable at 0 GPa. Removing the Na from NaGe$_2$ results in the hexagonal lonsdalite Ge allotrope, which has a direct band gap., Comment: 11 pages, 8 figures

This paper describes Meta's TestGen-LLM tool, which uses LLMs to automatically improve existing human-written tests. TestGen-LLM verifies that its generated test classes successfully clear a set of filters that assure measurable improvement over the original test suite, thereby eliminating problems due to LLM hallucination. We describe the deployment of TestGen-LLM at Meta test-a-thons for the Instagram and Facebook platforms. In an evaluation on Reels and Stories products for Instagram, 75% of TestGen-LLM's test cases built correctly, 57% passed reliably, and 25% increased coverage. During Meta's Instagram and Facebook test-a-thons, it improved 11.5% of all classes to which it was applied, with 73% of its recommendations being accepted for production deployment by Meta software engineers. We believe this is the first report on industrial scale deployment of LLM-generated code backed by such assurances of code improvement., Comment: 12 pages, 8 figures, 32nd ACM Symposium on the Foundations of Software Engineering (FSE 24)

In this paper we address the following question: How can we use Large Language Models (LLMs) to improve code independently of a human, while ensuring that the improved code - does not regress the properties of the original code? - improves the original in a verifiable and measurable way? To address this question, we advocate Assured LLM-Based Software Engineering; a generate-and-test approach, inspired by Genetic Improvement. Assured LLMSE applies a series of semantic filters that discard code that fails to meet these twin guarantees. This overcomes the potential problem of LLM's propensity to hallucinate. It allows us to generate code using LLMs, independently of any human. The human plays the role only of final code reviewer, as they would do with code generated by other human engineers. This paper is an outline of the content of the keynote by Mark Harman at the International Workshop on Interpretability, Robustness, and Benchmarking in Neural Software Engineering, Monday 15th April 2024, Lisbon, Portugal., Comment: 6 pages, 1 figure, InteNSE 24: ACM International Workshop on Interpretability, Robustness, and Benchmarking in Neural Software Engineering, April, 2024, Lisbon, Portugal

Quantum computing has the potential to provide solutions to problems that are intractable on classical computers, but the accuracy of the current generation of quantum computers suffer from the impact of noise or errors such as leakage, crosstalk, dephasing, and amplitude damping among others. As the access to quantum computers is almost exclusively in a shared environment through cloud-based services, it is possible that an adversary can exploit crosstalk noise to disrupt quantum computations on nearby qubits, even carefully designing quantum circuits to purposely lead to wrong answers. In this paper, we analyze the extent and characteristics of crosstalk noise through tomography conducted on IBM Quantum computers, leading to an enhanced crosstalk simulation model. Our results indicate that crosstalk noise is a significant source of errors on IBM quantum hardware, making crosstalk based attack a viable threat to quantum computing in a shared environment. Based on our crosstalk simulator benchmarked against IBM hardware, we assess the impact of crosstalk attacks and develop strategies for mitigating crosstalk effects. Through a systematic set of simulations, we assess the effectiveness of three crosstalk attack mitigation strategies, namely circuit separation, qubit allocation optimization via reinforcement learning, and the use of spectator qubits, and show that they all overcome crosstalk attacks with varying degrees of success and help to secure quantum computing in a shared platform., Comment: 13 pages, 7 figures

As governments try to navigate education system reform amid an increasingly uncertain future, it has become urgent to articulate how they can better maximize the co-benefits of efforts to build climate resilience, advance environmental sustainability and achieve quality education for all. This paper proposes a seven-dimension framework to leverage potential entry points and address gaps within national education systems to strengthen the resilience and relevance of education to climate change and environmental degradation. [For the background paper, "Quality Education with the Planet in Mind: Towards a Climate-Responsive and Nature-Positive Framework for the Education System of Lower-Income Countries," see ED631325.]

Fault-tolerant architectures aim to reduce the noise of a quantum computation. Despite such architectures being well studied a detailed understanding of how noise is transformed in a fault-tolerant primitive such as magic state injection is currently lacking. We use numerical simulations of logical process tomography on a fault-tolerant gadget that implements a logical T = $Z({\pi}/8)$ gate using magic state injection, to understand how noise characteristics at the physical level are transformed into noise characteristics at the logical level. We show how, in this gadget, a significant phase ($Z$) bias can arise in the logical noise, even with unbiased noise at the physical level. While the magic state injection gadget intrinsically induces biased noise, with extant phase bias being further amplified at the logical level, we identify noisy error correction circuits as a key limiting factor on the magnitude of this logical noise bias. Our approach provides a framework for assessing the detailed noise characteristics, as well as the overall performance, of fault-tolerant logical primitives., Comment: 11+5 pages, 17 figures. Comments welcome

For a static time slice of AdS$_3$ we describe a particular class of minimal surfaces which form trivalent networks of geodesics. Through geometric arguments we provide evidence that these surfaces describe a measure of multipartite entanglement. By relating these surfaces to Ryu-Takayanagi surfaces it can be shown that this multipartite contribution is related to the angles of intersection of the bulk geodesics. A proposed boundary dual, the multi-entropy, generalizes replica trick calculations involving twist operators by considering monodromies with finite group symmetry beyond the cyclic group used for the computation of entanglement entropy. We make progress by providing explicit calculations of Renyi multi-entropy in two dimensional CFTs and geometric descriptions of the replica surfaces for several cases with low genus. We also explore aspects of the free fermion and free scalar CFTs. For the free fermion CFT we examine subtleties in the definition of the twist operators used for the calculation of Renyi multi-entropy. In particular the standard bosonization procedure used for the calculation of the usual entanglement entropy fails and a different treatment is required., Comment: 50 pages, 21 figures. v2: Minor correction and typos v3: Added information about general multi-trace measures, minor corrections

We define a monoidal category $\operatorname{\mathbf{W}}$ and a closely related 2-category $\operatorname{\mathbf{2Weyl}}$ using diagrammatic methods. We show that $\operatorname{\mathbf{2Weyl}}$ acts on the category $\mathbf{TL} :=\bigoplus_n \operatorname{TL}_n\operatorname{-mod}$ of modules over Temperley-Lieb algebras, with its generating 1-morphisms acting by induction and restriction. The Grothendieck groups of $\operatorname{\mathbf{W}}$ and a third category we define $\operatorname{\mathbf W}^\infty$ are closely related to the Weyl algebra. We formulate a sense in which $K_0(\operatorname{\mathbf W}^\infty)$ acts asymptotically on $K_0(\mathbf{TL})$., Comment: 37 pages, many figures. Comments particularly encouraged!

One of the primary impediments hindering the widespread acceptance of autonomous vehicles (AVs) among pedestrians is their limited comprehension of AVs. This study employs virtual reality (VR) to provide pedestrians with an immersive environment for engaging with and comprehending AVs during unmarked midblock multilane crossings. Diverse AV driving behaviors were modeled to exhibit negotiation behavior with a yellow signal indication or non-yielding behavior with a blue signal indication. This paper aims to investigate the impact of various factors, such as AV behavior and signaling, pedestrian past behavior, etc., on pedestrians' perception change of AVs. Before and after the VR experiment, participants completed surveys assessing their perception of AVs, focusing on two main aspects: "Attitude" and "System Effectiveness." The Wilcoxon signed-rank test results demonstrated that both pedestrians' overall attitude score toward AVs and trust in the effectiveness of AV systems significantly increased following the VR experiment. Notably, individuals who exhibited a greater trust in the yellow signals were more inclined to display a higher attitude score toward AVs and to augment their trust in the effectiveness of AV systems. This indicates that the design of the yellow signal instills pedestrians with greater confidence in their interactions with AVs. Further, pedestrians who exhibit more aggressive crossing behavior are less likely to change their perception towards AVs as compared to those pedestrians with more positive crossing behaviors. It is concluded that integrating this paper's devised AV behavior and signaling within an immersive VR setting facilitated pedestrian engagement with AVs, thereby changing their perception of AVs.