Search

Machine learning is becoming a widely used technique with a impressive growth due to the diversity of problem of societal interest where it can offer practical solutions. This increase of applications and required resources start to become limited by present day hardware technologies. Indeed, novel machine learning subjects such as large language models or high resolution image recognition raise the question of large computing time and energy cost of the required computation. In this context, optical platforms have been designed for several years with the goal of developing more efficient hardware for machine learning. Among different explored platforms, optical free-space propagation offers various advantages: parallelism, low energy cost and computational speed. Here, we present a new design combining the strong and tunable nonlinear properties of a light beam propagating through a hot atomic vapor with an Extreme Learning Machine model. We numerically and experimentally demonstrate the enhancement of the training using such free-space nonlinear propagation on a MNIST image classification task. We point out different experimental hyperparameters that can be further optimized to improve the accuracy of the platform., Comment: 10 pages, 3 figures, 1 table

The evaluation of cell tracking results steers the development of tracking methods, significantly impacting biomedical research. This is quantitatively achieved by means of evaluation metrics. Unfortunately, current metrics favor local correctness and weakly reward global coherence, impeding high-level biological analysis. To also foster global coherence, we propose the CHOTA metric (Cell-specific Higher Order Tracking Accuracy) which unifies the evaluation of all relevant aspects of cell tracking: cell detections and local associations, global coherence, and lineage tracking. We achieve this by introducing a new definition of the term 'trajectory' that includes the entire cell lineage and by including this into the well-established HOTA metric from general multiple object tracking. Furthermore, we provide a detailed survey of contemporary cell tracking metrics to compare our novel CHOTA metric and to show its advantages. All metrics are extensively evaluated on state-of-the-art real-data cell tracking results and synthetic results that simulate specific tracking errors. We show that CHOTA is sensitive to all tracking errors and gives a good indication of the biologically relevant capability of a method to reconstruct the full lineage of cells. It introduces a robust and comprehensive alternative to the currently used metrics in cell tracking. Python code is available at https://github.com/CellTrackingChallenge/py-ctcmetrics ., Comment: Accepted at BIC Workshop at European Conference on Computer Vision 2024, 14 pages, 4 figures, 2 tables

Existing techniques for monocular 3D detection have a serious restriction. They tend to perform well only on a limited set of benchmarks, faring well either on ego-centric car views or on traffic camera views, but rarely on both. To encourage progress, this work advocates for an extended evaluation of 3D detection frameworks across different camera perspectives. We make two key contributions. First, we introduce the CARLA Drone dataset, CDrone. Simulating drone views, it substantially expands the diversity of camera perspectives in existing benchmarks. Despite its synthetic nature, CDrone represents a real-world challenge. To show this, we confirm that previous techniques struggle to perform well both on CDrone and a real-world 3D drone dataset. Second, we develop an effective data augmentation pipeline called GroundMix. Its distinguishing element is the use of the ground for creating 3D-consistent augmentation of a training image. GroundMix significantly boosts the detection accuracy of a lightweight one-stage detector. In our expanded evaluation, we achieve the average precision on par with or substantially higher than the previous state of the art across all tested datasets.

Dynamic Vision Sensors (DVS) have emerged as a revolutionary technology with a high temporal resolution that far surpasses RGB cameras. DVS technology draws biological inspiration from photoreceptors and the initial retinal synapse. Our research showcases the potential of additional retinal functionalities to extract visual features. We provide a domain-agnostic and efficient algorithm for ego-motion compensation based on Object Motion Sensitivity (OMS), one of the multiple robust features computed within the mammalian retina. We develop a framework based on experimental neuroscience that translates OMS' biological circuitry to a low-overhead algorithm. OMS processes DVS data from dynamic scenes to perform pixel-wise object motion segmentation. Using a real and a synthetic dataset, we highlight OMS' ability to differentiate object motion from ego-motion, bypassing the need for deep networks. This paper introduces a bio-inspired computer vision method that dramatically reduces the number of parameters by a factor of 1000 compared to prior works. Our work paves the way for robust, high-speed, and low-bandwidth decision-making for in-sensor computations.

The development of large language models leads to the formation of a pre-train-then-align paradigm, in which the model is typically pre-trained on a large text corpus and undergoes a tuning stage to align the model with human preference or downstream tasks. In this work, we investigate the relationship between pre-training and fine-tuning by fine-tuning multiple intermediate pre-trained model checkpoints. Our results on 18 datasets suggest that i) continual pre-training improves the model in a latent way that unveils after fine-tuning; ii) with extra fine-tuning, the datasets that the model does not demonstrate capability gain much more than those that the model performs well during the pre-training stage; iii) although model benefits significantly through supervised fine-tuning, it may forget previously known domain knowledge and the tasks that are not seen during fine-tuning; iv) the model resembles high sensitivity to evaluation prompts after supervised fine-tuning, but this sensitivity can be alleviated by more pre-training.

Quantum networks and distributed quantum computers rely on entanglement generation between photons and long-lived quantum memories. For large-scale architectures, one of the most crucial parameters is the efficiency at which entanglement can be created and detected. Here, we maximize the efficiency for the detection of hybrid entanglement between a nuclear spin qubit in diamond with a photonic time-bin qubit. Our approach relies on an optimal implementation of the photonic qubit analyzer, for which we use a high-speed electro-optic deflector to direct photons deterministically along the optimal interferometer paths. This way, we completely eliminate all cases in which photons are randomly lost due to the propagation in the wrong interferometer path. In this first demonstration experiment, we use nitrogen-vacancy center in diamond, for which we immediately demonstrate the presence of the entanglement. An extension to other spin-photon entanglement systems is straightforward. Further, our scheme can be used in the framework of quantum repeater networks, including spectral and temporal multiplexing strategies. Our results thus pave the way for the future high-performance quantum networks.

Recent advances in retinal neuroscience have fueled various hardware and algorithmic efforts to develop retina-inspired solutions for computer vision tasks. In this work, we focus on a fundamental visual feature within the mammalian retina, Object Motion Sensitivity (OMS). Using DVS data from EV-IMO dataset, we analyze the performance of an algorithmic implementation of OMS circuitry for motion segmentation in presence of ego-motion. This holistic analysis considers the underlying constraints arising from the hardware circuit implementation. We present novel CMOS circuits that implement OMS functionality inside image sensors, while providing run-time re-configurability for key algorithmic parameters. In-sensor technologies for dynamical environment adaptation are crucial for ensuring high system performance. Finally, we verify the functionality and re-configurability of the proposed CMOS circuit designs through Cadence simulations in 180nm technology. In summary, the presented work lays foundation for hardware-algorithm re-engineering of known biological circuits to suit application needs.

Pulsar-timing-array experiments have reported evidence for a stochastic background of nanohertz gravitational waves consistent with the signal expected from a population of supermassive--black-hole binaries. Those analyses assume power-law spectra for intrinsic pulsar noise and for the background, as well as a Hellings--Downs cross-correlation pattern among the gravitational-wave--induced residuals across pulsars. These assumptions are idealizations that may not be realized in actuality. We test them in the NANOGrav 15 yr data set using Bayesian posterior predictive checks: after fitting our fiducial model to real data, we generate a population of simulated data-set replications, and use them to assess whether the optimal-statistic significance, inter-pulsar correlations, and spectral coefficients assume extreme values for the real data when compared to the replications. We confirm that the NANOGrav 15 yr data set is consistent with power-law and Hellings--Downs assumptions. We also evaluate the evidence for the stochastic background using posterior-predictive versions of the frequentist optimal statistic and of Bayesian model comparison, and find comparable significance (3.2\ $\sigma$ and 3\ $\sigma$ respectively) to what was previously reported for the standard statistics. We conclude with novel visualizations of the reconstructed gravitational waveforms that enter the residuals for each pulsar. Our analysis strengthens confidence in the identification and characterization of the gravitational-wave background as reported by NANOGrav., Comment: 20 pages, 14 Figures

In the collective photon emission from atomic clouds both the atomic transition frequency and the decay rate are modified compared to a single isolated atom, leading to the effects of superradiance and subradiance. In this Letter, we analyse the properties of the Euclidean random matrix associated to the radiative dynamics of a cold atomic cloud, previously investigated in the contexts of photon localization and Dicke super- and subradiance. We present evidence of a new type of phase transition, surprisingly controlled by the cooperativeness parameter, rather than the spatial density or the diagonal disorder. The numerical results corroborate the occurrence of such a phase transition at a critical value of the cooperativeness parameter, above which the lower edge of the spectrum vanishes exhibiting a macroscopic accumulation of eigenvalues. Independent evaluations based on the two phenomena provide the same value of the critical cooperativeness parameter., Comment: 8 pages, 6 figures

Synthesizing NeRFs under arbitrary lighting has become a seminal problem in the last few years. Recent efforts tackle the problem via the extraction of physically-based parameters that can then be rendered under arbitrary lighting, but they are limited in the range of scenes they can handle, usually mishandling glossy scenes. We propose RRM, a method that can extract the materials, geometry, and environment lighting of a scene even in the presence of highly reflective objects. Our method consists of a physically-aware radiance field representation that informs physically-based parameters, and an expressive environment light structure based on a Laplacian Pyramid. We demonstrate that our contributions outperform the state-of-the-art on parameter retrieval tasks, leading to high-fidelity relighting and novel view synthesis on surfacic scenes., Comment: Paper accepted and presented at CGI 2024

Active Galactic Nuclei (AGN) are prime candidate sources of the high-energy, astrophysical neutrinos detected by IceCube. This is demonstrated by the real-time multi-messenger detection of the blazar TXS 0506+056 and the recent evidence of neutrino emission from NGC 1068 from a separate time-averaged study. However, the production mechanism of the astrophysical neutrinos in AGN is not well established which can be resolved via correlation studies with photon observations. For neutrinos produced due to photohadronic interactions in AGN, in addition to a correlation of neutrinos with high-energy photons, there would also be a correlation of neutrinos with photons emitted at radio wavelengths. In this work, we perform an in-depth stacking study of the correlation between 15 GHz radio observations of AGN reported in the MOJAVE XV catalog, and ten years of neutrino data from IceCube. We also use a time-dependent approach which improves the statistical power of the stacking analysis. No significant correlation was found for both analyses and upper limits are reported. When compared to the IceCube diffuse flux, at 100 TeV and for a spectral index of 2.5, the upper limits derived are $\sim3\%$ and $\sim9\%$ for the time-averaged and time-dependent case, respectively., Comment: 14 Pages 7 Figures

We present a search for an eV-scale sterile neutrino using 7.5 years of data from the IceCube DeepCore detector. The analysis uses a sample of 21,914 events with energies between 5 and 150 GeV to search for sterile neutrinos through atmospheric muon neutrino disappearance. Improvements in event selection and treatment of systematic uncertainties provide greater statistical power compared to previous DeepCore sterile neutrino searches. Our results are compatible with the absence of mixing between active and sterile neutrino states, and we place constraints on the mixing matrix elements $|U_{\mu 4}|^2 < 0.0534$ and $|U_{\tau 4}|^2 < 0.0574$ at 90% CL under the assumption that $\Delta m^2_{41}\geq 1\;\mathrm{eV^2}$. These null results add to the growing tension between anomalous appearance results and constraints from disappearance searches in the 3+1 sterile neutrino landscape., Comment: 11 pages, 5 figures. To be submitted to Physical Review D

Currently, there are many quantization methods appeared for LLM quantization, yet few are user-friendly and easy to be deployed locally. Packages like TensorRT and Quantohave many underlying structures and self-invoking internal functions, which are not conducive to developers' personalized development and learning for deployment. Therefore, we develop LLMEasyQuant, it is a package aiming to for easy quantization deployment which is user-friendly and suitable for beginners' learning.

Shifting to cycling in urban areas reduces greenhouse gas emissions and improves public health. Street-level bicycle volume information would aid cities in planning targeted infrastructure improvements to encourage cycling and provide civil society with evidence to advocate for cyclists' needs. Yet, the data currently available to cities and citizens often only comes from sparsely located counting stations. This paper extrapolates bicycle volume beyond these few locations to estimate bicycle volume for the entire city of Berlin. We predict daily and average annual daily street-level bicycle volumes using machine-learning techniques and various public data sources. These include app-based crowdsourced data, infrastructure, bike-sharing, motorized traffic, socioeconomic indicators, weather, and holiday data. Our analysis reveals that the best-performing model is XGBoost, and crowdsourced cycling and infrastructure data are most important for the prediction. We further simulate how collecting short-term counts at predicted locations improves performance. By providing ten days of such sample counts for each predicted location to the model, we are able to halve the error and greatly reduce the variability in performance among predicted locations.

Purpose: Phosphorus Magnetic Resonance Spectroscopy (31P MRS) enables non-invasive assessment of energy metabolism, yet its application is hindered by sensitivity limitations. To overcome this, often high magnetic fields are used, leading to challenges such as spatial B_1^+ inhomogeneity and therefore the need for accurate flip angle determination in accelerated acquisitions with short repetition times (T_R). In response to these challenges, we propose a novel short T_R and look-up table-based Double-Angle Method for fast 3D 31P B_1^+ mapping (fDAM). Methods: Our method incorporates 3D weighted stack of spiral gradient echo acquisitions and a frequency-selective pulse to enable efficient B_1^+ mapping based on the phosphocreatine signal at 7T. Protocols were optimised using simulations and validated through phantom experiments. The method was validated in phantom experiments and skeletal muscle applications using a birdcage 1H/31P volume coil. Results: The results of fDAM were compared to the classical DAM (cDAM). A good correlation (r=0.94) was obtained between the two B_1^+ maps. A 3D 31P B_1^+ mapping in the human calf muscle was achieved in about 10 min using a birdcage volume coil, with a 20% extended coverage relative to that of the cDAM (24 min). fDAM also enabled the first full brain coverage 31P 3D B_1^+ mapping in approx. 10 min using a 1 Tx/ 32 Rx coil. Conclusion: fDAM is an efficient method for 31P 3D B_1^+ mapping, showing promise for future applications in rapid 31P MRSI.

Roughly 2% of white dwarfs harbor planetary debris disks detectable via infrared excesses, but only a few percent of these disks show a gaseous component, distinguished by their double-peaked emission at the near-infrared calcium triplet. Previous studies found most debris disks around white dwarfs are variable at 3.4 and 4.5 $\mu$m, but they analyzed only a few of the now 21 published disks showing calcium emission. To test if most published calcium emission disks exhibit large-amplitude stochastic variability in the near-infrared, we use light curves generated from the unWISE images at 3.4 $\mu$m that are corrected for proper motion to characterize the near-infrared variability of these disks against samples of disks without calcium emission, highly variable cataclysmic variables, and 3215 isolated white dwarfs. We find most calcium emission disks are extremely variable: 6/11 with sufficient signal-to-noise show high-amplitude variability in their 3.4-$\mu$m light curves. These results lend further credence to the notion that disks showing gaseous debris in emission are the most collisionally active. Under the assumption that 3.4-$\mu$m variability is characteristic of white dwarfs with dusty debris disks, we generate a catalog of 104 high-confidence near-infrared variable white dwarfs, 84 of which are published as variable for the first time. We do near-infrared spectroscopic follow-up of seven new candidate 3.4-$\mu$m variables, confirming at least one new remnant planetary system, and posit that empirical near-infrared variability can be a discovery engine for debris disks showing gaseous emission., Comment: 27 pages, 8 figures, accepted to The Astrophysical Journal. Supplemental catalogs and light curves available at https://zenodo.org/doi/10.5281/zenodo.12538705

Nobel laureate Philip Anderson and Elihu Abrahams once stated that, "even if machines did contribute to normal science, we see no mechanism by which they could create a Kuhnian revolution and thereby establish a new physical law." In this Perspective, we draw upon insights from the philosophies of science and artificial intelligence (AI) to propose necessary conditions of precisely such a mechanism for generating revolutionary mathematical theories. Recent advancements in AI suggest that satisfying the proposed necessary conditions by machines may be plausible; thus, our proposed necessary conditions also define a moonshot challenge. We also propose a heuristic definition of the intelligibility of mathematical theories to accelerate the development of machine theorists.

Donor-Acceptor (D-A) type thermally activated delayed fluorescence (TADF) emitters which constitute the key functional units in proposed Gen3 Organic Light Emitting Diodes (OLEDs), are sensitive to the rigidity and polarity of their local environment. In particular, the torsional freedom of the D-A dihedral angle and the excited state dipole moments of the occurring charge transfer states, condition a distribution of TADF dynamics over the emitter ensemble, concealed in standard optical spectroscopy. Here we apply spectroscopy on the single molecule level to directly access individual emitter properties, and, thus, bypass the downside of ensemble averaging. By photon correlation data and locally resolved spectral information on single D-A type TADF molecules embedded in technological relevant host materials of different polarity and rigidity, we derive host-dependent characteristics and distributions in the TADF dynamics. Those can be related to the conformational freedom and the dielectric environment imposed by the specific local host rigidity and polarity, thereby pointing out new selection criteria for host-emitter combinations in OLEDs.

Aromatic molecules represent fundamental building blocks in prebiotic chemistry and are contemplated as vital precursors to DNA and RNA nitrogen bases. However, despite the identification of some 300 molecules in extraterrestrial environments, the pathways to pyridine (C5H5N), pyridinyl (C5H4N), and (iso)quinoline (C9H7N) the simplest representative of mono and bicyclic aromatic molecule carrying nitrogen are elusive. Here, we afford compelling evidence on the gas phase formation of methylene amidogen (H2CN) and cyanomethyl (H2CCN) radicals via molecular beam studies and electronic structure calculations. The modeling of the chemistries of Taurus Molecular Cloud (TMC 1) and Titans atmosphere contemplates a complex chain of reactions synthesizing pyridine, pyridinyl, and (iso)quinoline from H2CN and H2CCN at levels of up to 75%. This study affords unique entry points to precursors of DNA and RNA nitrogen bases in hydrocarbon rich extraterrestrial environments thus changing the way we think about the origin of prebiotic molecules in our Galaxy., Comment: Main paper and supplementary information file

The unsteady flow physics of wind-turbine wakes under dynamic forcing conditions are critical to the modeling and control of wind farms for optimal power density. Unsteady forcing in the streamwise direction may be generated by unsteady inflow conditions in the atmospheric boundary layer, dynamic induction control of the turbine, or streamwise surge motions of a floating offshore wind turbine due to floating-platform oscillations. This study seeks to identify the dominant flow mechanisms in unsteady wakes forced by a periodic upstream inflow condition. A theoretical framework for the problem is derived, which describes traveling-wave undulations in the wake radius and streamwise velocity. These dynamics encourage the aggregation of tip vortices into large structures that are advected along in the wake. Flow measurements in the wake of a periodically surging turbine were obtained in an optically accessible towing-tank facility, with an average diameter-based Reynolds number of 300,000 and with surge-velocity amplitudes of up to 40% of the mean inflow velocity. Qualitative agreement between trends in the measurements and model predictions is observed, supporting the validity of the theoretical analyses. The experiments also demonstrate large enhancements in the recovery of the wake relative to the steady-flow case, with wake-length reductions of up to 46.5% and improvements in the available power at 10 diameters downstream of up to 15.7%. These results provide fundamental insights into the dynamics of unsteady wakes and serve as additional evidence that unsteady fluid mechanics can be leveraged to increase the power density of wind farms.

Solid state quantum emitters are a prime candidate in distributed quantum technologies since they inherently provide a spin-photon interface. An ongoing challenge in the field, however, is the low photon extraction due to the high refractive index of typical host materials. This challenge can be overcome using photonic structures. Here, we report the integration of V2 centers in a cavity-based optical antenna. The structure consists of a silver-coated, 135 nm thin 4H-SiC membrane functioning as a planar cavity with a broadband resonance yielding a theoretical photon collection enhancement factor of 34. The planar geometry allows us to identify over 20 single V2 centers at room temperature with a mean (maximum) count rate enhancement factor of 9 (15). Moreover, we observe 10 V2 centers with a mean absorption linewidth below 80MHz at cryogenic temperatures. These results demonstrate a photon collection enhancement that is robust to the lateral emitter position.

The recent IceCube detection of TeV neutrino emission from the nearby active galaxy NGC 1068 suggests that active galactic nuclei (AGN) could make a sizable contribution to the diffuse flux of astrophysical neutrinos. The absence of TeV $\gamma$-rays from NGC 1068 indicates neutrino production in the vicinity of the supermassive black hole, where the high radiation density leads to $\gamma$-ray attenuation. Therefore, any potential neutrino emission from similar sources is not expected to correlate with high-energy $\gamma$-rays. Disk-corona models predict neutrino emission from Seyfert galaxies to correlate with keV X-rays, as they are tracers of coronal activity. Using through-going track events from the Northern Sky recorded by IceCube between 2011 and 2021, we report results from a search for individual and aggregated neutrino signals from 27 additional Seyfert galaxies that are contained in the BAT AGN Spectroscopic Survey (BASS). Besides the generic single power-law, we evaluate the spectra predicted by the disk-corona model. Assuming all sources to be intrinsically similar to NGC 1068, our findings constrain the collective neutrino emission from X-ray bright Seyfert galaxies in the Northern Hemisphere, but, at the same time, show excesses of neutrinos that could be associated with the objects NGC 4151 and CGCG 420-015. These excesses result in a 2.7$\sigma$ significance with respect to background expectations., Comment: 17 pages, 9 figures

Transiently accreting Low Mass X-Ray Binaries have the potential to probe the core composition of their neutron stars via deep crustal heating caused by nuclear reactions. We statistically assess this deep crustal heating scenario, taking into account the various microphysical and astrophysical uncertainties. We find that despite the sizable uncertainties there is the chance to discriminate different compositional scenarios. Several observed sources statistically challenge a minimal hadronic matter composition, where cooling proceeds exclusively via slow modified Urca reactions. Considering here two exemplary extended uniform compositions, namely ultra-dense hadronic matter with direct Urca emission and ungapped quark matter, we find that they are even within uncertainties distinguishable. We show that although exotic forms of matter are generally only expected in an inner core, which could in principle have any size, sufficiently large astrophysical data sets nonetheless have the potential to statistically discriminate compositional scenarios, in particular when further mass measurements become available., Comment: 15 pages, 3 figures

Active Galactic Nuclei (AGN) are promising candidate sources of high-energy astrophysical neutrinos since they provide environments rich in matter and photon targets where cosmic ray interactions may lead to the production of gamma rays and neutrinos. We searched for high-energy neutrino emission from AGN using the $\textit{Swift}$-BAT Spectroscopic Survey (BASS) catalog of hard X-ray sources and 12 years of IceCube muon track data. First, upon performing a stacked search, no significant emission was found. Second, we searched for neutrinos from a list of 43 candidate sources and found an excess from the direction of two sources, Seyfert galaxies NGC 1068 and NGC 4151. We observed NGC 1068 at flux $\phi_{\nu_{\mu}+\bar{\nu}_{\mu}}$ = $4.02_{-1.52}^{+1.58} \times 10^{-11}$ TeV$^{-1}$ cm$^{-2}$ s$^{-1}$ normalized at 1 TeV, with power-law spectral index, $\gamma$ = 3.10$^{+0.26}_{-0.22}$, consistent with previous IceCube results. The observation of a neutrino excess from the direction of NGC 4151 is at a post-trial significance of 2.9$\sigma$. If interpreted as an astrophysical signal, the excess observed from NGC 4151 corresponds to a flux $\phi_{\nu_{\mu}+\bar{\nu}_{\mu}}$ = $1.51_{-0.81}^{+0.99} \times 10^{-11}$ TeV$^{-1}$ cm$^{-2}$ s$^{-1}$ normalized at 1 TeV and $\gamma$ = 2.83$^{+0.35}_{-0.28}$.

The billiard table is modeled as an $n$-dimensional box $[0,a_1]\times [0,a_2]\times \ldots \times [0,a_n] \subset \mathbb{R}^n$, with each side having real-valued lengths $a_i$ that are pairwise commensurable. A ball is launched from the origin in direction $d=(1,1,\ldots,1)$. The ball is reflected if it hits the boundary of the billiard table. It comes to a halt when reaching a corner. We show that the number of intersections of the billiard curve at any given point on the table is either $0$ or a power of $2$. To prove this, we use algebraic and number theoretic tools to establish a bijection between the number of intersections of the billiard curve and the number of satisfying assignments of a specific constraint satisfaction problem.

Code Large Language Models (Code LLMs) have excelled at tasks like code completion but often miss deeper semantics such as execution effects and dynamic states. This paper aims to bridge the gap between Code LLMs' reliance on static text data and the need for thorough semantic understanding for complex tasks like debugging and program repair. We introduce a novel strategy to train Code LLMs with comprehensive semantics, encompassing high-level functional descriptions, local execution effects of individual statements, and overall input/output behavior, thereby linking static code text with dynamic execution states. We begin by collecting PyX, a clean code corpus of fully executable samples with functional descriptions and execution tracing. We propose training Code LLMs to write code and represent and reason about execution behaviors using natural language, mimicking human verbal debugging. This approach led to the development of SemCoder, a Code LLM with only 6.7B parameters, which shows competitive performance with GPT-3.5-turbo on code generation and execution reasoning tasks. SemCoder achieves 81.1% on HumanEval (GPT-3.5-turbo: 76.8%) and 54.5% on CRUXEval-I (GPT-3.5-turbo: 50.3%). We also study the effectiveness of SemCoder's monologue-style execution reasoning compared to concrete scratchpad reasoning, showing that our approach integrates semantics from multiple dimensions more smoothly. Finally, we demonstrate the potential of applying learned semantics to improve Code LLMs' debugging and self-refining capabilities.

We present the first three-parameter fit to a 3+1 sterile neutrino model using 7.634 years of data from the IceCube Neutrino Observatory on $\nu_\mu+\overline{\nu}_\mu$ charged-current interactions in the energy range 500-9976 GeV. Our analysis is sensitive to the mass-squared splitting between the heaviest and lightest mass state ($\Delta m_{41}^2$), the mixing matrix element connecting muon flavor to the fourth mass state ($|U_{\mu4}|^2$), and the element connecting tau flavor to the fourth mass state ($|U_{\tau4}|^2$). Predicted propagation effects in matter enhance the signature through a resonance as atmospheric neutrinos from the Northern Hemisphere traverse the Earth to the IceCube detector at the South Pole. The result is consistent with the no-sterile neutrino hypothesis with a probability of 4.3 %. Profiling the likelihood of each parameter yields the 90 % confidence levels: $ 2.4\,\mathrm{eV}^{2} < \Delta m_{41}^2 <9.6\,\mathrm{eV}^{2} $ , $0.0081 < |U_{\mu4}|^2 < 0.10$ , and $|U_{\tau4}|^2< 0.035$, which narrows the allowed parameter-space for $|U_{\tau4}|^2$. However, the primary result of this analysis is the first map of the 3+1 parameter space exploring the interdependence of $\Delta m_{41}^2$, $|U_{\mu4}|^2$, and $|U_{\tau4}|^2$.

We study Leaky ResNets, which interpolate between ResNets ($\tilde{L}=0$) and Fully-Connected nets ($\tilde{L}\to\infty$) depending on an 'effective depth' hyper-parameter $\tilde{L}$. In the infinite depth limit, we study 'representation geodesics' $A_{p}$: continuous paths in representation space (similar to NeuralODEs) from input $p=0$ to output $p=1$ that minimize the parameter norm of the network. We give a Lagrangian and Hamiltonian reformulation, which highlight the importance of two terms: a kinetic energy which favors small layer derivatives $\partial_{p}A_{p}$ and a potential energy that favors low-dimensional representations, as measured by the 'Cost of Identity'. The balance between these two forces offers an intuitive understanding of feature learning in ResNets. We leverage this intuition to explain the emergence of a bottleneck structure, as observed in previous work: for large $\tilde{L}$ the potential energy dominates and leads to a separation of timescales, where the representation jumps rapidly from the high dimensional inputs to a low-dimensional representation, move slowly inside the space of low-dimensional representations, before jumping back to the potentially high-dimensional outputs. Inspired by this phenomenon, we train with an adaptive layer step-size to adapt to the separation of timescales.

Electronic logbooks contain valuable information about activities and events concerning their associated particle accelerator facilities. However, the highly technical nature of logbook entries can hinder their usability and automation. As natural language processing (NLP) continues advancing, it offers opportunities to address various challenges that logbooks present. This work explores jointly testing a tailored Retrieval Augmented Generation (RAG) model for enhancing the usability of particle accelerator logbooks at institutes like DESY, BESSY, Fermilab, BNL, SLAC, LBNL, and CERN. The RAG model uses a corpus built on logbook contributions and aims to unlock insights from these logbooks by leveraging retrieval over facility datasets, including discussion about potential multimodal sources. Our goals are to increase the FAIR-ness (findability, accessibility, interoperability, and reusability) of logbooks by exploiting their information content to streamline everyday use, enable macro-analysis for root cause analysis, and facilitate problem-solving automation., Comment: 5 pages, 1 figure, 15th International Particle Accelerator Conference

We demonstrate that "natural inflation", also known as "axion inflation", can be compatible with Planck 2018 measurements of the cosmic microwave background, while predicting an exponentially small tensor-to-scalar ratio, e.g., $r\sim 10^{-15}$. The strong suppression of $r$ arises from dynamics of the radial component of the complex scalar field, whose phase is the axion. Such tiny values of $r$ remain well below the threshold for detection by CMB-S4 or Simons Observatory B-mode searches. The model is testable with the running $\alpha_s$ of the spectral index, which is within reach of next-generation CMB and large-scale structure experiments, motivating the running as a primary science goal for future experiments., Comment: 5 pages, 3 figures

This study evaluates the capabilities of ChatGPT versions 3.5 and 4 in generating code across a diverse range of programming languages. Our objective is to assess the effectiveness of these AI models for generating scientific programs. To this end, we asked ChatGPT to generate three distinct codes: a simple numerical integration, a conjugate gradient solver, and a parallel 1D stencil-based heat equation solver. The focus of our analysis was on the compilation, runtime performance, and accuracy of the codes. While both versions of ChatGPT successfully created codes that compiled and ran (with some help), some languages were easier for the AI to use than others (possibly because of the size of the training sets used). Parallel codes -- even the simple example we chose to study here -- also difficult for the AI to generate correctly., Comment: 9 pages, 3 figures

Autonomous tuning of particle accelerators is an active and challenging field of research with the goal of enabling novel accelerator technologies cutting-edge high-impact applications, such as physics discovery, cancer research and material sciences. A key challenge with autonomous accelerator tuning remains that the most capable algorithms require an expert in optimisation, machine learning or a similar field to implement the algorithm for every new tuning task. In this work, we propose the use of large language models (LLMs) to tune particle accelerators. We demonstrate on a proof-of-principle example the ability of LLMs to successfully and autonomously tune a particle accelerator subsystem based on nothing more than a natural language prompt from the operator, and compare the performance of our LLM-based solution to state-of-the-art optimisation algorithms, such as Bayesian optimisation (BO) and reinforcement learning-trained optimisation (RLO). In doing so, we also show how LLMs can perform numerical optimisation of a highly non-linear real-world objective function. Ultimately, this work represents yet another complex task that LLMs are capable of solving and promises to help accelerate the deployment of autonomous tuning algorithms to the day-to-day operations of particle accelerators., Comment: 22 pages, 5 figures

Primordial black holes (PBHs) are hypothetical objects that could have originated from density fluctuations in a very early phase of our Universe. Recent observations restrict the masses that such PBHs could have, if they are to constitute all of dark matter today: $10^{17} \, {\rm g} \leq m \leq 10^{23} \, {\rm g}$. With such low masses, general relativity predicts that the corresponding radii for the PBHs would be atomic or subatomic in size. When captured by a star, such a tiny PBH could exhibit an orbit completely or partially inside the body of the star, without significantly changing its mass for quite a long time. Here we examine the possible trajectories of a PBH that is captured by a Sun-like star. When in motion in the interior of the star, the amount of stellar mass that effectively interacts with the PBH will be a function of its distance to the center of the star. As a consequence, a strong effect on the shape of the orbits emerges, leading to PBH trajectories that could be open or closed, and exhibiting a rich variety of patterns., Comment: 11 pages, 15 figures

We provide supporting details for the search for a 3+1 sterile neutrino using data collected over eleven years at the IceCube Neutrino Observatory. The analysis uses atmospheric muon-flavored neutrinos from 0.5 to 100\, TeV that traverse the Earth to reach the IceCube detector, and finds a best-fit point at $\sin^2(2\theta_{24}) = 0.16$ and $\Delta m^{2}_{41} = 3.5$ eV$^2$ with a goodness-of-fit p-value of 12\% and consistency with the null hypothesis of no oscillations to sterile neutrinos with a p-value of 3.1\%. Several improvements were made over past analyses, which are reviewed in this article, including upgrades to the reconstruction and the study of sources of systematic uncertainty. We provide details of the fit quality and discuss stability tests that split the data for separate samples, comparing results. We find that the fits are consistent between split data sets., Comment: 18 pages, 17 figures, 2 tables. This long-form paper is a companion to the letter "A search for an eV-scale sterile neutrino using improved high-energy {\nu}{\mu} event reconstruction in IceCube."

This Letter presents the result of a 3+1 sterile neutrino search using 10.7 years of IceCube data. We analyze atmospheric muon neutrinos that traverse the Earth with energies ranging from 0.5 to 100 TeV, incorporating significant improvements in modeling neutrino flux and detector response compared to earlier studies. Notably, for the first time, we categorize data into starting and through-going events, distinguishing neutrino interactions with vertices inside or outside the instrumented volume, to improve energy resolution. The best-fit point for a 3+1 model is found to be at $\sin^2(2\theta_{24}) = 0.16$ and $\Delta m^{2}_{41} = 3.5$ eV$^2$, which agrees with previous iterations of this study. The result is consistent with the null hypothesis of no sterile neutrinos with a p-value of 3.1\%., Comment: 9 pages, 3 figures. This letter is supported by the long-form paper "Methods and stability tests associated with the sterile neutrino search using improved high-energy $\nu_\mu$ event reconstruction in IceCube," also appearing on arXiv

Background: Two strains of the endoparasitoid Cotesia typhae present a differential parasitism success on the host, Sesamia nonagrioides. One is virulent on both permissive and resistant host populations, and the other only on the permissive host. This interaction provides a very interesting frame for studying virulence factors. Here, we used a combination of comparative transcriptomic and proteomic analyses to unravel the molecular basis underlying virulence differences between the strains.Results: First, we report that virulence genes are mostly expressed during the nymphal stage of the parasitoid. Especially, proviral genes are broadly up-regulated at this stage, while their expression is only expected in the host. Parasitoid gene expression in the host increases with time, indicating the production of more virulence factors. Secondly, comparison between strains reveals differences in venom composition, with 12 proteins showing differential abundance. Proviral expression in the host displays a strong temporal variability, along with differential patterns between strains. Notably, a subset of proviral genes including protein-tyrosine phosphatases is specifically over-expressed in the resistant host parasitized by the less virulent strain, 24 hours after parasitism. This result particularly hints at host modulation of proviral expression.Conclusions: This study sheds light on the temporal expression of virulence factors of Cotesia typhae, both in the host and in the parasitoid. It also identifies potential molecular candidates driving differences in parasitism success between two strains. Together, those findings provide a path for further exploration of virulence mechanisms in parasitoid wasps, and offer insights into host-parasitoid coevolution.

Cotesia typhae is an eastern African endoparasitoid braconid wasp that targets the larval stage of the lepidopteran stem borer, Sesamia nonagrioides, a maize crop pest in Europe. The French host population is partially resistant to the Makindu strain of the wasp, allowing its development in only 40% of the cases. Resistant larvae can encapsulate the parasitoid and survive the infection. This interaction provides a very interesting frame for investigating the impact of parasitism on host cellular resistance. We characterized the parasitoid ovolarval development in a permissive host and studied the encapsulation process in a resistant host by dissection and histological sectioning compared to that of inert chromatography beads. We measured the total hemocyte count in parasitized and bead-injected larvae over time to monitor the magnitude of the immune reaction. Our results show that parasitism of resistant hosts delayed encapsulation but did not affect immune abilities towards inert beads. Moreover, while bead injection increased total hemocyte count, it remained constant in resistant and permissive larvae. We conclude that while Cotesia spp virulence factors are known to impair the host immune system, our results suggest that passive evasion could also occur.

In recent years, interest in RISC-V computing architectures has moved from academic to mainstream, especially in the field of High Performance Computing where energy limitations are increasingly a concern. As of this year, the first single board RISC-V CPUs implementing the finalized ratified vector specification are being released. The RISC-V vector specification follows in the tradition of vector processors found in the CDC STAR-100, the Cray-1, the Convex C-Series, and the NEC SX machines and accelerators. The family of vector processors offers support for variable-length array processing as opposed to the fixed-length processing functionality offered by SIMD. Vector processors offer opportunities to perform vector-chaining which allows temporary results to be used without the need to resolve memory references. In this work, we use the Octo-Tiger multi-physics, multi-scale, 3D adaptive mesh refinement astrophysics application to study these early RISC-V chips with vector machine support. We report on our experience in porting this modern C++ code (which is built upon several open-source libraries such as HPX and Kokkos) to RISC-V. In addition, we show the impact of the RISC-V Vector extension on a RISC-V single board computer by implementing the std::experimental:simd interface and integrating it with our code. We also compare the application's performance, scalability, and power consumption on desktop-grade RISC-V computer to an A64FX system.

We consider the free propagation geometry of a light beam (or fluid of light) in a multimode waveguide. As a result of the effective photon-photon interactions, the photon fluid thermalizes to an equilibrium state during its conservative propagation. In this configuration, Rayleigh-Jeans (RJ) thermalization and condensation of classical light waves have been recently observed experimentally in graded index multimode optical fibers characterized by a 2D parabolic trapping potential. As well-known, the properties of RJ condensation differ substantially from those of Bose-Einstein (BE) condensation: The condensate fraction decreases quadratically with the temperature for BE condensation, while it decreases linearly for RJ condensation. Furthermore, for quantum particles the heat capacity tends to zero at small temperatures, and it takes a constant value in the classical particle limit at high temperatures. This is in contrast with classical RJ waves, where the specific heat takes a constant value at small temperatures, and tends to vanish above the condensation transition in the normal (uncondensed) state. Here, we reconcile the thermodynamic properties of BE and RJ condensation: By introducing a frequency cut-off inherent to light propagation in a waveguide, we derive generalized expressions of the thermodynamic properties that include the RJ and BE limits as particular cases. We extend the approach to encompass negative temperatures. In contrast to positive temperatures, the specific heat does not display a singular behavior at negative temperatures, reflecting the non-critical nature of the transition to a macroscopic population of the highest energy level. Our work contributes to understanding the quantum-to-classical crossover in the equilibrium properties of light, within a versatile experimental platform based on nonlinear optical propagation in multimode waveguides., Comment: 17 pages, 8 figures