Search

Industrial embedded systems often require specialized hardware. However, software engineers have access to such domain-specific hardware only at the continuous integration (CI) stage and have to use simulated hardware otherwise. This results in a higher proportion of compilation errors at the CI stage than in other types of systems, warranting a deeper study. To this end, we create a CI diagnostics solution called ``Shadow Job'' that analyzes our industrial CI system. We collected over 40000 builds from 4 projects from the product source code and categorized the compilation errors into 14 error types, showing that the five most common ones comprise 89 % of all compilation errors. Additionally, we analyze the resolution time, size, and distance for each error type, to see if different types of compilation errors are easier to localize or repair than others. Our results show that the resolution time, size, and distance are independent of each other. Our research also provides insights into the human effort required to fix the most common industrial compilation errors. We also identify the most promising directions for future research on fault localization., Comment: 12 pages, 10 figures, ICST 2024

Planar Josephson junctions (JJs) have emerged as a promising platform for the realization of topological superconductivity and Majorana zero modes. To obtain robust quasi one-dimensional (1D) topological superconducting states using planar JJs, limiting the number of 1D Andreev bound states' subbands that can be present, and increasing the size of the topological superconducting gap, are two fundamental challenges. It has been suggested that both problems can be addressed by properly designing the interfaces between the JJ's normal region and the superconducting leads. We fabricated Josephson junctions with periodic hole structures on the superconducting contact leads on InAs heterostructures with epitaxial superconducting Al. By depleting the chemical potential inside the hole region with a top gate, we observed an enhancement of the supercurrent across the junction. Such an enhancement is reproduced in theoretical simulations. The theoretical analysis shows that the enhancement of the JJ's critical current is achieved when the hole depletion is such to optimize the matching of quasiparticles' wave-function at the normal/superconductor interface. These results show how the combination of carefully designed patterns for the Al coverage, and external gates, can be successfully used to tune the density and wave functions' profiles in the normal region of the JJ, and therefore open a new avenue to tune some of the critical properties, such as number of subbands and size of the topological gap, that must be optimized to obtain robust quasi 1D superconducting states supporting Majorana bound states.

Parameter-efficient fine-tuning (PEFT) has emerged as an effective method for adapting pre-trained language models to various tasks efficiently. Recently, there has been a growing interest in transferring knowledge from one or multiple tasks to the downstream target task to achieve performance improvements. However, current approaches typically either train adapters on individual tasks or distill shared knowledge from source tasks, failing to fully exploit task-specific knowledge and the correlation between source and target tasks. To overcome these limitations, we propose PEMT, a novel parameter-efficient fine-tuning framework based on multi-task transfer learning. PEMT extends the mixture-of-experts (MoE) framework to capture the transferable knowledge as a weighted combination of adapters trained on source tasks. These weights are determined by a gated unit, measuring the correlation between the target and each source task using task description prompt vectors. To fully exploit the task-specific knowledge, we also propose the Task Sparsity Loss to improve the sparsity of the gated unit. We conduct experiments on a broad range of tasks over 17 datasets. The experimental results demonstrate our PEMT yields stable improvements over full fine-tuning, and state-of-the-art PEFT and knowledge transferring methods on various tasks. The results highlight the effectiveness of our method which is capable of sufficiently exploiting the knowledge and correlation features across multiple tasks., Comment: Accepted to Findings of the ACL 2024

Recent years have witnessed the success of introducing deep learning models to time series forecasting. From a data generation perspective, we illustrate that existing models are susceptible to distribution shifts driven by temporal contexts, whether observed or unobserved. Such context-driven distribution shift (CDS) introduces biases in predictions within specific contexts and poses challenges for conventional training paradigms. In this paper, we introduce a universal calibration methodology for the detection and adaptation of CDS with a trained model. To this end, we propose a novel CDS detector, termed the "residual-based CDS detector" or "Reconditionor", which quantifies the model's vulnerability to CDS by evaluating the mutual information between prediction residuals and their corresponding contexts. A high Reconditionor score indicates a severe susceptibility, thereby necessitating model adaptation. In this circumstance, we put forth a straightforward yet potent adapter framework for model calibration, termed the "sample-level contextualized adapter" or "SOLID". This framework involves the curation of a contextually similar dataset to the provided test sample and the subsequent fine-tuning of the model's prediction layer with a limited number of steps. Our theoretical analysis demonstrates that this adaptation strategy can achieve an optimal bias-variance trade-off. Notably, our proposed Reconditionor and SOLID are model-agnostic and readily adaptable to a wide range of models. Extensive experiments show that SOLID consistently enhances the performance of current forecasting models on real-world datasets, especially on cases with substantial CDS detected by the proposed Reconditionor, thus validating the effectiveness of the calibration approach., Comment: KDD'24 research paper

In cosmological evolution, it is the homogeneous scalar field (inflaton) that drives the universe to expand isotropically and to generate standard model particles. However, to simulate cosmology, atomic gas research has focused on the dynamics of Bose-Einstein condensates (BEC) with continuously applied forces. In this paper we argue a complementary approach needs also to be pursued; we, thus, consider the analogue BEC experiments in a non-driven, closed atomic system. We implement this using a BEC in an optical lattice which, after a quench, freely transitions from an unstable to a stable state. This dynamical evolution displays the counterpart "preheating", "reheating" and "thermalization" phases of cosmology. Importantly, our studies of these analogue processes yield tractable analytic models. Of great utility to the cold atom community, such understanding elucidates the dynamics of non-adiabatic condensate preparation., Comment: 14 pages, 7 figures

We construct a set of hyperonic equations of state (EoS) by assuming SU(3) symmetry within the baryon octet and by using a covariant density functional (CDF) theory approach. The low-density regions of our EoS are constrained by terrestrial experiments, while the high-density regime is modeled by systematically varying the nuclear matter skewness coefficient $Q_{\rm sat}$ and the symmetry energy slope $L_{\rm sym}$. The sensitivity of the EoS predictions is explored in terms of $z$ parameter of the SU(3) symmetric model that modifies the meson-hyperon coupling constants away from their SU(6) symmetric values. Our results show that model EoS based on our approach can support static Tolman-Oppenheimer-Volkof (TOV) masses in the range $2.3$-$2.5\,M_{\odot}$ in the large-$Q_{\rm sat}$ and small-$z$ regime, however, such stars contain only a trace amount of hyperons compared to SU(6) models. We also construct uniformly rotating Keplerian configurations for our model EoS for which the masses of stellar sequences may reach up to $3.0\,M_{\odot}$. These results are used to explore the systematic dependence of the ratio of maximum masses of rotating and static stars, the lower bound on the rotational frequency of the models that will allow secondary masses in the gravitational waves events to be compact stars with $M_2 \lesssim 3.0\,M_{\odot}$ and the strangeness fraction on the model parameters. We conclude that very massive stellar models can be, in principle, constructed within the SU(3) symmetric model, however, they are nucleonic-like as their strangeness fraction drops below 3\%., Comment: 11 pages, 7 figures

Machine learning is applied to investigate the phase transition of two-dimensional complex plasmas. The Langevin dynamics simulation is employed to prepare particle suspensions in various thermodynamic states. Based on the resulted particle positions in two extreme conditions, bitmap images are synthesized and imported to a convolutional neural network (ConvNet) as training sample. As a result, a phase diagram is obtained. This trained ConvNet model can be directly applied to the sequence of the recorded images using video microscopy in the experiments to study the melting.

The creation of a Hamiltonian in the quantum regime which has non-trivial topological features is a central goal of the cold-atom community, enabling widespread exploration of novel phases of quantum matter. A general scheme to synthesize such Hamiltonians is based on dynamical modulation of optical lattices which thereby generate vector potentials. At the same time the modulation can lead to heating and serious difficulties with equilibration. Here we show that these challenges can be overcome by demonstrating how a Hofstadter Bose-Einstein condensate (BEC) can be dynamically realized, using experimental protocols. From Gross-Pitaevskii simulations our study reveals a complex, multistage evolution; this includes a chaotic intermediate "heating" stage followed by a spontaneous reentrance to the BEC. The observed behavior is reminiscent of evolution in cosmological models.

Health economic evaluations often require predictions of survival rates beyond the follow-up period. Parametric survival models can be more convenient for economic modelling than the Cox model. The generalized gamma (GG) and generalized F (GF) distributions are extensive families that contain almost all commonly used distributions with various hazard shapes and arbitrary complexity. In this study, we present a new SAS macro for implementing a wide variety of flexible parametric models including the GG and GF distributions and their special cases, as well as the Gompertz distribution. Proper custom distributions are also supported. Different from existing SAS procedures, this macro not only supports regression on the location parameter but also on ancillary parameters, which greatly increases model flexibility. In addition, the SAS macro supports weighted regression, stratified regression and robust inference. This study demonstrates with several examples how the SAS macro can be used for flexible survival modeling and extrapolation., Comment: 15 pages, 1 figure, 10 tables, accepted by The Clinical Data Science Conference - PHUSE US Connect 2021

Software bugs are common and correcting them accounts for a significant part of costs in the software development and maintenance process. This calls for automatic techniques to deal with them. One promising direction towards this goal is gaining repair knowledge from historical bug fixing examples. Retrieving insights from software development history is particularly appealing with the constant progress of machine learning paradigms and skyrocketing `big' bug fixing data generated through Continuous Integration (CI). In this paper, we present R-Hero, a novel software repair bot that applies continual learning to acquire bug fixing strategies from continuous streams of source code changes, implemented for the single development platform Github/Travis CI. We describe R-Hero, our novel system for learning how to fix bugs based on continual training, and we uncover initial successes as well as novel research challenges for the community.

Nowadays, driven by the increasing concern on diet and health, food computing has attracted enormous attention from both industry and research community. One of the most popular research topics in this domain is Food Retrieval, due to its profound influence on health-oriented applications. In this paper, we focus on the task of cross-modal retrieval between food images and cooking recipes. We present Modality-Consistent Embedding Network (MCEN) that learns modality-invariant representations by projecting images and texts to the same embedding space. To capture the latent alignments between modalities, we incorporate stochastic latent variables to explicitly exploit the interactions between textual and visual features. Importantly, our method learns the cross-modal alignments during training but computes embeddings of different modalities independently at inference time for the sake of efficiency. Extensive experimental results clearly demonstrate that the proposed MCEN outperforms all existing approaches on the benchmark Recipe1M dataset and requires less computational cost., Comment: Accepted to CVPR 2020

Paper documents are widely used as an irreplaceable channel of information in many fields, especially in financial industry, fostering a great amount of demand for systems which can convert document images into structured data representations. In this paper, we present a machine learning framework for data ingestion in document images, which processes the images uploaded by users and return fine-grained data in JSON format. Details of model architectures, design strategies, distinctions with existing solutions and lessons learned during development are elaborated. We conduct abundant experiments on both synthetic and real-world data in State Street. The experimental results indicate the effectiveness and efficiency of our methods.

We show that jet emission from a Bose condensate with periodically driven interactions, a.k.a. "Bose fireworks", contains essential information on the condensate wavefunction, which is difficult to obtain using standard detection methods. We illustrate the underlying physics with two examples. When condensates acquire phase patterns from external potentials or from vortices, the jets display novel sub-structure, such as oscillations or spirals, in their correlations. Through a comparison of theory, numerical simulations and experiments, we show how one can quantitatively extract the phase and the helicity of a condensate from the emission pattern. Our work demonstrating the strong link between jet emission and the underlying quantum system, bears on the recent emphasis on jet sub-structure in particle physics.

Electronic compressibility, the second derivative of ground state energy with respect to total electron number, is a measurable quantity that reveals the interaction strength of a system and can be used to characterize the orderly crystalline lattice of electrons known as the Wigner crystal. Here, we measure the electronic compressibility of individual suspended ultraclean carbon nanotubes in the low-density Wigner crystal regime. Using low-temperature quantum transport measurements, we determine the compressibility as a function of carrier number in nanotubes with varying band gaps. We observe two qualitatively different trends in compressibility versus carrier number, both of which can be explained using a theoretical model of a Wigner crystal that accounts for both the band gap and the confining potential experienced by charge carriers. We extract the interaction strength as a function of carrier number for individual nanotubes and show that the compressibility can be used to distinguish between strongly and weakly interacting regimes., Comment: 20 pages, 6 figures