Search

The availability of pre-trained models (PTMs) has enabled faster deployment of machine learning across applications by reducing the need for extensive training. Techniques like quantization and distillation have further expanded PTM applicability to resource-constrained IoT hardware. Given the many PTM options for any given task, engineers often find it too costly to evaluate each model's suitability. Approaches such as LogME, LEEP, and ModelSpider help streamline model selection by estimating task relevance without exhaustive tuning. However, these methods largely leave hardware constraints as future work-a significant limitation in IoT settings. In this paper, we identify the limitations of current model recommendation approaches regarding hardware constraints and introduce a novel, hardware-aware method for PTM selection. We also propose a research agenda to guide the development of effective, hardware-conscious model recommendation systems for IoT applications., Comment: Accepted at SERP4IOT'25

Energy-efficient software helps improve mobile device experiences and reduce the carbon footprint of data centers. However, energy goals are often de-prioritized in order to meet other requirements. We take inspiration from recent work exploring the use of large language models (LLMs) for different software engineering activities. We propose a novel application of LLMs: as code optimizers for energy efficiency. We describe and evaluate a prototype, finding that over 6 small programs our system can improve energy efficiency in 3 of them, up to 2x better than compiler optimizations alone. From our experience, we identify some of the challenges of energy-efficient LLM code optimization and propose a research agenda.

Due to the high dimensionality or multimodality that is common in modern astronomy, sampling Bayesian posteriors can be challenging. Several publicly available codes based on different sampling algorithms can solve these complex models, but the execution of the code is not always efficient or fast enough. The article introduces a C language general-purpose code, Nii-C (https://github.com/shengjin/nii-c.git), that implements a framework of Automatic Parallel Tempering Markov Chain Monte Carlo. Automatic in this context means that the parameters that ensure an efficient parallel tempering process can be set by a control system during the initial stages of a sampling process. The auto-tuned parameters consist of two parts, the temperature ladders of all parallel tempering Markov chains and the proposal distributions for all model parameters across all parallel tempering chains. In order to reduce dependencies in the compilation process and increase the code's execution speed, Nii-C code is constructed entirely in the C language and parallelised using the Message-Passing Interface protocol to optimise the efficiency of parallel sampling. These implementations facilitate rapid convergence in the sampling of high-dimensional and multi-modal distributions, as well as expeditious code execution time. The Nii-C code can be used in various research areas to trace complex distributions due to its high sampling efficiency and quick execution speed. This article presents a few applications of the Nii-C code., Comment: 21 pages, 9 figures, accepted for publication in ApJS

Background: Collaborative Software Package Registries (SPRs) are an integral part of the software supply chain. Much engineering work synthesizes SPR package into applications. Prior research has examined SPRs for traditional software, such as NPM (JavaScript) and PyPI (Python). Pre-Trained Model (PTM) Registries are an emerging class of SPR of increasing importance, because they support the deep learning supply chain. Aims: Recent empirical research has examined PTM registries in ways such as vulnerabilities, reuse processes, and evolution. However, no existing research synthesizes them to provide a systematic understanding of the current knowledge. Some of the existing research includes qualitative claims lacking quantitative analysis. Our research fills these gaps by providing a knowledge synthesis and quantitative analyses. Methods: We first conduct a systematic literature review (SLR). We then observe that some of the claims are qualitative. We identify quantifiable metrics associated with those claims, and measure in order to substantiate these claims. Results: From our SLR, we identify 12 claims about PTM reuse on the HuggingFace platform, 4 of which lack quantitative validation. We successfully test 3 of these claims through a quantitative analysis, and directly compare one with traditional software. Our findings corroborate qualitative claims with quantitative measurements. Our findings are: (1) PTMs have a much higher turnover rate than traditional software, indicating a dynamic and rapidly evolving reuse environment within the PTM ecosystem; and (2) There is a strong correlation between documentation quality and PTM popularity. Conclusions: We confirm qualitative research claims with concrete metrics, supporting prior qualitative and case study research. Our measures show further dynamics of PTM reuse, inspiring research infrastructure and new measures., Comment: [ESEM'24] Proceedings of the 18th ACM/IEEE International Symposium on Empirical Software Engineering and Measurement (ESEM) 2024

This paper undertakes the task of replicating the MaskFormer model a universal image segmentation model originally developed using the PyTorch framework, within the TensorFlow ecosystem, specifically optimized for execution on Tensor Processing Units (TPUs). Our implementation exploits the modular constructs available within the TensorFlow Model Garden (TFMG), encompassing elements such as the data loader, training orchestrator, and various architectural components, tailored and adapted to meet the specifications of the MaskFormer model. We address key challenges encountered during the replication, non-convergence issues, slow training, adaptation of loss functions, and the integration of TPU-specific functionalities. We verify our reproduced implementation and present qualitative results on the COCO dataset. Although our implementation meets some of the objectives for end-to-end reproducibility, we encountered challenges in replicating the PyTorch version of MaskFormer in TensorFlow. This replication process is not straightforward and requires substantial engineering efforts. Specifically, it necessitates the customization of various components within the TFMG, alongside thorough verification and hyper-parameter tuning. The replication is available at: https://github.com/PurdueDualityLab/tf-maskformer/tree/main/official/projects/maskformer

Deep neural networks (DNNs) achieve state-of-the-art performance in many areas, including computer vision, system configuration, and question-answering. However, DNNs are expensive to develop, both in intellectual effort (e.g., devising new architectures) and computational costs (e.g., training). Reusing DNNs is a promising direction to amortize costs within a company and across the computing industry. As with any new technology, however, there are many challenges in reusing DNNs. These challenges include both missing technical capabilities and missing engineering practices. This vision paper describes challenges in current approaches to DNN re-use. We summarize studies of re-use failures across the spectrum of re-use techniques, including conceptual (e.g., reusing based on a research paper), adaptation (e.g., re-using by building on an existing implementation), and deployment (e.g., direct re-use on a new device). We outline possible advances that would improve each kind of re-use., Comment: Proceedings of the IEEE John Vincent Atanasoff Symposium on Modern Computing (JVA'23) 2023

The development and training of deep learning models have become increasingly costly and complex. Consequently, software engineers are adopting pre-trained models (PTMs) for their downstream applications. The dynamics of the PTM supply chain remain largely unexplored, signaling a clear need for structured datasets that document not only the metadata but also the subsequent applications of these models. Without such data, the MSR community cannot comprehensively understand the impact of PTM adoption and reuse. This paper presents the PeaTMOSS dataset, which comprises metadata for 281,638 PTMs and detailed snapshots for all PTMs with over 50 monthly downloads (14,296 PTMs), along with 28,575 open-source software repositories from GitHub that utilize these models. Additionally, the dataset includes 44,337 mappings from 15,129 downstream GitHub repositories to the 2,530 PTMs they use. To enhance the dataset's comprehensiveness, we developed prompts for a large language model to automatically extract model metadata, including the model's training datasets, parameters, and evaluation metrics. Our analysis of this dataset provides the first summary statistics for the PTM supply chain, showing the trend of PTM development and common shortcomings of PTM package documentation. Our example application reveals inconsistencies in software licenses across PTMs and their dependent projects. PeaTMOSS lays the foundation for future research, offering rich opportunities to investigate the PTM supply chain. We outline mining opportunities on PTMs, their downstream usage, and cross-cutting questions., Comment: Accepted at MSR'24

Developing and training deep learning models is expensive, so software engineers have begun to reuse pre-trained deep learning models (PTMs) and fine-tune them for downstream tasks. Despite the wide-spread use of PTMs, we know little about the corresponding software engineering behaviors and challenges. To enable the study of software engineering with PTMs, we present the PeaTMOSS dataset: Pre-Trained Models in Open-Source Software. PeaTMOSS has three parts: a snapshot of (1) 281,638 PTMs, (2) 27,270 open-source software repositories that use PTMs, and (3) a mapping between PTMs and the projects that use them. We challenge PeaTMOSS miners to discover software engineering practices around PTMs. A demo and link to the full dataset are available at: https://github.com/PurdueDualityLab/PeaTMOSS-Demos.

As innovation in deep learning continues, many engineers seek to adopt Pre-Trained Models (PTMs) as components in computer systems. Researchers publish PTMs, which engineers adapt for quality or performance prior to deployment. PTM authors should choose appropriate names for their PTMs, which would facilitate model discovery and reuse. However, prior research has reported that model names are not always well chosen - and are sometimes erroneous. The naming for PTM packages has not been systematically studied. In this paper, we frame and conduct the first empirical investigation of PTM naming practices in the Hugging Face PTM registry. We initiated our study with a survey of 108 Hugging Face users to understand the practices in PTM naming. From our survey analysis, we highlight discrepancies from traditional software package naming, and present findings on naming practices. Our findings indicate there is a great mismatch between engineers' preferences and practical practices of PTM naming. We also present practices on detecting naming anomalies and introduce a novel automated DNN ARchitecture Assessment technique (DARA), capable of detecting PTM naming anomalies. We envision future works on leveraging meta-features of PTMs to improve model reuse and trustworthiness., Comment: 21 pages

Software engineers develop, fine-tune, and deploy deep learning (DL) models using a variety of development frameworks and runtime environments. DL model converters move models between frameworks and to runtime environments. Conversion errors compromise model quality and disrupt deployment. However, the failure characteristics of DL model converters are unknown, adding risk when using DL interoperability technologies. This paper analyzes failures in DL model converters. We survey software engineers about DL interoperability tools, use cases, and pain points (N=92). Then, we characterize failures in model converters associated with the main interoperability tool, ONNX (N=200 issues in PyTorch and TensorFlow). Finally, we formulate and test two hypotheses about structural causes for the failures we studied. We find that the node conversion stage of a model converter accounts for ~75% of the defects and 33% of reported failure are related to semantically incorrect models. The cause of semantically incorrect models is elusive, but models with behaviour inconsistencies share operator sequences. Our results motivate future research on making DL interoperability software simpler to maintain, extend, and validate. Research into behavioural tolerances and architectural coverage metrics could be fruitful., Comment: [ISSTA'24] Proceedings of the 33rd ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) 2024

Due to the cost of developing and training deep learning models from scratch, machine learning engineers have begun to reuse pre-trained models (PTMs) and fine-tune them for downstream tasks. PTM registries known as "model hubs" support engineers in distributing and reusing deep learning models. PTM packages include pre-trained weights, documentation, model architectures, datasets, and metadata. Mining the information in PTM packages will enable the discovery of engineering phenomena and tools to support software engineers. However, accessing this information is difficult - there are many PTM registries, and both the registries and the individual packages may have rate limiting for accessing the data. We present an open-source dataset, PTMTorrent, to facilitate the evaluation and understanding of PTM packages. This paper describes the creation, structure, usage, and limitations of the dataset. The dataset includes a snapshot of 5 model hubs and a total of 15,913 PTM packages. These packages are represented in a uniform data schema for cross-hub mining. We describe prior uses of this data and suggest research opportunities for mining using our dataset. The PTMTorrent dataset (v1) is available at: https://app.globus.org/file-manager?origin_id=55e17a6e-9d8f-11ed-a2a2-8383522b48d9&origin_path=%2F~%2F. Our dataset generation tools are available on GitHub: https://doi.org/10.5281/zenodo.7570357., Comment: 5 pages, 2 figures, Accepted to MSR'23

Many engineering organizations are reimplementing and extending deep neural networks from the research community. We describe this process as deep learning model reengineering. Deep learning model reengineering - reusing, reproducing, adapting, and enhancing state-of-the-art deep learning approaches - is challenging for reasons including under-documented reference models, changing requirements, and the cost of implementation and testing. In addition, individual engineers may lack expertise in software engineering, yet teams must apply knowledge of software engineering and deep learning to succeed. Prior work has examined on DL systems from a "product" view, examining defects from projects regardless of the engineers' purpose. Our study is focused on reengineering activities from a "process" view, and focuses on engineers specifically engaged in the reengineering process. Our goal is to understand the characteristics and challenges of deep learning model reengineering. We conducted a case study of this phenomenon, focusing on the context of computer vision. Our results draw from two data sources: defects reported in open-source reeengineering projects, and interviews conducted with open-source project contributors and the leaders of a reengineering team. Our results describe how deep learning-based computer vision techniques are reengineered, analyze the distribution of defects in this process, and discuss challenges and practices. Integrating our quantitative and qualitative data, we proposed a novel reengineering workflow. Our findings inform several future directions, including: measuring additional unknown aspects of model reengineering; standardizing engineering practices to facilitate reengineering; and developing tools to support model reengineering and model reuse., Comment: Under submission to EMSE

Deep Neural Networks (DNNs) are being adopted as components in software systems. Creating and specializing DNNs from scratch has grown increasingly difficult as state-of-the-art architectures grow more complex. Following the path of traditional software engineering, machine learning engineers have begun to reuse large-scale pre-trained models (PTMs) and fine-tune these models for downstream tasks. Prior works have studied reuse practices for traditional software packages to guide software engineers towards better package maintenance and dependency management. We lack a similar foundation of knowledge to guide behaviors in pre-trained model ecosystems. In this work, we present the first empirical investigation of PTM reuse. We interviewed 12 practitioners from the most popular PTM ecosystem, Hugging Face, to learn the practices and challenges of PTM reuse. From this data, we model the decision-making process for PTM reuse. Based on the identified practices, we describe useful attributes for model reuse, including provenance, reproducibility, and portability. Three challenges for PTM reuse are missing attributes, discrepancies between claimed and actual performance, and model risks. We substantiate these identified challenges with systematic measurements in the Hugging Face ecosystem. Our work informs future directions on optimizing deep learning ecosystems by automated measuring useful attributes and potential attacks, and envision future research on infrastructure and standardization for model registries., Comment: Proceedings of the ACM/IEEE 45th International Conference on Software Engineering (ICSE) 2023

Training deep neural networks (DNNs) takes signifcant time and resources. A practice for expedited deployment is to use pre-trained deep neural networks (PTNNs), often from model zoos -- collections of PTNNs; yet, the reliability of model zoos remains unexamined. In the absence of an industry standard for the implementation and performance of PTNNs, engineers cannot confidently incorporate them into production systems. As a first step, discovering potential discrepancies between PTNNs across model zoos would reveal a threat to model zoo reliability. Prior works indicated existing variances in deep learning systems in terms of accuracy. However, broader measures of reliability for PTNNs from model zoos are unexplored. This work measures notable discrepancies between accuracy, latency, and architecture of 36 PTNNs across four model zoos. Among the top 10 discrepancies, we find differences of 1.23%-2.62% in accuracy and 9%-131% in latency. We also fnd mismatches in architecture for well-known DNN architectures (e.g., ResNet and AlexNet). Our findings call for future works on empirical validation, automated tools for measurement, and best practices for implementation., Comment: Proceedings of the 30th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering: Ideas, Visions, and Reflections track (ESEC/FSE-IVR) 2022

Software metrics capture information about software development processes and products. These metrics support decision-making, e.g., in team management or dependency selection. However, existing metrics tools measure only a snapshot of a software project. Little attention has been given to enabling engineers to reason about metric trends over time -- longitudinal metrics that give insight about process, not just product. In this work, we present PRiME (PRocess MEtrics), a tool for computing and visualizing process metrics. The currently-supported metrics include productivity, issue density, issue spoilage, and bus factor. We illustrate the value of longitudinal data and conclude with a research agenda. The tool's demo video can be watched at https://youtu.be/YigEHy3_JCo. The source code can be found at https://github.com/SoftwareSystemsLaboratory/prime., Comment: Accepted at ASE 2022 Tool Demonstrations

We comment on a recent TKDE paper "Linear Approximation of F-measure for the Performance Evaluation of Classification Algorithms on Imbalanced Data Sets", and make two improvements related to comparison of F-measures for two prediction rules.

Machine learning techniques are becoming a fundamental tool for scientific and engineering progress. These techniques are applied in contexts as diverse as astronomy and spam filtering. However, correctly applying these techniques requires careful engineering. Much attention has been paid to the technical potential; relatively little attention has been paid to the software engineering process required to bring research-based machine learning techniques into practical utility. Technology companies have supported the engineering community through machine learning frameworks such as TensorFLow and PyTorch, but the details of how to engineer complex machine learning models in these frameworks have remained hidden. To promote best practices within the engineering community, academic institutions and Google have partnered to launch a Special Interest Group on Machine Learning Models (SIGMODELS) whose goal is to develop exemplary implementations of prominent machine learning models in community locations such as the TensorFlow Model Garden (TFMG). The purpose of this report is to define a process for reproducing a state-of-the-art machine learning model at a level of quality suitable for inclusion in the TFMG. We define the engineering process and elaborate on each step, from paper analysis to model release. We report on our experiences implementing the YOLO model family with a team of 26 student researchers, share the tools we developed, and describe the lessons we learned along the way., Comment: Technical Report

We provide analytic formulas for the standard error and confidence intervals for the F measures, based on a property of asymptotic normality in the large sample limit. The formula can be applied for sample size planning in order to achieve accurate enough estimation of these F measures.

In this article, we propose a novel variable screening method for linear models named as conditional screening via ordinary least squares projection (COLP). COLP can take advantage of prior knowledge concerning certain active predictors by eliminating the adverse impact of their coefficients in the estimation of remaining ones and thus significantly enhance the screening accuracy. We prove its sure-screening property under reasonable assumptions and demonstrate its utility in an application to a leukemia dataset. Moreover, based on the conditional approach, we introduce an iterative algorithm named as forward screening via ordinary least squares projection (FOLP), which not only could exploit the prior information more effectively, but also has promising performance when no prior knowledge is available using a data-driven conditioning set. Extensive simulation studies are carried out to demonstrate the competence of both proposed methods.

Word embeddings capture syntactic and semantic information about words. Definition modeling aims to make the semantic content in each embedding explicit, by outputting a natural language definition based on the embedding. However, existing definition models are limited in their ability to generate accurate definitions for different senses of the same word. In this paper, we introduce a new method that enables definition modeling for multiple senses. We show how a Gumble-Softmax approach outperforms baselines at matching sense-specific embeddings to definitions during training. In experiments, our multi-sense definition model improves recall over a state-of-the-art single-sense definition model by a factor of three, without harming precision.

Fan and Lv (2008) proposed the path-breaking theory of sure independence screening (SIS) and an iterative algorithm (ISIS) to effectively reduce the predictor dimension for further variable selection approaches. Fan et al. (2009) extended ISIS to generalized linear models and introduced the Vanilla ISIS (Van-ISIS) algorithm, allowing selected predictors to be screened out in upcoming iterations. The success of SIS depends on its sure screening property, which was obtained by Fan and Lv (2008) under the marginal correlation assumption. However, despite wide applications of ISIS and Van-ISIS in various scientific fields, their sure screening properties have not been proved during the past decade. To fill this gap, we prove the sure screening properties of three different types of iterative algorithms for linear models without relying on the marginal correlation assumption, where ISIS and Van-ISIS can be regarded as two special cases of them.

Ecological inference (EI) is the process of learning about individual behavior from aggregate data. We study a partially identified linear contextual effects model for EI and describe how to estimate the district level parameter averaging over many precincts in the presence of the non-identified parameter of the contextual effect. This may be regarded as a first attempt in this venerable literature to limit the scope of the key form of non-identifiability in EI. To study the operating characteristics of our model, we have amassed the largest collection of data with known ground truth ever applied to evaluate solutions to the EI problem. We collect and study 459 datasets from a variety of fields including public health, political science, and sociology. The datasets contain a total of 2,370,854 geographic units (e.g., precincts), with an average of 5,165 geographic units per dataset. Our replication data are publicly available via the Harvard Dataverse (Jiang et al. 2018) and may serve as a useful resource for future researchers. For all real data sets in our collection that fit our proposed rules, our approach reduces the width of the Duncan and Davis (1953) deterministic bound, on average, by about 45\%, while still capturing the true district level parameter in excess of 97\% of the time. .

We establish the limiting distribution (in total variation) of the quasi posteriors based on moment conditions, which only partially identify the parameters of interest. Some examples are discussed.

Community detection has been an active research area for decades. Among all probabilistic models, Stochastic Block Model has been the most popular one. This paper introduces a novel probabilistic model: RW-HDP, based on random walks and Hierarchical Dirichlet Process, for community extraction. In RW-HDP, random walks conducted in a social network are treated as documents; nodes are treated as words. By using Hierarchical Dirichlet Process, a nonparametric Bayesian model, we are not only able to cluster nodes into different communities, but also determine the number of communities automatically. We use Stochastic Variational Inference for our model inference, which makes our method time efficient and can be easily extended to an online learning algorithm.