@article{li_MLbindings,

title = {Bridging the language gap: an empirical study of bindings for open source machine learning libraries across software package ecosystems},

author = {Hao Li and Cor-Paul Bezemer},

year  = {2024},

date = {2024-10-18},

urldate = {2024-10-18},

journal = {Empirical Software Engineering},

volume = {30},

number = {6},

abstract = {Open source machine learning (ML) libraries enable developers to

integrate advanced ML functionality into their own applications. However,

popular ML libraries, such as TensorFlow, are not available natively in all

programming languages and software package ecosystems. Hence, developers

who wish to use an ML library which is not available in their programming lan-

guage or ecosystem of choice, may need to resort to using a so-called binding

library (or binding). Bindings provide support across programming languages

and package ecosystems for reusing a host library. For example, the Keras

.NET binding provides support for the Keras library in the NuGet (.NET)

ecosystem even though the Keras library was written in Python. In this pa-

per, we collect 2,436 cross-ecosystem bindings for 546 ML libraries across 13

software package ecosystems by using an approach called BindFind, which can

automatically identify bindings and link them to their host libraries. Further-

more, we conduct an in-depth study of 133 cross-ecosystem bindings and their

development for 40 popular open source ML libraries. Our findings reveal that

the majority of ML library bindings are maintained by the community, with

npm being the most popular ecosystem for these bindings. Our study also

indicates that most bindings cover only a limited range of the host library’s

releases, often experience considerable delays in supporting new releases, and

have widespread technical lag. Our findings highlight key factors to consider

for developers integrating bindings for ML libraries and open avenues for re-

searchers to further investigate bindings in software package ecosystems.},

keywords = {Library bindings, Machine learning, SE4AI, SE4ML},

pubstate = {published},

tppubtype = {article}

}

@article{Li_BindingsQuality,

title = {Studying the Impact of TensorFlow and PyTorch Bindings on Machine Learning Software Quality},

author = {Hao Li and Gopi Krishnan Rajbahadur and Cor-Paul Bezemer},

year  = {2024},

date = {2024-07-07},

journal = {ACM Transactions on Software Engineering and Methodology},

abstract = {Bindings for machine learning frameworks (such as TensorFlow and PyTorch) allow developers to integrate a framework’s functionality using a programming language different from the framework’s default language (usually Python). In this paper, we study the impact of using TensorFlow and PyTorch bindings in C#, Rust, Python and JavaScript on the software quality in terms of correctness (training and test accuracy) and time cost (training and inference time) when training and performing inference on five widely used deep learning models. Our experiments show that a model can be trained in one binding and used for inference in another binding for the same framework without losing accuracy. Our study is the first to show that using a non-default binding can help improve machine learning software quality from the time cost perspective compared to the default Python binding while still achieving the same level of correctness.},

keywords = {Library bindings, Machine learning, SE4AI, SE4ML, Software quality},

pubstate = {published},

tppubtype = {article}

}

@article{Li_Overfitting,

title = {Keeping Deep Learning Models in Check: A History-Based Approach to Mitigate Overfitting},

author = {Hao Li and Gopi Krishnan Rajbahadur and Dayi Lin and Cor-Paul Bezemer and Zhen Ming (Jack) Jiang},

year  = {2024},

date = {2024-05-17},

journal = {IEEE Access},

volume = {12},

pages = {70676--70689},

abstract = {In software engineering, deep learning models are increasingly deployed for critical tasks such as bug detection and code review. However, overfitting remains a challenge that affects the quality, reliability, and trustworthiness of software systems that utilize deep learning models. Overfitting can be (1) prevented (e.g., using dropout or early stopping) or (2) detected in a trained model (e.g., using correlation-based approaches). Both overfitting detection and prevention approaches that are currently used have constraints (e.g., requiring modification of the model structure, and high computing resources). In this paper, we propose a simple, yet powerful approach that can both detect and prevent overfitting based on the training history (i.e., validation losses). Our approach first trains a time series classifier on training histories of overfit models. This classifier is then used to detect if a trained model is overfit. In addition, our trained classifier can be used to prevent overfitting by identifying the optimal point to stop a model’s training. We evaluate our approach on its ability to identify and prevent overfitting in real-world samples. We compare our approach against correlation-based detection approaches and the most commonly used prevention approach (i.e., early stopping). Our approach achieves an F1 score of 0.91 which is at least 5% higher than the current best-performing non-intrusive overfitting detection approach. Furthermore, our approach can stop training to avoid overfitting at least 32% of the times earlier than early stopping and has the same or a better rate of returning the best model.},

keywords = {Machine learning, Overfitting},

pubstate = {published},

tppubtype = {article}

}

@article{Sabuhi_MicroFL,

title = {Micro-FL: A Fault-Tolerant Scalable Microservice Based Platform for Federated Learning},

author = {Mikael Sabuhi and Petr Musilek and Cor-Paul Bezemer },

year  = {2024},

date = {2024-02-19},

journal = {Future Internet},

volume = {16},

number = {3},

pages = {1-19},

abstract = {As the number of machine learning applications increases, growing concerns about data privacy expose the limitations of traditional cloud-based machine learning methods that rely on centralized data collection and processing. Federated learning emerges as a promising alternative, offering a novel approach to training machine learning models that safeguards data privacy. Federated learning facilitates collaborative model training across various entities. In this approach, each user trains models locally and shares only the local model parameters with a central server, which then generates a global model based on these individual updates. This approach ensures data privacy since the training data itself is never directly shared with a central entity. However, existing federated machine learning frameworks are not without challenges. In terms of server design, these frameworks exhibit limited scalability with an increasing number of clients and are highly vulnerable to system faults, particularly as the central server becomes a single point of failure. This paper introduces Micro-FL, a federated learning framework that uses a microservices architecture to implement the federated learning system. It demonstrates that the framework is fault-tolerant and scalable, showing its ability to handle an increasing number of clients. A comprehensive performance evaluation confirms that Micro-FL proficiently handles component faults, enabling a smooth and uninterrupted operation.},

keywords = {Federated learning, Machine learning, Microservices},

pubstate = {published},

tppubtype = {article}

}