bioRxiv Science⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/2021.12.06.470764

EpyNN: Educational python for Neural Networks

Malard, F.·Danner, L.·Rouzies, E.·Meyer, J. G.·Lescop, E.·Olivier-Van Stichelen, S.
View original source ↗Export metadata ↓

Abstract

SummaryArtificial Neural Networks (ANNs) have achieved unequaled performance for numerous problems in many areas of Science, Business, Public Policy, and more. While experts are familiar with performance-oriented software and underlying theory, ANNs are difficult to comprehend for non-experts because it requires skills in programming, background in mathematics and knowledge of terminology and concepts. In this work, we release EpyNN, an educational python resource meant for a public willing to understand key concepts and practical implementation of scalable ANN architectures from concise, homogeneous and idiomatic source code. EpyNN contains an educational Application Programming Interface (API), educational workflows from data preparation to ANN training and a documentation website setting side-by-side code, mathematics, graphical representation and text to facilitate learning and provide teaching material. Overall, EpyNN provides basics for python-fluent individuals who wish to learn, teach or develop from scratch. AvailabilityEpyNN documentation is available at https://epynn.net and repository can be retrieved from https://github.com/synthaze/epynn. ContactStephanie Olivier-Van-Stichelen, solivier@mcw.edu. Supplementary InformationSupplementary files and listings.

Source connections

Explore related subjects

scientific communication and educationPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Malard, F., Danner, L., Rouzies, E., Meyer, J. G., Lescop, E., Olivier-Van Stichelen, S.. 2021-12-07. EpyNN: Educational python for Neural Networks. https://doi.org/10.1101/2021.12.06.470764

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Evaluating Large Language Models as Tools to Navigate Researchers in Rapidly Evolving Research Landscapes: A Case Study in Cancer Drug Response Prediction

Large Language Models (LLMs) have emerged as promising tools for assisting researchers in automating and accelerating the synthesis of literature reviews. However, their reliability is a significant concern due to issues like factual inaccuracies and hallucinations. The key question is whether LLMs can reliably provide comprehensive, up-to-date overviews and analyses. This study evaluates the performance of three leading LLMs (OpenAI's ChatGPT, Google's Gemini, and DeepSeek) on the complex task of generating a comprehensive survey paper on deep learning for cancer Drug Response Prediction (DRP). By testing both standard and Deep Research (DR) / Deep Think (DT) modes of LLMs with prompts of varying detail, this paper assesses key academic dimensions, including reference management, content quality, and analytical depth. Key findings reveal that while DR modes of LLMs significantly improve reliability by eliminating hallucinations, performance variations exist across models and prompts. A trade-off between reference quantity and integration quality was observed, and even the best-performing models lacked the analytical depth of human experts, often requiring extensive human supervision. The study concludes that LLMs currently serve as powerful assistive tools but still cannot replace the critical validation and synthesis provided by human researchers. Choosing the best LLM to use depends on the task in hand, while several strategies can be implemented to improve the produced output.

scientific communication and education↗
bioRxiv2022

Preparing future STEM faculty nationwide through flexible teaching professional development

We report on a five-year initiative that has prepared thousands of future STEM faculty around the world to adopt evidence-based instructional practices by participating in two massive open online courses (MOOCs) and facilitated in-person learning communities. This novel combination of asynchronous online and coordinated, structured face-to-face learning community experiences provides flexible options for STEM graduate students and postdoctoral fellows to pursue teaching professional development, while leveraging the affordances of educational technologies and the geographically clustered nature of this target learner demographic. A total of 14,977 participants enrolled in seven offerings of the introductory course held 2014-2018, with 1,725 participants from approximately 60 countries completing at an average course completion rate of 13%. The preparation of future STEM faculty makes an important difference in establishing high-quality instruction that meets the diverse needs of all undergraduate students, and the initiative described here can serve as a model for increasing access to such preparation.

scientific communication and education↗
bioRxiv2022

Sexual health knowledge acquisition processes among very young adolescent girls in rural Malawi: Implications for sexual and reproductive health programs

Early adolescence is a period characterized by enormous biological, cognitive, sexual, emotional, and social changes. Sexual curiosity and the desire to acquire sexual health (SH) information are part of these developments. Understanding the SH knowledge acquisition process is critical for designing interventions that can best support very young adolescents (VYAs). This study explored the SH knowledge acquisition processes among VYA girls aged 10 to 14 years who attended the DREAMs Girl Only Clubs (GOCs) and those who did not. The GOCs were a part of a larger comprehensive HIV prevention project called DREAMS (Determined, Resilient, Empowered, AIDS-free, Mentored, and Safe) which provided an evidence-based core package of interventions to VYAs to prevent HIV. In-depth interviews were conducted with 43 VYA girls aged 10-14 years in two rural southern districts of Zomba and Machinga in Malawi. Twenty-three VYA girls were GOC participants and 20 VYAs were not. Guided by the Social Ecological Model, a thematic analysis approach was used to analyze the data with the assistance of Nvivo 12 software. The SH knowledge acquisition processes were the interaction of various factors at the microsystem (biological factors, attitudes), mesosystem (parents, older relatives, friends), and exosystem levels (schools and GOCs). Factors influencing this interaction included the beginning of menstruation, supportive parents/adults on SH issues, trust and confidentiality in SH information sources, and preference for well-informed sources of information. Compared to Non-GOC participants, GOC participants demonstrated an in-depth knowledge of SH issues, and positive sexual behaviors such as HIV testing. Limitations to SH knowledge acquisition were adult messages focus on sexual relationship avoidance and on girls who have started menstruation; perception of adults not being knowledgeable about SH and school teachers hiding some SH information. VYA SH interventions that provide VYA with accurate SH information such as the GOCs have the potential to address the limitations that other sources have in reaching VYAs. Integrating such interventions with empowering parents, other adults, and teachers with comprehensive SH information and with skills on how to deliver SH information to VYAs can enhance their SH knowledge acquisition and influence positive behavior change.

scientific communication and education↗