Melissa Humphries’ Post

"Excited to share that our paper, " Improving Pneumonia Detection in X-ray Images with Hybrid Deep Learning Techniques", has been initially accepted at the 6th International Conference on Sustainable Technologies for Industry 5.0 (STI 2024)! This prestigious platform motivates me further in my research journey. Special thanks to my senior, Amir Sohel sir, for his invaluable support and guidance as a major co-author. Keep me in your thoughts and prayers!"

Are you a university researcher thinking of using AI and LLMs in your work? Check out the resources we’ve put together at the Accelerate Programme for Scientific Discovery to help you get started 👇🏻

Preprint of a new article: Machine Learning for Leprosy Suspicion Questionnaire: A Low-Cost Tool for High Quality Leprosy New Case Screening Under review by Scientific reports - Nature https://lnkd.in/e6m89qxH

New blog now live! Catherine Breslin and Ryan Daniels talk about work to develop resources for researchers interested in using LLMs in research & emerging best practice. From getting started to fine tuning models, Accelerate Science are developing resources to help researchers use LLMs at all stages of the research journey. Read the blog: https://bit.ly/3XYclJu & check out our resources on GitHub: https://bit.ly/4dTaPOL This blog is the first in a series examining LLMs in research, keep an eye out for upcoming posts featuring researchers across the University using LLMs in their research area!

Why Science and Technology Matter |A Scientist’s Perspective As a scientist deeply involved in the world of science and technology, it might seem contradictory to assert that I don’t like these fields. However, this statement comes with a twist — it’s not that I dislike science and technology per se, but rather that I recognize the challenges and complexities they present. My journey has been one of both fascination and frustration, and through this article, I aim to share my personal insights to encourage a deeper appreciation and understanding of science and technology among the community. https://lnkd.in/gSNKSWpX

Science Myths: Busted! 👻 Since it's the spookiest month of the year 🎃 It's the perfect time to shine a light on common myths about careers in science that might be keeping future scientists in the dark! 👀 ✨ ❌ Myth: "All scientists wear white lab coats." 🥼 ✅ Fact: Not every scientist works in a lab! In fact, the classic white lab coat is only one tiny part of the picture! Scientists work everywhere - from research labs and field sites to classrooms, corporate offices, and even deep in our oceans, deserts and rainforests. Want to know more about what it's really like to work in science? Check out our website ➡️ https://lnkd.in/eTj-f3V4

AGU Session SY053 - So Much Talking. When Do Scientists Listen? Now inviting abstracts that focus on how *listening* changed what you study, helped you find new applications of your work or changed you and how you perceive your science. Submit at: https://lnkd.in/gkfsw654 Scientists seeking impact have been urged to embrace "dialog." But dialog is more than two parties talking at each other. It requires active listening and an openness to changing one's own behavior. It also requires dedicated time and effort. A 2023 survey of researchers in the Long Term Ecological Research Network (Besley and Downs 2024) found that scientists prioritize listening to the needs and desires of community members and resource managers when they perceive benefits to themselves, their science, or their site. In this session, we aim to attract and share examples of the benefits of listening. Has listening changed what you study or how you study it? Has listening helped you find new applications of your work? Has it changed you and how you perceive your science? Come share your stories of the impact of listening on science and scientists.

Academic research must be more robust, reproducible, and transparent to improve the public's trust in science. This was a great series from Freakonomics highlighting some of the perverse incentives that can push individuals to commit academic fraud. It also speaks to progress many individuals and institutions are making to improve the quality of scholarship including work from the Data Colada team, Retraction Watch, and others. https://lnkd.in/esPGJfae

Creativity Here's the third column of my 4 part series on Creativity in the Jersey Evening Post - my usual evidence-based cognitve science analysis and evaluation - focusing on practices and tecniques.

The dualization problem for positive (or monotone) boolean functions has polynomial time complexity. Preprint of the full paper submitted to Quantum Machine Intelligence https://lnkd.in/daWxXVht