Thinking Allowed

medical / technology / education / art / flub

showing posts for 'models'

Nightshade, the free tool that ‘poisons’ AI models, is now available for artists to use

blog post image "The tool's creators are seeking to make it so that AI model developers must pay artists to train on data from them that is uncorrupted." Artists can now use this software to modify their art and influence the big tech owners of AI tools to properly recognise the original works. Growing the poisonous...
Source: venturebeat.com

Predicting students’ academic progress and related attributes in first-year medical students: an analysis with artificial

"Background Dropout and poor academic performance are persistent problems in medical schools in emerging economies. Identifying at-risk students early and knowing the factors that contribute to their success would be useful for designing educational interventions. Educational Data Mining (EDM) methods...
Source: biomedcentral.com

What (else) happened? A key question for learning programmes.

blog post image Good paper from 2013 on the need to go beyond just asking 'did our programme work?' "It is clear that programme evaluations using traditional ‘outcomes-based’ models are inadequate for the health professions context. Consequently, the scholarship in health professions education has begun to incorporate...
Source: wiley.com

Discovery of a structural class of antibiotics with explainable deep learning.

One of the challenges with deep learning (neural networks) is that although they find patterns the reasoning disappears into an endless detail of numbers. In this paper the researchers built an 'explainable' AI to discover antibiotics instead of such a 'black box'. "The discovery of novel structural...
Source: nature.com

This AI Trained on the Life Events of Every Person in Denmark. It Can Now Predict Their Future.

Artificial intelligence is very good at finding patterns. Given enough data it will start making telling you what's going to happen. "The AI can make highly accurate predictions about people’s lives, including how likely they are to die in a given time window and their personality traits." ... ...
Source: singularityhub.com

(We are not) using eHealth Data to Inform CPD for Medical Practitioners

"There is no formal or well-established correlation between individual performance data obtained through eHealth data analysis and CPD planning and programming for medical practitioners; in particular, the literature shows no consistency in type of eHealth data to analyze, software and tools to use,...
Source: nih.gov

Coding excursions

blog post image Anonymity with encryption At Outcomes Engine we are working on techniques to gather data from learners, analyse the data, and share the data whilst maintaining anonymity. I was involved in some work in my previous company (pharmaceutical) with the security of personal data - in our case it was data...

The Biggest Discoveries in Computer Science in 2023

"Artificial intelligence learned how to generate text and art better than ever before, while computer scientists developed algorithms that solved long-standing problems." Links to further papers and discussion on topics including: Tackling "P versus NP" Emergent behaviours in large language models...
Source: quantamagazine.org

Large Language Models: Don't Believe the Hype

We've reached the Skynet moment for AI, it would seem. GPT-4, cast in the role of Terminator, has not only achieved consciousness, but it's coming for
Source: datanami.com

ChatGPT performs well in the USMLE (nearly passes with no training)

This pre-print paper suggests ChatGPT could change how assessments might be done. “We evaluated the performance of a large language model called ChatGPT on the United States Medical Licensing Exam (USMLE), which consists of three exams: Step 1, Step 2CK, and Step 3. ChatGPT performed at or near the...

Biopharma futures

blog post image I've been looking at a few industry reports on the future of biopharma as part of a course I'm on with INSEAD Business School on Business Strategy and Financial Performance. I thought I'd share some of my ramblings on Biopharma Futures. Expectations are high Pharmaceutical companies are operating...
Source: deanjenkins.me

The future of biopharma

Shifts in how diseases are diagnosed and prevented, custom treatments, curative therapies, digital therapeutics, and precision intervention could upend current business models in biopharma. Deloitte's Five forces to 2040 Prevention and early detection Customized treatments - interventions for very...
Source: deloitte.com

Volvo Cars to go fully electric by 2030

"The Swedish firm is to phase out all models with internal combustion engines over the next few years."
Source: bbc.com

Increasing Students Critical Thinking Skills and Learning Motivation Using Inquiry Mind Map

These researchers from Indonesia explore the use of inquiry-based models in encouraging critical thinking in medical students. "Critical thinking skills are very important to have for students given the rapid distribution of information. To promote the critical thinking skills of the student, it could...
Source: online-journals.org

Phrase of the day: Scaling Kubernetes to 7,500 Nodes

Scaling Kubernetes to 7,500 Nodes: We've scaled Kubernetes clusters to 7,500 nodes, producing a scalable infrastructure for large models like GPT-3, CLIP, and DALL·E, but also for rapid small-scale iterative research such as Scaling Laws for Neural Language Models. Scaling a single Kubernetes cluster...
Source: openai.com

MMR, smoking, and 'in-vitro' models in COVID-19

Good summary of current science news topics for coronavirus in Reuters. Strong responders to MMR II vaccine (since 1979) have less severe disease with COVID-19, cigarette smoke increases risk, and a useful 'in-vivo' model for further study of how SARS-CoV-2 and other viruses might be treated.
Source: reuters.com

Why Use Ensemble Learning? - Machine Learning Mastery: What are the Benefits of Ensemble Methods for Machine Learning? Ensembles

Why Use Ensemble Learning? - Machine Learning Mastery: What are the Benefits of Ensemble Methods for Machine Learning? Ensembles are predictive models that combine predictions from two or more other models. Ensemble learning methods are popular and the go-to technique when the best performance on...
Source: machinelearningmastery.com

Augmented reality could be the geology classroom

Augmented reality could be the geology classroom’s killer app "Geology is a very spatial science and can require a lot of 3-D visualization. Simple physical models (not to mention rocks) have long been used to aid teaching about things like faults or crystalline mineral structure. But these things...
Source: arstechnica.com

Learning to Summarize with Human Feedback: We've applied reinforcement learning from human feedback to train language models

Learning to Summarize with Human Feedback: We've applied reinforcement learning from human feedback to train language models that are better at summarization. Our models generate summaries that are better than summaries from 10x larger models trained only with supervised learning. Even though we train...
Source: openai.com

Sustainable Open Access – What’s Next?  - The Scholarly Kitchen: How can collective action models to support open access,

Sustainable Open Access – What’s Next?  - The Scholarly Kitchen: How can collective action models to support open access, like Subscribe to Open, be applied to academic publishing? An interview with Raym Crow.
Source: sspnet.org