LucianoSphere (Luciano Abriata, PhD)

AI

Predictive Power Score: Calculation, Pros, Cons, and JavaScript Code

A project to learn about correlation in general, and to test neural networks in web browsers with Brain.js and Tensorflow.js LucianoSphere (Luciano Abriata, PhD) · Follow Published in Towards Data Science · 7 min read · 3 hours ago — There’s clearly some relationship between the values plotted in X and Y, but regular correlation coefficients like Pearson’s would return a score close to 0. However, the Predicted Power Score coupled to a proper model

Read More »
AI

Why the 2024 Nobel Prize in (AI for) Chemistry Matters So Much

To Demis Hassabis and John Jumper, from DeepMind, and to David Baker, leader of the Institute for Protein Design LucianoSphere (Luciano Abriata, PhD) · Follow Published in Towards Data Science · 5 min read · 15 hours ago — Figure from the Royal Swedish Academy of Sciences (in the public domain) In summary: because the research and development behind this Nobel Prize unlocked not only new biology but also new computer science, all directly leading

Read More »
AI

How AI Could Soon Take Human-Computer Interaction to New Levels

As AI models reach excellence in speech recognition and synthesis, text processing, and multimodalism, the ultimate voice-user interfaces could soon be ubiquitous LucianoSphere (Luciano Abriata, PhD) · Follow Published in Towards Data Science · 18 min read · 1 hour ago — Voice User Interface (VUI) for natural speech-based human-computer interaction as imagined by Dall-E 3 via ChatGPT. It was a typical Friday afternoon right at the end of a long week of work on

Read More »
AI

New Approach for Training Physical (as Opposed to Computer-Based) Artificial Neural Networks

Neural networks built from light waves could allow for much more versatile, scalable, and energy-efficient AI systems LucianoSphere (Luciano Abriata, PhD) · Follow Published in Towards Data Science · 6 min read · 15 hours ago — Picture generated by Dall-E 3 via ChatGPT. Traditional AI systems relying on deep artificial neural networks that run inside computers require vast amounts of computational resources for their training, leading to concerns about their sustainability. One promising avenue

Read More »
AI

A New Method to Detect “Confabulations” Hallucinated by Large Language Models

By calculating semantic entropy with a second LLM, we can better flag answers as unreliable due to lack of knowledge LucianoSphere (Luciano Abriata, PhD) · Follow Published in Towards Data Science · 10 min read · 9 hours ago — Photo by Jametlene Reskp on Unsplash As you surely know, AI has made huge strides in the last two years with the development and mass-scale deployment of large language models…

Read More »
AI

“Sparks of Chemical Intuition”—and Gross Limitations!—in AlphaFold 3

Observations after 3 weeks of DeepMind releasing its hitherto most advanced model for biomolecular structure prediction. LucianoSphere (Luciano Abriata, PhD) · Follow Published in Towards Data Science · 16 min read · 2 hours ago — Figure composed by the author. Index – Introduction– What AlphaFold 3 is and how it Stands out From Previous Versions Beyond proteins Community…

Read More »