Workplace AI tools can do tasks by themselves. Getting them to stop is the problem.
Category: robotics/AI – Page 260
The Hubbard model is a studied model in condensed matter theory and a formidable quantum problem. A team of physicists used deep learning to condense this problem, which previously required 100,000 equations, into just four equations without sacrificing accuracy. The study, titled “Deep Learning the Functional Renormalization Group,” was published on September 21 in Physical Review Letters.
Dominique Di Sante is the lead author of this study. Since 2021, he holds the position of Assistant Professor (tenure track) at the Department of Physics and Astronomy, University of Bologna. At the same time, he is a Visiting Professor at the Center for Computational Quantum Physics (CCQ) at the Flatiron Institute, New York, as part of a Marie Sklodowska-Curie Actions (MSCA) grant that encourages, among other things, the mobility of researchers.
He and colleagues at the Flatiron Institute and other international researchers conducted the study, which has the potential to revolutionize the way scientists study systems containing many interacting electrons. In addition, if they can adapt the method to other problems, the approach could help design materials with desirable properties, such as superconductivity, or contribute to clean energy production.
Artificially intelligent Maxwell’s demon for optimal control of open…
A team of researchers used reinforcement learning (RL) to optimize feedback control strategies in quantum systems.
Prof. Tim Rocktäschel, AI researcher at UCL and Google DeepMind, talks about open-ended AI systems. These systems aim to keep learning and improving on their own, like evolution does in nature.
TOC:
00:00:00 Introduction to Open-Ended AI and Key Concepts.
00:01:37 Tim Rocktäschel’s Background and Research Focus.
00:06:25 Defining Open-Endedness in AI Systems.
00:10:39 Subjective Nature of Interestingness and Learnability.
00:16:22 Open-Endedness in Practice: Examples and Limitations.
00:17:50 Assessing Novelty in Open-ended AI Systems.
00:20:05 Adversarial Attacks and AI Robustness.
00:24:05 Rainbow Teaming and LLM Safety.
00:25:48 Open-ended Research Approaches in AI
00:29:05 Balancing Long-term Vision and Exploration in AI Research.
00:37:25 LLMs in Program Synthesis and Open-Ended Learning.
00:37:55 Transition from Human-Based to Novel AI Strategies.
00:39:00 Expanding Context Windows and Prompt Evolution.
00:40:17 AI Intelligibility and Human-AI Interfaces.
00:46:04 Self-Improvement and Evolution in AI Systems.
Show notes (New!) https://www.dropbox.com/scl/fi/5avpsy…
REFS:
00:01:47 — UCL DARK Lab (Rocktäschel) — AI research lab focusing on RL and open-ended learning — https://ucldark.com/
00:02:31 — GENIE (Bruce) — Generative interactive environment from unlabelled videos — https://arxiv.org/abs/2402.
00:02:42 — Promptbreeder (Fernando) — Self-referential LLM prompt evolution — https://arxiv.org/abs/2309.
The startup from MIT’s CSAIL says its Liquid Foundation Models have smaller memory needs thanks to a post-transformer architecture.
Artificial intelligence applications are experiencing a boom and expected to be mainstream technologies in the near future. However, these applications run on classic computing hardware and are extremely power-hungry.
With the help of AI, 15 passages have been deciphered from a 2,000-year-old unrolled Herculaneum scroll, providing a glimpse into the thoughts of an ancient philosopher.
As technology advances, Big Data will play an increasingly important role in protecting Earth from asteroids. By harnessing the power of data analytics, AI, and machine learning, scientists can monitor and predict asteroid movements with greater accuracy than ever before. This enables us to develop early warning systems and potentially deflect asteroids before they can cause harm. Aspiring data scientists interested in contributing to such significant fields can gain the necessary skills by enrolling in a data science course in Chennai, where they can learn to utilize these advanced tools and techniques.
What methods can be developed to help identify symptoms of myopia and its more serious version, myopic maculopathy? This is what a recent study published in JAMA Ophthalmology hopes to address as an international team of researchers investigated how artificial intelligence (AI) algorithms can be used to identify early signs of myopic maculopathy, as left untreated it can lead to irreversible damage to a person’s eyes. This study holds the potential to help researchers develop more effective options for identifying this worldwide disease, as it is estimated that approximately 50 percent of the global population will suffer from myopia by 2050.
“AI is ushering in a revolution that leverages global knowledge to improves diagnosis accuracy, especially in its earliest stage of the disease,” said Dr. Yalin Wang, who is a professor in the School of Computing and Augmented Intelligence at Arizona State University and a co-author on the study. “These advancements will reduce medical costs and improve the quality of life for entire societies.”
For the study, the researchers used a novel AI algorithm known as NN-MobileNet to scan retinal images and classify the severity of myopic maculopathy, which currently has five levels of severity in the medical field. The team then used deep neural networks to determine what’s known as the spherical equivalent, which is how eye doctors prescribe glasses and contacts to their patients. Combining these two methods enabled researchers to create a new AI algorithm capable of identifying early signs of myopic maculopathy.
Researchers have developed a new architecture for optical computing called diffraction casting, offering power-efficient processing by using light waves.
This method promises better integration and flexibility for high-performance computing tasks and could be used in fields like AI and machine learning.
Optical Computing