Quality-Diversity (QD) approaches are a promising direction to develop open-ended processes as they can discover archives of high-quality solutions across diverse niches. While already successful in many applications, QD approaches usually rely on combining only one or two solutions to generate new candidate solutions. As observed in open-ended processes such as technological evolution, wisely combining large diversity of these solutions could lead to more innovative solutions and potentially boost the productivity of QD search. In this work, we propose to exploit the pattern-matching capabilities of generative models to enable such efficient solution combinations. We introduce In-context QD, a framework of techniques that aim to elicit the in-context capabilities of pre-trained Large Language Models (LLMs) to generate interesting solutions using few-shot and many-shot prompting with quality-diverse examples from the QD archive as context. Applied to a series of common QD domains, In-context QD displays promising results compared to both QD baselines and similar strategies developed for single-objective optimization. Additionally, this result holds across multiple values of parameter sizes and archive population sizes, as well as across domains with distinct characteristics from BBO functions to policy search. Finally, we perform an extensive ablation that highlights the key prompt design considerations that encourage the generation of promising solutions for QD.

It's part of a field of work that is building machines that can provide real-time help using only limited data as input. Standard machine-learning algorithms often need to process thousands of possibilities before deciding on a solution, which may be impractical in pressurised scenarios where fast adaptation is critical. After Japan's Fukushima nuclear disaster in 2011, for example, robots were sent into the power plant to clear up radioactive debris in conditions far too dangerous for humans. The problem, says robotics researcher Professor Jean-Baptiste Mouret is that the robots kept breaking down or came across hazards that stopped them in their tracks. As part of the ResiBots initiative, he is designing a lower-cost robot that can last long periods without needing constant human maintenance for breakages and are better at overcoming unexpected obstacles.

Last month, the ScanPyramids project, led by a team of researchers from the University of Cairo's Faculty of Engineering in Egypt and the HIP Institute in France, announced that they'd used muon imaging to discover a large void hidden deep inside the Khufu's Pyramid (also known as the Great Pyramid of Giza, since it's the big one). Nobody knows what's inside, or if there's anything inside at all, or even if maybe that's where the Stargate is stashed. Obviously there's a lot of interest in what may or may not be hiding out in here, and it could help solve mysteries like how and why exactly the pyramids were built. The problem is that (understandably) we're not going to just start blowing holes in the Great Pyramid to see what's going on. In 2002, Egyptologists used a custom exploration robot (made by iRobot, in fact) to explore a small shaft leading out of the Queen's Chamber in the Great Pyramid that was sealed by a door. Rather than try to open the door, likely destroying it in the process, the robot drilled a tiny hole just large enough to poke a camera through in an effort to do the minimum amount of irreversible damage to the only wonder of the ancient world that we've got left.

Researchers have developed robots that learn to live with damaged parts in less than a minute, instead of the many hours needed by traditional self-learning systems. The system paves the way for robots to be used in a wide variety of settings, coping with damage that occurs in the real world. Possible applications include robots looking after the elderly, rescuing earthquake victims or doing housework. We marvel at the robots we see in films: some try to wipe out humanity, such as Ultron in the recent Avengers film, while others like C-3PO in Star Wars are helpful albeit slightly annoying. But they are the still the stuff of science fiction, partly because if the slightest thing goes wrong with a real-life robot it usually stops working altogether.