AM-IA brings together news, research papers and applications relating to artificial intelligence published around the world into a single feed — and makes them available in your European language at the click of a button.
Scroll to browse · Click to filter
A system combining physics and machine learning reduces forecasting errors by between 20 and 60 per cent compared with the European reference centre, over the challenging two- to six-week forecast window.
▶
Tina Huang passe en revue les grandes familles de modèles d'IA et leurs usages respectifs, dans une mise à jour de sa vidéo de référence sur le sujet.
Mark Zuckerberg defends an AI that is widely disseminated rather than concentrated in a few hands, and promises "personal superintelligence" for everyone.
A survey of nearly 800 architects and designers reveals widespread experimentation and real time savings, but integration remains very limited.
The Israeli-American start-up claims that general-purpose models succeed in 88 per cent of their simulated attacks but detect only 12 per cent of them when asked to defend against them.
▶
Le deuxième épisode de la revue hebdomadaire revient sur les annonces de Google DeepMind, une initiative à 5 milliards de dollars et les portefeuilles de paiement destinés aux agents IA.
▶
Le point quotidien sur l'actualité de l'intelligence artificielle, condensé en une minute : annonces de modèles, financements et faits marquants du jour.
With micro-reactors that can be deployed within eighteen months and conventional reactors that take more than ten years to build, the timescales for nuclear power are poorly aligned with those for AI.
The Danish laboratory wants to shorten the path between the identification of a therapeutic target and the first administration in humans using agentic systems.
Dubai City Council is automating the review of villa plans to speed up the issuing of planning permission, whilst freeing up inspectors to focus on complex projects.
Thanks to 4-bit quantisation, Meta is able to fit a complete agentic model into 18 to 20 GB of VRAM, making it usable on a standard PC or Mac fitted with a consumer-grade graphics card.
An evaluation of 779 dialogues on classic algebra errors shows that chatbots do not tutor spontaneously, and that the quality obtained has almost no relation to the price of the model.
Based on more than 7 million ice thickness measurements, the model estimates the volume of ice outside the ice caps at 150,000 km³, or 32.3 cm of potential sea level rise.
The Northeast & Islands OER has a very thin body of evidence on the real effect of AI tools on learning, and proposes a series of safeguards.
A NBER experiment on more than 6,000 middle school students shows that the help of AI slows down progress but improves recovery after mistakes — provided that it is coupled with a logic of mastery.
Researchers are exploiting the interchangeability of encrypted blocks within the same provider to force a less protected model to render the reasoning of an advanced model in plain text.
Faced with the increase in attacks carried out by autonomous agents, OpenAI is opening its Daybreak cyber defense service to a specialized model, currently reserved for a handful of partners.
A new, ready-to-use AI system published in Nature Cancer identifies eighteen different types of cancers with high accuracy from a reduced number of histological slides.
By replaying the experiments of 168 papers presented at a machine learning conference, AI agents were only able to confirm a minority of the claimed results.
OpenAI cannot rule out that its next model will cross the critical threshold of cybersecurity risk and confine it to isolated test environments.