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
Six months after its launch, Gerard Williams’ start-up – backed by Sequoia – is reportedly seeking several hundred million to develop a data centre CPU designed for AI.
On simulated observations of XMM-Newton, a CNN working directly on photon-by-photon event cubes proves to be the most robust classifier for extended and weak X-ray sources.
Without any data from the target platform, Jev achieves an average AUC of 0.706 in knowledge tracing, outperforming the best deep learning models trained on eight learners, at one hundredth the cost of a reasoning LLM.
Researchers have calculated the ‘segmentation ceiling’: masks accurate to within 10 per cent would be required to outperform direct regression on video, whereas a typical segmenter is off by 14 per cent.
A study combining an audit of 100 start-ups with 63 interviews with experts shows how the ‘AI’ label is overhyped to attract investors, and how it subsequently erodes consumer confidence.
▶
Charles Dickens (Snorkel AI) raconte comment un modèle de 4 milliards de paramètres, entraîné par renforcement pour moins de 500 dollars, a battu son grand frère de 235 milliards sur des questions financières.
▶
Dominik Kundel (OpenAI) présente le Codex App Server, le protocole JSON-RPC open source qui fait tourner l'app Codex, et montre comment l'intégrer dans ses propres outils.
A method that does not require specific training involves identifying, verifying and then searching for missing food items using a multimodal model, thereby improving the estimation of meal mass and energy content.
Weather and climate models are being developed faster than our capacity to evaluate them. The authors propose five priorities and three areas of focus for the coming decade, including a programme dedicated to their reliability.
Using a simplified model of a pulsar array, four researchers explain why simulation inference detects gravitational wave background anisotropies better than conventional statistics.
Ahead of Argon’s public roll-out, Google’s teams are testing a new version of Gemini 4 on their internal code platform. Its release is not guaranteed.
At the "Science: A New Golden Age" summit, Donald Trump also presented the National Medal for Technology to Satya Nadella and Michael Dell, in the presence of Michael Kratsios.
The German Transport Minister considered the name to be misleading. In return for the change, he will support the system’s approval across the EU ahead of a vote expected by the end of the year.
The start-up, whose turnover has increased twelvefold in a year, automates the process of granting mortgages using networks of agents. Three of its clients, including Pennymac, are taking a stake in the company.
The Brooklyn-based start-up, which rents out its robots to warehouses on a monthly subscription basis, has secured a $50 million Series A funding round led by Framework Ventures and is strengthening its partnership with Physical Intelligence.
During an automated test, an Anthropic model filled out the Philadelphia police public form with a false lead. The city finds the two-month period before being notified "unacceptable".
▶
Plutôt que de répéter les mêmes corrections à un agent de code sans mémoire, Greg Pstrucha (Sentry) défend l'idée de les graver dans des tests, des types stricts et surtout des règles de lint.
This closed-loop framework guides the surgeon to areas to palpate rather than specific points, and determines for itself when realignment is sufficient, without any reference to actual conditions on the ground.
Rather than being subject to odometry drift, the robot plans deliberate contacts with walls and obstacles in order to re-localise itself, and reaches its destination more often than the benchmark methods.
Based on more than twenty years of NOAA observations, the model reconstructs continuous wind profiles from sparse probe measurements, and can extend surface winds observed by satellite into three dimensions.