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
Thought for celestial navigation when GNSS is unavailable, the method is more resistant than its predecessors to the disturbances of a terrestrial observation, at a comparable time of calculation.
Driven by self-supervision on colored H&E cuts, the model reaches a AUC of 0.91 to distinguish tumour and normal tissue, then groups tumours into four pseudo-stages.
▶
Une ligne de prompt modifiée a rendu un agent de support 17 points plus sûr… et 17 points moins efficace. Sofia Jakovcevic (Arize AI) détaille une boucle de non-régression construite sur les traces de production.
Pre-recorded study on a vision-driven challenge: reconnecting recovery memory to a LoRA policy adds 34 points, distillation alone barely 1 point.
Guided by major climate modes (MJO, NAO, ENSO), the model improves the probabilistic score of weekly rain and temperature forecasts by 22 to 34%, without calibrating the extremes.
Abstracts and derived states keep track of what has been deleted: only one question is enough to find the deleted secret in Qwen2.5-7B in about two-thirds of cases.
The Israeli start-up, supported by Red Dot Capital Partners, wants AI to investigate alerts directly in existing tools, without forcing companies to centralize their data.
Turnover per employee and margins are increasing among the Seven Magnificents, not in the rest of the S&P 500: for Apollo's chief economist, tokens become a cost centre.
Google's new privacy policy opens an exception: what the user voluntarily connects to certain services to customize Gemini or Search can also customize ads.
Announced with the budget, this national laboratory must strengthen British defences against biological threats; OpenAI and Anthropic bring 200 million pounds of credit to use their models.
▶
Le fonds Bridgewater raconte la construction du « Vault », un corpus de plus de 100 millions de documents interrogeable par ses analystes comme par ses agents IA.
32 participants learned science with GPT-5-mini assistants, some of which had been falsified: scores dropped from 7.4 to 2.9 out of 10.
A French team publishes a "grammar" of civil cassation (2016-2025) which shows what means come to pass, what texts they invoke and how it evolves.
Based on the energy diagnostics alone, a CatBoost model estimates the savings of each work with their uncertainty and shows that the order of construction sites counts.
A study attempts to measure the "alternative boundary": the length of tasks that agents succeed doubles every 129 days, but market signals remain mixed.
Tested on 10,497 confirmed transients, the Fritz production module recovers 83% of non-taxonomy objects by inspecting only 15% of the flow.
Four controls defined by two clinicians, formally verified and applied to the responses: each verdict is recalculated on the basis of a named rule and sixteen facts.
Meteorologists in Munich re-use the model decoder to read their internal states during calculation: quality progresses layer after layer, from large scales to small ones.
By alternating the lighting, the sensor alternates to capture an image through its membrane and a touch print, from the same point of view and without calibration between cameras.
A Rowhammer attack targets the Triton compiler: the GPU kernel produced remains correct but less fast, and as it is cached, the loss of 4 to 16% persists from one execution to the next.