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
Tested on a FastNet configuration producing 1.5 km forecasts on the UK, a wavelet loss function gives more realistic precipitation than quadratic error.
Where BurnP3+ requires days or weeks per region, the network estimates in seconds the likelihood of burning, intensity and rate of spread.
Tested on 11 Sen1Floods11 floods, the NASA and IBM model remains slightly behind a conventional architecture in terms of precision, but divides the calculation time by more than half.
In Tucson, the Project Blue project brought together opponents from all sides. The city blocked the annexation of the land and the county frozen the new projects for 120 days, but the site continues.
By bringing 134 published estimates back to the same measuring area, TokWatt sets reference values on six orders of magnitude, from the simple response to the generated video.
Up to 24 hours of maturity, FitzRoy equals or exceeds the state of art without the blurred rain fields typical of AI, and reports record cumulations of the Texas and Missouri floods.
Following the decline in inverter production by inverter, a multimodal model anticipates cloud passage and advances 14 reference methods, including overcast or changing skies.
A forecast says where the storm will hit, not what crops will suffer. WxCrop links weather indices, causal analysis and insurance data to estimate losses, to within 2% over years of testing.
This deep model, guided by the cause-effect bonds discovered on 522 cyclones in the Northwest Pacific, better alerts the most dangerous phase than four reference systems.
By crossing satellite maps of floods and telephone data, researchers find in India and Indonesia the pattern observed in rich countries: the disadvantage weighs on recovery.
Noise Steering controls both the orientation and the power of each turbine; where conventional bridging costs 17% production, it loses only 2% on average.
Driven over nine years of 3-km-resolution analysis, HRRCast has been running in near real time for the past year. Over seven months of independent verification, its overall average exceeds the operational model on several surface variables.
The chemical mechanisms of the atmosphere circulate in a format designed for Fortran and C. KPPax translate them into driftable JAX code, connecting them directly to neuron networks.
Driven across eight GPU H200, this quarter-degree overall model beats the entire ECMWF European centre on the 600 combinations of variables, levels and deadlines tested.
Send each query to the least carbonized region forgets the cost of recalculating history. By taking this into account, CarbonRouter reduces emissions by an average of 16%.
A study measures the energy of 11 open models on EffiBench's 1,000 problems: the "optimized" code consumes up to 886% more, and optimization itself is almost never profitable.
Electricity, water, emissions and electronic waste: researchers extend the notions of green AI and red AI to the entire IA-data-telecom pile and classify remedies into four levels.
An index learned by total graph variation, combined with the humidity of Indochina soils, predicts summer rains in the Yangtse basin better than the El Niño indicators.
On 56,078 windows of two solar installations, combining LightGBM with seven weighted neuron networks is better than the simple average or the indistinct addition of modules.
Researchers propose estimating diffuse solar radiation without instruments on the ground, using fisheye images of the sky and calculated atmospheric parameters.