From physics to generative AI: An AI model for advanced pattern generation

MIT News  September 27, 2023 Researchers at MIT have introduced a new family of physics-inspired generative models termed PFGM++ that unifies diffusion models and Poisson Flow Generative Models (PFGM). The models produced generative trajectories for N dimensional data by embedding paths in N+D dimensional space while still controlling the progression with a simple scalar norm of the D additional variables. The new model reduced to PFGM when D=1 and to diffusion models when D→∞. The flexibility of choosing D allowed them to trade off robustness against rigidity as increasing D resulted in more concentrated coupling between the data and the […]

Improved microphysics modeling of clouds

Phys.org  May 11, 2023 An international team of researchers (USA – Brookhaven National Laboratory, NCAR, UCAR, Canada, Japan) concentrated on several topics that are understudied but hold great potential for further advancing bulk microphysics parameterizations such as multi-moment bulk microphysics parameterizations and the role of the spectral shape of hydrometeor size distributions; discrete vs “continuous” representation of hydrometeor types; turbulence-microphysics interactions etc. and theoretical foundations for the mathematical expressions used to describe hydrometeor size distributions and hydrometeor morphology; and approaches for developing bulk microphysics parameterizations. They presented spectral bin scheme and particle-based scheme (especially, super-droplet method) for representing explicit microphysics […]

Researchers use AI to explore potential zoonotic diseases

Phys.org  April 26, 2023 When certain conditions are met, the passage of viruses from one species to another can ultimately lead to the emergence of a zoonosis. To make better predictions of interactions between mammals and viruses in general, an international team of researchers (Canada, UK, USA – University of Oklahoma, Georgetown University, Washington University) developed an algorithm to sample thousands of species of mammals and even more thousands of viruses to identify which host-virus interactions to explore further. The algorithm represents the system as a network of interactions between viruses and mammals that the algorithm must then complete. It […]

New approach to epidemic modeling could speed up pandemic simulations

Science Daily  January 5, 2023 Simulating large-scale epidemics requires substantial computational resources and in many cases is practically infeasible. One way to reduce the computational cost of simulating epidemics on the networks derived from modern datasets is sparsification, where a representative subset of edges is selected based on some measure of their importance. Researchers at Santa Fe Institute used the effective resistance, which takes both local and global connectivity into account. They tested their method in simulations on a U.S.-wide mobility network and fond that it preserved epidemic dynamics with high fidelity. According to the researchers combined with efficient epidemic […]

New tool helps researchers investigate clouds, rain, and climate change

Phys.org  October 12, 2022 A team of researchers in the US (Pennsylvania State University, Argonne National Laboratory, NASA Goddard Institute for Space Studies, New York, Texas A&M, College Station) developed the Earth Model Column Collaboratory (EMC2), an open-source ground-based lidar and radar instrument simulator and subcolumn generator, specifically designed for large-scale models, in particular climate models, but also applicable to high-resolution model output. It provides a flexible framework enabling direct comparison of model output with ground-based observations, including generation of subcolumns that may statistically represent finer model spatial resolutions and EMC2 large-scale models’ physical assumptions implemented in their cloud or […]

Filling a vital gap in climate models

Phys.org  September 13, 2022 The persistent Southern Ocean (SO) shortwave radiation causes biases in climate models. An international team of researchers (Australia, USA – University of Oklahoma) investigated and characterized cloud and precipitation processes over the Southern Ocean from synoptic to micro scales using the observations collected during the Clouds Aerosols Precipitation Radiation and atmospheric Composition Over the Southern Ocean 2016 and 2018 field campaigns. The cloud and precipitation regimes corresponded to the seven thermodynamic clusters established using a K-means clustering technique, while less distinctions were evident using the cyclone and (cold) front compositing methods. Cloud radar and disdrometer data […]

Mathematical model offers new insights into spread of epidemics

Phys.org  October 8, 2021 The complexity of real-world scenarios still poses new theoretical challenges for mathematical modeling of epidemic spreading. Existing network models of epidemic spreading often focus on contacts between pairs of individuals while co-location of individuals such as in the workplace, restaurants, or gym typically include more than two people. An international team of researchers (Canada, UK, USA – University of Vermont) has developed a new approach to epidemic modeling that considers interactions between two or more people in the same location and for different amounts of time and considered emerging evidence that suggests a minimal infective dose […]

Researchers develop sensitive new way of detecting transistor defects

Nanowerk  October 11, 2021 A team of researchers in the US (NIST, Pennsylvania State University) concentrated on the boundary, or channel, between the thin oxide layer and the bulk semiconductor body which controls the resistance of the device from source to drain. They used electrically detected magnetic resonance (EDMR) to detect defects in the channel. To focus exclusively on activity in the channel, researchers use a technique called bipolar amplification effect (BAE). The parameter of importance is the interface defect density, which is a number that describes how many defects are within some area of the semiconductor-oxide interface. The BAE […]