Air pollution is a global hazard, and as of 2023, 94\% of the world's population is exposed to unsafe pollution levels. Surface Ozone (O3), an important pollutant, and the drivers of its trends are difficult to model, and traditional physics-based models fall short in their practical use for scales relevant to human-health impacts. Deep Learning-based emulators have shown promise in capturing complex climate patterns, but overall lack the interpretability necessary to support critical decision making for policy changes and public health measures. We implement an uncertainty-aware U-Net architecture to predict the Multi-mOdel Multi-cOnstituent Chemical data assimilation (MOMO-Chem) model's surface ozone residuals (bias) using Bayesian and quantile regression methods. We demonstrate the capability of our techniques in regional estimation of bias in North America and Europe for June 2019. We highlight the uncertainty quantification (UQ) scores between our two UQ methodologies and discern which ground stations are optimal and sub-optimal candidates for MOMO-Chem bias correction, and evaluate the impact of land-use information in surface ozone residual modeling.

Spatiotemporal modeling of real-world data poses a challenging problem due to inherent high dimensionality, measurement noise, and expensive data collection procedures. In this paper, we present Sparse Identification of Nonlinear Dynamics with SHallow REcurrent Decoder networks (SINDy-SHRED), a method to jointly solve the sensing and model identification problems with simple implementation, efficient computation, and robust performance. SINDy-SHRED uses Gated Recurrent Units (GRUs) to model the temporal sequence of sensor measurements along with a shallow decoder network to reconstruct the full spatiotemporal field from the latent state space using only a few available sensors. Our proposed algorithm introduces a SINDy-based regularization; beginning with an arbitrary latent state space, the dynamics of the latent space progressively converges to a SINDy-class functional, provided the projection remains within the set. In restricting SINDy to a linear model, the architecture produces a Koopman-SHRED model which enforces a linear latent space dynamics. We conduct a systematic experimental study including synthetic PDE data, real-world sensor measurements for sea surface temperature, and direct video data. With no explicit encoder, SINDy-SHRED and Koopman-SHRED enable efficient training with minimal hyperparameter tuning and laptop-level computing; further, it demonstrates robust generalization in a variety of applications with minimal to no hyperparameter adjustments. Finally, the interpretable SINDy and Koopman models of latent state dynamics enables accurate long-term video predictions, achieving state-of-the-art performance and outperforming all baseline methods considered, including Convolutional LSTM, PredRNN, ResNet, and SimVP.

In order to achieve high data rate and ubiquitous connectivity in future wireless networks, a key task is to efficiently manage the radio resource by judicious beamforming and power allocation. Unfortunately, the iterative nature of the commonly applied optimization-based algorithms cannot meet the low latency requirements due to the high computational complexity. For real-time implementations, deep learning-based approaches, especially the graph neural networks (GNNs), have been demonstrated with good scalability and generalization performance due to the permutation equivariance (PE) property. However, the current architectures are only equipped with the node-update mechanism, which prohibits the applications to a more general setup, where the unknown variables are also defined on the graph edges. To fill this gap, we propose an edge-update mechanism, which enables GNNs to handle both node and edge variables and prove its PE property with respect to both transmitters and receivers. Simulation results on typical radio resource management problems demonstrate that the proposed method achieves higher sum rate but with much shorter computation time than state-of-the-art methods and generalizes well on different numbers of base stations and users, different noise variances, interference levels, and transmit power budgets. Yunqi Wang is with the Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, and also with Shenzhen Research Institute of Big Data, Shenzhen 518172, China (email: yunqi9@connect.hku.hk). Yang Li is with Shenzhen Research Institute of Big Data, Shenzhen 518172, China (e-mail: liyang@sribd.cn).

Mike Pondsmith started playing Dungeons & Dragons in the late seventies, as an undergraduate at the University of California, Davis. The game, published just a few years before, popularized a newish form of entertainment: tabletop role-playing, in which players, typically using dice and a set of rule books, create characters who pursue open-ended quests within an established world. "The most stimulating part of the game is the fact that anything can happen," an early D&D review noted. Soon, other such games hit the market, including Traveller, a sci-fi game published in 1977, the year that "Star Wars" came out. Pondsmith, a tall Black man who grew up in multiple countries because his dad was in the Air Force, loved sci-fi, and fancied himself a bit like Lando Calrissian, the smooth-talking "Star Wars" rogue played by Billy Dee Williams.

This analysis on: The Hustle newsletter (lead story), Hacker News, The Growth Newsletter (issue #021), Social Blade, Data Science Weekly newsletter (issue of Jul 09 2020), Tubefilter. Around 1.5 years ago, I did an analysis of YouTube trending videos in US. That analysis was performed on trending videos of some months in 2017 and 2018. The analysis received a lot of interest on Kaggle and Reddit; I also received some emails praising the work done. That was 1.5 years ago. Today, I present an improved and expanded version of that analysis. This analysis is more advanced and contains new interesting elements. In this analysis, All trending videos for the whole year of 2019 were analyzed (More than 70,000 videos). Titles, descriptions, thumbnails, tags, views, likes/dislikes, and comments were all analyzed to produce the results shown in this post. Continue reading to know more about the analysis and the data or you can jump directly to the results section. The main goal of the analysis is to find interesting facts and patterns by exploring the data and by using effective visualizations. If you don't have time to read the full analysis, here are some of the insights that were extracted from the data along with links to their sections in the analysis: YouTube, as you know, is the most popular and most used video platform in the world today.

The world is about to get a whole lot smarter. As the new decade begins, we're hearing predictions on everything from fully remote workforces to quantum computing. However, one emerging trend is scarcely mentioned on tech blogs – one that may be small in form but has the potential to be massive in implication. There are 250 billion microcontrollers in the world today. Perhaps we are getting a bit ahead of ourselves though, because you may not know exactly what we mean by microcontrollers.

If you were looking for the oddest dish being served in an American restaurant right now, you should probably start with the fish course at Jordan Kahn's new Vespertine, a dish that nudges the idea of culinary abstraction dangerously close to the singularity. It doesn't look like fish, for one thing -- it looks rather like an empty bowl, coarse and pebbly inside and out, of a blackness deep enough to suck up all light, your dreams and your soul. If this were Coi or Alinea, to name two modernist temples, your server would instruct you on how to eat the dish, or at least on where you might direct your spoon. At Vespertine, the server, wearing a severe frock like something out of "The Handmaid's Tale," does not. If you prompt her, she may whisper the word hirame, which in a sushi bar can mean either flounder or halibut.