Providing better language tools for low-resource and endangered languages is imperative for equitable growth. Recent progress with massively multilingual pretrained models has proven surprisingly effective at performing zero-shot transfer to a wide variety of languages. However, this transfer is not universal, with many languages not currently understood by multilingual approaches. It is estimated that only 72 languages possess a "small set of labeled datasets" on which we could test a model's performance, the vast majority of languages not having the resources available to simply evaluate performances on. In this work, we attempt to clarify which languages do and do not currently benefit from such transfer. To that end, we develop a general approach that requires only unlabelled text to detect which languages are not well understood by a cross-lingual model. Our approach is derived from the hypothesis that if a model's understanding is insensitive to perturbations to text in a language, it is likely to have a limited understanding of that language. We construct a cross-lingual sentence similarity task to evaluate our approach empirically on 350, primarily low-resource, languages.

About three weeks ago, I received an email with an access code for an app that allowed me to take rides in a robotaxi. The app comes from Cruise, a startup that was acquired by General Motors in 2016 for just under a billion dollars. With it, I could now use the Cruise robotaxis, which have been operating in San Francisco since August 2021, for myself. What makes them special is that these robotaxis drive driverless. There is no one in the car when it picks up passengers. Thanks to a friend who had been given access to the app a few months earlier, I was able to make my first two trips as early as the beginning of July. I reported on that, especially because two coyotes had crossed our path.

Can continuous diffusion models bring the same performance breakthrough on natural language they did for image generation? To circumvent the discrete nature of text data, we can simply project tokens in a continuous space of embeddings, as is standard in language modeling. Through qualitative and quantitative evaluation, we show that our text diffusion models generate samples comparable with those produced by standard autoregressive language models -- while being in theory more efficient on accelerator hardware at inference time. Our work paves the way for scaling up diffusion models for text, similarly to autoregressive models, and for improving performance with recent refinements to continuous diffusion. Continuous diffusion models (Sohl-Dickstein et al., 2015) have taken the world of image generation by storm, advancing the state of the art further than ever before (Rombach et al., 2021; Ramesh et al., 2022). Diffusion for language is indeed an attractive prospect. Compared to autoregressive (AR) models (Bengio et al., 2000; Sutskever et al., 2011; Austin et al., 2021; Hoffmann et al., 2022), diffusion models can predict all tokens in a sequence at once. This allows for bidirectional, rather than causal attention-- increasing interactions between tokens, potentially leading to more coherent samples. Diffusion models can make a better usage of hardware accelerators during inference than AR models, since computations are parallelizable over the sequence axis.

Recent lay language generation systems have used Transformer models trained on a parallel corpus to increase health information accessibility. However, the applicability of these models is constrained by the limited size and topical breadth of available corpora. We introduce CELLS, the largest (63k pairs) and broadest-ranging (12 journals) parallel corpus for lay language generation. The abstract and the corresponding lay language summary are written by domain experts, assuring the quality of our dataset. Furthermore, qualitative evaluation of expert-authored plain language summaries has revealed background explanation as a key strategy to increase accessibility. Such explanation is challenging for neural models to generate because it goes beyond simplification by adding content absent from the source. We derive two specialized paired corpora from CELLS to address key challenges in lay language generation: generating background explanations and simplifying the original abstract. We adopt retrieval-augmented models as an intuitive fit for the task of background explanation generation, and show improvements in summary quality and simplicity while maintaining factual correctness. Taken together, this work presents the first comprehensive study of background explanation for lay language generation, paving the path for disseminating scientific knowledge to a broader audience. CELLS is publicly available at: https://github.com/LinguisticAnomalies/pls_retrieval.

The Cyberguy, Kurt Knutsson, discussed how Microsoft is set to have ten thousand employees in China and warns about rogue fraudulent apps on phone app stores, on'Fox & Friends Weekend.' Microsoft is buying Activision Blizzard – the video game company behind titles like "Guitar Hero," "Candy Crush," "World of Warcraft" and "Call of Duty" – in the largest tech acquisitions in history. Antitrust regulators are assessing whether the deal could hurt competition in the booming global video game industry, for numerous reasons including the fact that Microsoft already produces the widely used Xbox gaming console. Further consolidation in the tech industry elicits well deserved skepticism from regulators, but they should also consider how this deal could help America's greatest geopolitical adversary: China. To understand how, it's first important to understand Microsoft's long and close relationship with the People's Republic of China and its ruling Chinese Communist Party (CCP). Microsoft has operated in China for three decades, boasting on its website that its "most complete subsidiary and largest R&D center outside the United States is in China."

Time Series Forecasting has been an active area of research due to its many applications ranging from network usage prediction, resource allocation, anomaly detection, and predictive maintenance. Numerous publications published in the last five years have proposed diverse sets of objective loss functions to address cases such as biased data, long-term forecasting, multicollinear features, etc. In this paper, we have summarized 14 well-known regression loss functions commonly used for time series forecasting and listed out the circumstances where their application can aid in faster and better model convergence. We have also demonstrated how certain categories of loss functions perform well across all data sets and can be considered as a baseline objective function in circumstances where the distribution of the data is unknown. Our code is available at GitHub: https://github.com/aryan-jadon/Regression-Loss-Functions-in-Time-Series-Forecasting-Tensorflow.

Fox News Flash top headlines are here. Check out what's clicking on Foxnews.com. Chief diplomats from the world's leading democracies rallied together in a joint statement condemning global adversaries like Iran and North Korea and called on Russia and China to remember their security commitments to the United Nations. After two days of meetings, officials from the Group of 7 (G7) released a lengthy statement Friday in an address to its top geopolitical challengers, warning them to adhere to international laws. United States Secretary of States Antony Blinken and Foreign Minister Yoshimasa Hayashi of Japan, right, meet for bilateral talks at the G7 Foreign Ministers' Meeting in Muenster, Germany, Friday, Nov. 4, 2022.

Analytics and business intelligence (BI) have long been understood to be fundamental to business success. Today, powerful technologies, including artificial intelligence (AI) and machine learning (ML), make it possible to gain deeper insights into all areas of business activity in order to drive efficiency, reduce waste and gain a better understanding of customers. Truly benefiting from analytics – particularly the most advanced and powerful analytics techniques involving AI – requires developing a top-to-bottom culture of data literacy throughout an organization and this, in my experience, is where many businesses are still failing. This is highlighted by one particular statistic that came up during my recent webinar conversation with Amir Orad, CEO of Sisense. Orad told me that according to his observations, 80 percent of employees in the average organization simply aren't leveraging the analytics that, in theory, are available to them.

In this paper, we present Pangu-Weather, a deep learning based system for fast and accurate global weather forecast. For this purpose, we establish a data-driven environment by downloading $43$ years of hourly global weather data from the 5th generation of ECMWF reanalysis (ERA5) data and train a few deep neural networks with about $256$ million parameters in total. The spatial resolution of forecast is $0.25^\circ\times0.25^\circ$, comparable to the ECMWF Integrated Forecast Systems (IFS). More importantly, for the first time, an AI-based method outperforms state-of-the-art numerical weather prediction (NWP) methods in terms of accuracy (latitude-weighted RMSE and ACC) of all factors (e.g., geopotential, specific humidity, wind speed, temperature, etc.) and in all time ranges (from one hour to one week). There are two key strategies to improve the prediction accuracy: (i) designing a 3D Earth Specific Transformer (3DEST) architecture that formulates the height (pressure level) information into cubic data, and (ii) applying a hierarchical temporal aggregation algorithm to alleviate cumulative forecast errors. In deterministic forecast, Pangu-Weather shows great advantages for short to medium-range forecast (i.e., forecast time ranges from one hour to one week). Pangu-Weather supports a wide range of downstream forecast scenarios, including extreme weather forecast (e.g., tropical cyclone tracking) and large-member ensemble forecast in real-time. Pangu-Weather not only ends the debate on whether AI-based methods can surpass conventional NWP methods, but also reveals novel directions for improving deep learning weather forecast systems.

Yellow Submarine, Ringo Starr's turn on Revolver, has been a gateway for children into the music of the Beatles since its release in 1966. A new reissue of the album makes that relationship more explicit: Giles Martin, son of original producer George and the sonic custodian of the Beatles catalogue, says his "de-mixing" of the album – using AI to separate individual instruments that were originally squeezed together on four tracks – was done in part with a playlist-listening younger audience in mind. Martin recently told Variety that his teenage children listen to old and new music side by side, veering from Fleetwood Mac to Billie Eilish and Olivia Rodrigo. "[W]hat I want to make sure is that when people hear the Beatles, that it has the same dynamic as the other stuff they're listening to," he said. He added that 1969's Abbey Road, recorded on a then luxuriant eight tracks and the first Beatles album not released in mono, stands out from the band's catalogue as "it sounds more hi-fi than the other Beatles albums".