We propose a Distributional Approach to address Controlled Text Generation from pre-trained Language Models (LMs). This view permits to define, in a single formal framework, "pointwise" and "distributional" constraints over the target LM -- to our knowledge, this is the first approach with such generality -- while minimizing KL divergence with the initial LM distribution. The optimal target distribution is then uniquely determined as an explicit EBM (Energy-Based Model) representation. From that optimal representation we then train the target controlled autoregressive LM through an adaptive distributional variant of Policy Gradient. We conduct a first set of experiments over pointwise constraints showing the advantages of our approach over a set of baselines, in terms of obtaining a controlled LM balancing constraint satisfaction with divergence from the initial LM (GPT-2). We then perform experiments over distributional constraints, a unique feature of our approach, demonstrating its potential as a remedy to the problem of Bias in Language Models. Through an ablation study we show the effectiveness of our adaptive technique for obtaining faster convergence.

Health misinformation on social media devastates physical and mental health, invalidates health gains, and potentially costs lives. Understanding how health misinformation is transmitted is an urgent goal for researchers, social media platforms, health sectors, and policymakers to mitigate those ramifications. Deep learning methods have been deployed to predict the spread of misinformation. While achieving the state-of-the-art predictive performance, deep learning methods lack the interpretability due to their blackbox nature. To remedy this gap, this study proposes a novel interpretable deep learning approach, Generative Adversarial Network based Piecewise Wide and Attention Deep Learning (GAN-PiWAD), to predict health misinformation transmission in social media. Improving upon state-of-the-art interpretable methods, GAN-PiWAD captures the interactions among multi-modal data, offers unbiased estimation of the total effect of each feature, and models the dynamic total effect of each feature when its value varies. We select features according to social exchange theory and evaluate GAN-PiWAD on 4,445 misinformation videos. The proposed approach outperformed strong benchmarks. Interpretation of GAN-PiWAD indicates video description, negative video content, and channel credibility are key features that drive viral transmission of misinformation. This study contributes to IS with a novel interpretable deep learning method that is generalizable to understand other human decision factors. Our findings provide direct implications for social media platforms and policymakers to design proactive interventions to identify misinformation, control transmissions, and manage infodemics.

You won't have to go out of your way to catch Queen Elizabeth II's annual Christmas Day message if you have an Echo (or a similar device) on hand. The Guardian reports that the Queen's message will be available on smart speakers for the first time through Amazon's Alexa. So long as you live in an English-speaking country, you can ask Alexa to "play the Queen's Christmas day message" after 3PM GMT (10AM ET) and get the inspiring speech while you're finishing a holiday meal. Google Assistant and HomePod users are out of luck for the on-demand message, but you can always stream BBC Radio 4 on your speaker to get the live broadcast. It's a relatively late move when smart speakers have been around for several years.

An increasing variety of technologies such as artificial intelligence, drones and high-quality 4K video cameras is being introduced in the field of security amid a serious shortage of personnel in the field. A virtual "AI guard" developed by major Japanese security firm Secom Co. was tested at Ogikubo Hospital in Tokyo in late October. An animated character displayed on an electric panel at the hospital entrance takes visitors' temperatures and then welcomes those without fevers into the facility. The character has been programmed to respond verbally to basic inquiries and can tell visitors where the bathrooms are located and what time their buses will arrive. It is also able to make eye contact with visitors and lean down when approached by children or people in wheelchairs.

Fox News Flash top headlines are here. Check out what's clicking on Foxnews.com. Two women were caught on video fighting over a PS5 video game console at Walmart. The video, which has since gone viral on social media, was taken at a Walmart location in Charlotte, North Carolina, according to TMZ. "What the f–k are you gonna do?" one woman can be heard saying in the video as she takes off her purse and jacket. One man can be heard yelling, "Call the cops!" as a crowd gathered around the two brawling women.

Early one morning at the beginning of December, I rushed to the nearest newsstand to purchase a copy of the Nikkei Marketing Journal. It was somehow reassuring once again to see, emblazoned atop page one, the publication's traditional sumo-style banzuke (ranking sheet) -- a layout virtually unchanged since 1971 -- listing Japan's top-selling hit products of 2020. Before dissecting the 2020 rankings, it's worth examining how this annual list began. The 1970s marked the time when discerning consumers in Japan began showing a preference for greater variety. Prior to that time, manufacturers had been content to sell their rice cookers, washing machines, TV sets and other mass-produced household items by appealing mainly through brand affinity.

The Sonos Arc is a pricy but peerless home audio package. Sonos' first-ever Dolby Atmos soundbar is more than just a pretty (and tubular) face. This powerful hunk of well fashioned plastic is loaded with 11 individually powered drivers, including dual speakers pointed upward to bounce off your ceiling, creating an impressively potent example of the hemispheric immersion for which Dolby's Atmos sound format is so highly praised. Its well-tuned drivers offer a rich and smooth sound signature that's fantastic for anything you play, from sitcoms to streaming music. Speaking of streaming, like all Sonos speakers, the Arc offers WiFi connection through the Sonos app, along with the ability to connect with other Sonos speakers, either in a group, or as part of a surround sound setup. This allows you to add dual surround speakers and a Sonos subwoofer if you want, though the Arc offers impressive bass response even without one. Like the Sonos Beam, it also sports built-in microphones and your choice of Google Assistant or Amazon Alexa support, making it both a soundbar and a powerful smart speaker. The drawback of all these features is, of course, the price of $799, which puts the bar beyond the average budget.

On Dec. 11, the Wall Street Journal published an op-ed by Joseph Epstein claiming that Jill Biden's use of the title "Dr." feels "fraudulent, even comic," and referring to her as "kiddo." The tone of the piece is all too familiar to women, especially Black and brown women, across academia, who know what it feels like to be constantly questioned about their expertise from peers and strangers alike. The article led to an uproar, with academics across disciplines calling out its sexism and snobbery. Responses in the Atlantic, the New York Times, Forbes, and elsewhere argued that Biden should ignore Epstein's advice to "consider stowing" her degree and stick to her honorific. Academia is no stranger to overt and subtle sexism, which can manifest through student evaluations or even how women academics are introduced during speaking engagements.

Two weeks after the departure of AI researcher Dr. Timnit Gebru, Google's Ethical AI team has issued a list of demands to its management. In a letter obtained by NBC News and Bloomberg, the division calls for the removal of Megan Kacholia, vice president of engineering at Google, from the group's reporting structure. "We have lost trust in her as a leader," the team says of Kacholia, according to Bloomberg. The team also calls for Google to bring back Gebru "at a higher level" than the one she had before her departure. They also asked that Kacholia and AI Chief Jeff Dean apologize to Gebru over their handling of the situation and that Google put in place a racial literacy training program for management.

Teachers and students (or anyone else in the public, for that matter) can now explore more than 1.5 million historical newspaper images online using artificial intelligence. The latest machine learning experience from LC Labs, Newspaper Navigator allows users to search visual content in American newspapers dating from 1789-1963. The user begins by entering a keyword that returns a selection of photos. Then the user can choose photos to search against, allowing the discovery of related images that were previously undetectable by search engines. For decades, partners across the United States have collaborated to digitize newspapers through the Library's Chronicling America website, a database of historical U.S. newspapers.