We study the $L_1$-regularized maximum likelihood estimator/estimation (MLE) problem for discrete Markov random fields (MRFs), where efficient and scalable learning requires both sparse regularization and approximate inference. To address these challenges, we consider a stochastic learning framework called stochastic proximal gradient (SPG; Honorio 2012a, Atchade et al. 2014,Miasojedow and Rejchel 2016). SPG is an inexact proximal gradient algorithm [Schmidtet al., 2011], whose inexactness stems from the stochastic oracle (Gibbs sampling) for gradient approximation - exact gradient evaluation is infeasible in general due to the NP-hard inference problem for discrete MRFs [Koller and Friedman, 2009]. Theoretically, we provide novel verifiable bounds to inspect and control the quality of gradient approximation. Empirically, we propose the tighten asymptotically (TAY) learning strategy based on the verifiable bounds to boost the performance of SPG.

Neural language models are becoming the prevailing methodology for the tasks of query answering, text classification, disambiguation, completion and translation. Commonly comprised of hundreds of millions of parameters, these neural network models offer state-of-the-art performance at the cost of interpretability; humans are no longer capable of tracing and understanding how decisions are being made. The attention mechanism, introduced initially for the task of translation, has been successfully adopted for other language-related tasks. We propose AttViz, an online toolkit for exploration of self-attention---real values associated with individual text tokens. We show how existing deep learning pipelines can produce outputs suitable for AttViz, offering novel visualizations of the attention heads and their aggregations with minimal effort, online. We show on examples of news segments how the proposed system can be used to inspect and potentially better understand what a model has learned (or emphasized).

More than 73 million user records have been stolen from a number of online services and are being sold on the dark web for a total of $18,000. The hacker group, named'ShinyHunters', scraped data from at least 10 companies including Zoosk, Chatbooks and the StarTribune. The online dating app, Zoosk, had the largest breach, as the cybercriminals grabbed 30 million user record and Chatbooks came in second with 15 million, according to ZDNet. Chatbooks is the only firm to acknowledge the attack, which said login credentials including names, email address and passwords were taken in the haul, along with some phone numbers and FacebookIDs. More than 73 million user records have been stolen from a number of online services and are being sold on the dark web for a total of $18,000.

Horses can form deep and long lasting connections with their human keepers according to new research from France. The experiment, led by Léa Lansade of the French National Research Institute for Agriculture, Food and Environment, found that horses were able to identify their keepers when presented with a photo of them and a random human about 75 percent of the time. The results were surprising not just because they suggest horses form emotional attachments to their human companions, but because it shows horses understand photographs as symbolic representations. According to a new study from France, horses seem to be able to recognize photos of their human handlers, and can even identify former handlers they haven't seen or worked with in six months or more Past research has shown horses can identify their keepers based on smells or sound cues, according to a report in Scientific American, but Lansade's study is the first to show two-dimensional images can also have significance to horses. '[T]hese results show that horses have advanced face-recognition abilities, and are able, like humans, to differentiate between a photograph of a familiar and unfamiliar individual, even when the faces did not belong to their own species,' the team writes.

We've seen AI outperform doctors in spotting breast cancer, lung cancer and skin cancer. Now, researchers from Intel and the University of Pennsylvania are turning their attention to brain tumors. Using Intel's AI hardware and software, Penn Medicine will lead 29 international healthcare and research institutions in creating an AI model trained on the largest brain tumor dataset ever -- and will do so without sharing sensitive patient data. The project is based on a technique called federated learning, which trains an algorithm across decentralized servers, so that hospitals can work together without actually sharing patient data. This will allow the institutions -- from the US, Canada, the UK, Germany, the Netherlands, Switzerland and India -- to create a much larger data set than any one institution would be able to on its own.

Almost two years after it announced the game, you might think Bethesda is ready to share more information about its latest mainline Elder Scrolls project. But that would be a wrong assumption. Prompted by a fan over the weekend, Bethesda senior vice president of marketing and communications Pete Hines said the company won't have anything substantial to say about The Elder Scrolls 6 for a couple more years. "It's after Starfield, which you pretty much know nothing about," he said. "So if you're coming at me for details now and not years from now, I'm failing to properly manage your expectations."

The lectern stands before a vast, grassy amphitheater. Above, the sky is a pristine blue, the cloudless quintessence of a spring morning. Perhaps the most arresting feature in this panorama is the stage: a towering neoclassical structure with Doric columns that soar heavenward. Suspended from the columns are banners emblazoned with the letters QU, for Quaranteen University. The name is a wink from the worldbuilders--college and high school students, homebound for weeks now--about the surreal circumstances of their situation.

As countries around the world are gradually reopening following lockdowns, government authorities are using surveillance drones in an attempt to enforce social distancing rules. In India, police are using AI-equipped drones developed by US start-up Skylark Labs to monitor evening curfews and the distance between people who are outside during the day. The drones are being flown in six cities in the northern state of Punjab, and are also being trialled in the southern city of Bangalore, says Skylark Labs CEO Amarjot Singh. Each drone is fitted with a camera and an AI that can detect humans within a range of 150 metres to 1 kilometre. If it spots people it can send an alert to police in the district located nearest to the sighting.

For many ASEAN nations, manufacturing takes up a sizable chunk of the national GDP. For example, manufacturing in Singapore contributed 20.9 percent of Singapore's GDP in 2019, according to the department of statistics Singapore. Successful companies readily acknowledge one key factor contributing to their achievements – hardworking, committed and skilled employees who are the foundation of the companies. However, manufacturers today face the primary challenge of filling open positions with skilled workers, which in turn affects overall productivity and growth. In addition to the human capital challenge, manufacturers are facing immense costs associated with workplace injuries.

Yeah, so the entire project is a coordinated effort by the White House Office of Science and Technology Policy. I think some time in early March a group at Georgetown, the Center for Security in Emerging Technology (CSET) reached out to us at Allen AI to help coordinate the release of this dataset, along with a couple of different organizations. You mentioned MSR (Microsoft Research), Chan Zuckerberg, Kaggle was also involved, and the National Library of Medicine, which is part of the NIH. So all these groups - we're going to come together to essentially create this dataset to help create text mining and information retrieval tools that could assist medical experts in understanding more of what was going on with the epidemic. For Allen AI, the way that we got involved is we had recently created a new pipeline to revamp our open research corpus.