The estimation of causal effects with observational data continues to be a very active research area. In recent years, researchers have developed new frameworks which use machine learning to relax classical assumptions necessary for the estimation of causal effects. In this paper, we review one of the most prominent methods - "double/debiased machine learning" (DML) - and empirically evaluate it by comparing its performance on simulated data relative to more traditional statistical methods, before applying it to real-world data. Our findings indicate that the application of a suitably flexible machine learning algorithm within DML improves the adjustment for various nonlinear confounding relationships. This advantage enables a departure from traditional functional form assumptions typically necessary in causal effect estimation. However, we demonstrate that the method continues to critically depend on standard assumptions about causal structure and identification. When estimating the effects of air pollution on housing prices in our application, we find that DML estimates are consistently larger than estimates of less flexible methods. From our overall results, we provide actionable recommendations for specific choices researchers must make when applying DML in practice.

Bayesian approaches for training deep neural networks (BNNs) have received significant interest and have been effectively utilized in a wide range of applications. There have been several studies on the properties of posterior concentrations of BNNs. However, most of these studies only demonstrate results in BNN models with sparse or heavy-tailed priors. Surprisingly, no theoretical results currently exist for BNNs using Gaussian priors, which are the most commonly used one. The lack of theory arises from the absence of approximation results of Deep Neural Networks (DNNs) that are non-sparse and have bounded parameters. In this paper, we present a new approximation theory for non-sparse DNNs with bounded parameters. Additionally, based on the approximation theory, we show that BNNs with non-sparse general priors can achieve near-minimax optimal posterior concentration rates to the true model.

A Swampscott, Massachusetts-based IRS agent with specialized knowledge of accounting and investigative auditing techniques was arrested on Wednesday for allegedly filing false tax returns for three years in a row. The Department of Justice (DOJ) announced the arrest of 67-year-old Ndeye Amy Thioub, who has been charged with filing false tax returns in 2017, 2018 and 2019. Thioub has worked for the IRS for more than 17 years and is currently working as a revenue agent with the Large Business and International Division of the IRS. In her role, the DOJ said, Thioub does independent examinations in the field, along with investigations into complex income tax returns of large businesses, corporations and organizations. Charging documents allege that Thioub has specialized knowledge of accounting techniques, practices and investigative auditing techniques.

During the torture ritual that was doing my taxes this year, I was surprised to find myself giddy after reading these words: "You are now chatting with IRS Representative-1004671045." I had gotten stuck trying to parse my W-2, which, under "Box 14: Other," contained a mysterious 389.70 deduction from my overall pay last year. I tapped the chat button on my tax software for help, expecting to be sucked into customer-service hell. Instead, a real IRS employee answered my question in less than two minutes. The program is not TurboTax, or any one of its many competitors that will give you the white-glove treatment only after you pony up. It is Direct File, a new pilot program made by the IRS.

Michael Cohen will not face sanctions after he cited fake legal cases in a court filing generated by artificial intelligence, a federal judge said Wednesday. Cohen, former President Trump's onetime fixer and lawyer, had pleaded guilty to tax and campaign finance violations and is currently under supervised release. He has repeatedly sought to have his sentence reduced, and in his most recent attempt, Cohen provided his attorney with fabricated case citations he later admitted were generated by Google's AI chatbot, formerly known as Bard. U.S. District Judge Jesse Furman said the false citations were "embarrassing and certainly negligent" in a 13-page order that denied Cohen's fourth motion for early termination of supervised release. But the judge found that Cohen, who had said he misunderstood how AI works and did not intend to cite fake cases, had not acted in "bad faith" and that neither he nor his lawyer, David Schwartz, should be sanctioned.

A group of researchers backed by the French government have released what is thought to be the largest AI training dataset composed entirely of text that is in the public domain. "There's no fundamental reason why someone couldn't train an LLM fairly," says Ed Newton-Rex, CEO of Fairly Trained. He founded the nonprofit in January 2024 after quitting his executive role at image generation startup Stability AI because he disagreed with its policy of scraping content without permission. Fairly Trained offers a certification to companies willing to prove that they've trained their AI models on data that they either own, have licensed, or is in the public domain. When the nonprofit launched, some critics pointed out that it hadn't yet identified a large language model that met those requirements.

'Seen and Unseen': Fox News' Raymond Arroyo has the latest on President Biden's dog Commander and the controversy surrounding Google's A.I.-generated historical images on'The Ingraham Angle.' BEHIND THE CURTAIN: Google abandoned "fairness" and took major "shortcuts" to launch the Gemini artificial intelligence chatbot despite internal concerns, according to a former high-level employee. IN THE WORKS: Apple is in talks with Google to use its new Gemini artificial intelligence models to power the AI features for iPhones after previously discussing the prospect with ChatGPT maker OpenAI, according to a new report. 'CRAZY AND WEIRD': LSU women's basketball star Angel Reese took to social media Monday to call out those allegedly creating AI-generated photos of the college basketball player. Angel Reese #10 of the LSU Lady Tigers looks on against the Tennessee Lady Vols in the first quarter at Thompson-Boling Arena on February 25, 2024 in Knoxville, Tennessee. LAWN BEAST: Imagine a future where the hum of lawn mowers and the rustle of leaves being raked are sounds of the past, replaced by quiet and efficient robots.

Artificial intelligence (AI) could take over more than eight in 10 repetitive jobs performed by civil servants, a study has found. From processing passports to registering to vote, at least 120 million tasks across government have the potential to be automated. Every minute AI helped cut per transaction would save hundreds of thousands of hours of manual work by human staff. The Alan Turing Institute, which carried out the research, said it would free up officials from never-ending bureaucracy and spend more time dealing with the public. Last month, the Deputy Prime Minister promised AI would end'timewasting, pencil-pushing, computer-saysno' frustrations of dealing with public services.

An interactive exhibit opening Wednesday at the National WWII Museum will use artificial intelligence to let visitors hold virtual conversations with images of veterans, including a Medal of Honor winner who died in 2022. Voices From the Front will also enable visitors to the New Orleans museum to ask questions of war-era home front heroes and supporters of the U.S. war effort -- including a military nurse who served in the Philippines, an aircraft factory worker, and Margaret Kerry, a dancer who performed at USO shows and, after the war, was a model for the Tinker Bell character in Disney productions. Four years in the making, the project incorporates video-recorded interviews with 18 veterans of the war or the support effort -- each of them having sat for as many as a thousand questions about the war and their personal lives. Among the participants was Marine Corps veteran Hershel Woodrow "Woody" Wilson, a Medal of Honor Winner who fought at Iwo Jima, Japan. He died in June 2022 after recording his responses.

France's competition watchdog has fined Google 250 million euros ( 272m) for breaching commitments to media companies on content licensing. The French Competition Authority said on Wednesday that it was imposing the fine as part of additional measures over a 2019 case that organisations representing French magazines and newspapers had lodged against the United States tech giant and other online platforms. The media outlets accused the tech companies of making billions from their content without sharing the revenue with those who gathered it. In 2021, the watchdog fined Google 500 million euros ( 592m) for failing to negotiate in good faith. The dispute appeared to be resolved in 2022 when the company dropped its appeal against the fine.