We study nonstationary newsvendor problems under nonparametric demand models and general distributional measures of nonstationarity, addressing the practical challenges of unknown degree of nonstationarity and demand censoring. We propose a novel distributional-detection-and-restart framework for learning in nonstationary environments, and instantiate it through two efficient algorithms for the uncensored and censored demand settings. The algorithms are fully adaptive, requiring no prior knowledge of the degree and type of nonstationarity, and offer a flexible yet powerful approach to handling both abrupt and gradual changes in nonstationary environments. We establish a comprehensive optimality theory for our algorithms by deriving matching regret upper and lower bounds under both general and refined structural conditions with nontrivial proof techniques that are of independent interest. Numerical experiments using real-world datasets, including nurse staffing data for emergency departments and COVID-19 test demand data, showcase the algorithms' superior and robust empirical performance. While motivated by the newsvendor problem, the distributional-detection-and-restart framework applies broadly to a wide class of nonstationary stochastic optimization problems. Managerially, our framework provides a practical, easy-to-deploy, and theoretically grounded solution for decision-making under nonstationarity.

When the Rabbit R1 arrived earlier this year, it was an unfinished product. Engadget's own Devindra Hardawar called it "a toy that fails at almost everything." Most of the features Rabbit promised, including its signature "large action model" (LAM), were either missing at launch or didn't work as promised. Now, after more 20 software updates since the spring, Rabbit is releasing its most substantial update yet. Starting today, every R1 user now has beta access to teach mode, a feature that allows you to train Rabbit's AI model to automate tasks for you on any website you can visit from your computer.

Sometimes there are no background noises when there should be. Or, in the case of the robocall, there's a lot of noise mixed into the background almost to give an air of realness that actually sounds unnatural. With photos, it helps to zoom in and examine closely for any "inconsistencies with the physical world or human pathology", like buildings with crooked lines or hands with six fingers, Lyu said. Little details like hair, mouths and shadows can hold clues to whether something is real. Hands were once a clearer tell for AI-generated images because they would more frequently end up with extra appendages, though the technology has improved and that's becoming less common, Lyu said.

This story was supposed to have a different beginning. You were supposed to hear about how, earlier this week, I attended a splashy launch party for a new AI gadget--the Rabbit R1--in New York City, and then, standing on a windy curb outside the venue, pressed a button on the device to summon an Uber home. Instead, after maybe an hour of getting it set up and fidgeting with it, the connection failed. The R1 is a bright-orange chunk of a device, with a camera, a mic, and a small screen. Press and hold its single button, ask it a question or give it a command using your voice, and the cute bouncing rabbit on screen will perk up its ears, then talk back to you.

At CES this January, startup Rabbit unveiled its first device, just in time for the end of the year of the rabbit according to the lunar calendar. It's a cute little orange square that was positioned as a "pocket companion that moves AI from words to action." In other words, it's basically a dedicated AI machine that acts kind of like a walkie talkie to a virtual assistant. You're probably thinking of the Humane AI Pin, which was announced last year and started shipping this month. I awarded it a score of 50 (out of 100) earlier this month, while outlets like Wired and The Verge gave it similarly low marks of 4 out of 10.

The pathway leading into Rabbit's venue--for the launch event of the R1, an artificial intelligence-powered device announced at CES 2024--was paved with gadgets from the past. First was the orange JVC Videosphere, then the Sony Walkman, a Tamagotchi, a transparent GameBoy Color, heck, even the original Pokédex toy from 1998. At the very end of the hall was Teenage Engineering's Pocket Operator, and across from it, a few concept prototypes of the Rabbit R1. If the Pocket Operator stands out, seeing as it's barely a decade old, that's because the Swedish design-firm Teenage Engineering helped design the R1. And at the launch event, CEO Jesse Lyu announced on stage that Jesper Kouthoofd, founder of Teenage Engineering, has joined Rabbit as its chief design officer (while still maintaining his role as CEO of TE).

At least, that was my takeaway after my first chat with the founder of Rabbit Inc., a new AI startup debuting a pocket-friendly device called the R1 at CES 2024. Instead of taking out your smartphone to complete some task, hunting for the right app, and then tapping around inside it, Lyu wants us to just ask the R1 via a push-to-talk button. Then a series of automated scripts called "rabbits" will carry out the task so you can go about your day. The R1 is a red-orange, square-ish device about the size of a stack of Post-It notes. It was designed in collaboration with the Swedish firm Teenage Engineering.

We focus on analyzing the classical stochastic projected gradient methods under a general dependent data sampling scheme for constrained smooth nonconvex optimization. We show the worst-case rate of convergence $\tilde{O}(t^{-1/4})$ and complexity $\tilde{O}(\varepsilon^{-4})$ for achieving an $\varepsilon$-near stationary point in terms of the norm of the gradient of Moreau envelope and gradient mapping. While classical convergence guarantee requires i.i.d. data sampling from the target distribution, we only require a mild mixing condition of the conditional distribution, which holds for a wide class of Markov chain sampling algorithms. This improves the existing complexity for the constrained smooth nonconvex optimization with dependent data from $\tilde{O}(\varepsilon^{-8})$ to $\tilde{O}(\varepsilon^{-4})$ with a significantly simpler analysis. We illustrate the generality of our approach by deriving convergence results with dependent data for stochastic proximal gradient methods, adaptive stochastic gradient algorithm AdaGrad and stochastic gradient algorithm with heavy ball momentum. As an application, we obtain first online nonnegative matrix factorization algorithms for dependent data based on stochastic projected gradient methods with adaptive step sizes and optimal rate of convergence.

No one wants to be falsely accused of saying or doing something that will destroy their reputation. Even more nightmarish is a scenario where, despite being innocent, the fabricated "evidence" against a person is so convincing that they are unable to save themselves. Yet thanks to a rapidly advancing type of artificial intelligence (AI) known as "deepfake" technology, our near-future society will be one where everyone is at great risk of having exactly that nightmare come true. Deepfakes -- or videos that have been altered to make a person's face or body appear to do something they did not in fact do -- are increasingly used to spread misinformation and smear their targets. Political, religious and business leaders are already expressing alarm by the viral spread of deepfakes that maligned prominent figures like former US President Donald Trump, Pope Francis and Twitter CEO Elon Musk.

In the third week of Russia's war in Ukraine, Volodymyr Zelensky appeared in a video, dressed in a dark green shirt, speaking slowly and deliberately while standing behind a white presidential podium featuring his country's coat of arms. Except for his head, the Ukrainian president's body barely moved as he spoke. His voice sounded distorted and almost gravelly as he appeared to tell Ukrainians to surrender to Russia. "I ask you to lay down your weapons and go back to your families," he appeared to say in Ukrainian in the clip, which was quickly identified as a deepfake. "This war is not worth dying for. I suggest you to keep on living, and I am going to do the same."