Breakthroughs, discoveries, and DIY tips sent every weekday. The interstellar comet 3I/ATLAS has captured the imaginations of both amateur and professional skygazers, but it's not the only icy space rock to recently speed past Earth. In October, a pair of comets known as Lemmon and SWAN also left trails of dust and gas as they continued along their vast orbits through the solar system. As luck had it, their timing perfectly aligned with a wave of vibrant auroras generated by one of this year's largest solar eruptions. And judging from NASA's recently released photos, few people had a better vantage point than the astronauts aboard the International Space Station.

There Is Only One AI Company. As Nvidia, OpenAI, Google, and Microsoft forge partnerships and deals, the AI industry is looking more like one interconnected machine. What does that mean for all of us? It all began, as many things do, with Elon Musk . In the early 2010s he realized that AI was on a track to become perhaps the most powerful technology of all time.

Recent work on deriving $O(\log T)$ anytime regret bounds for stochastic dueling bandit problems has considered mostly Condorcet winners, which do not always exist, and more recently, winners defined by the Copeland set, which do always exist. In this work, we consider a broad notion of winners defined by tournament solutions in social choice theory, which include the Copeland set as a special case but also include several other notions of winners such as the top cycle, uncovered set, and Banks set, and which, like the Copeland set, always exist. We develop a family of UCB-style dueling bandit algorithms for such general tournament solutions, and show $O(\log T)$ anytime regret bounds for them. Experiments confirm the ability of our algorithms to achieve low regret relative to the target winning set of interest.

Fake AI apps disguised as ChatGPT and DALL·E clones are flooding app stores with sophisticated spyware that steals personal data and monitors users.

Data augmentation is a ubiquitous technique for increasing the size of labeled training sets by leveraging task-specific data transformations that preserve class labels.

For example, replicating the straggling task on another available node is a common approach to deal with stragglers (e.g., [

We study fairness in collaborative-filtering recommender systems, which are sensitive to discrimination that exists in historical data.

Multi-armed bandit is a useful tool to model the process of incrementally collecting data for multiple objects in a decision space.

We instead rely on weak supervision sources having some structure by virtue of being encoded programmatically.

We study how the word usage of U.S. Congressional speeches varies across states and party affiliation, how words are used differently across sections of the ArXiv, and how the co-purchase patterns of groceries can vary across seasons.