The semantics of emoji has, to date, been considered from a static perspective. We offer the first longitudinal study of how emoji semantics changes over time, applying techniques from computational linguistics to six years of Twitter data. We identify five patterns in emoji semantic development and find evidence that the less abstract an emoji is, the more likely it is to undergo semantic change. In addition, we analyse select emoji in more detail, examining the effect of seasonal-ity and world events on emoji semantics. To aid future work on emoji and semantics, we make our data publicly available along with a web-based interface that anyone can use to explore semantic change in emoji.

Estimating the tangent spaces of a data manifold is a fundamental problem in data analysis. The standard approach, Local Principal Component Analysis (LPCA), struggles in high-noise settings due to a critical trade-off in choosing the neighborhood size. Selecting an optimal size requires prior knowledge of the geometric and noise characteristics of the data that are often unavailable. In this paper, we propose a spectral method, Laplacian Eigenvector Gradient Orthogonalization (LEGO), that utilizes the global structure of the data to guide local tangent space estimation. Instead of relying solely on local neighborhoods, LEGO estimates the tangent space at each data point by orthogonalizing the gradients of low-frequency eigenvectors of the graph Laplacian. We provide two theoretical justifications of our method. First, a differential geometric analysis on a tubular neighborhood of a manifold shows that gradients of the low-frequency Laplacian eigenfunctions of the tube align closely with the manifold's tangent bundle, while an eigenfunction with high gradient in directions orthogonal to the manifold lie deeper in the spectrum. Second, a random matrix theoretic analysis also demonstrates that low-frequency eigenvectors are robust to sub-Gaussian noise. Through comprehensive experiments, we demonstrate that LEGO yields tangent space estimates that are significantly more robust to noise than those from LPCA, resulting in marked improvements in downstream tasks such as manifold learning, boundary detection, and local intrinsic dimension estimation.

As any adult who loves video games knows, there are simply too many of them – 19,000 games were released in 2024 on PC games storefront Steam alone, not counting all the playable delights on consoles and smartphones. Most of us have backlogs of unplayed classics that make us feel guilty about buying newer games. Finding things that are actually good, meanwhile, can feel totally impossible. At least 50% of the questions people send in for this newsletter are a variant of "Help, what should I play?" We do our best to help, but even though it's my job to know about games, I still don't have infinite time to play them.

Armed with drones, long-distance camera lenses and metal detectors, a hillside in Eaton Canyon has become the focus of intense scrutiny over the last month by teams of private investigators now seeking clues on whether Southern California Edison equipment caused the massive fire that destroyed large swaths of Altadena. Some of the findings and theories of these privately hired teams of fire investigators and electrical engineers have emerged in more than 40 lawsuits that residents have filed against the utility. Much of the focus has been centered on a group of transmission towers where the first flames were seen just as the Eaton fire exploded. Earlier this week, a new lawsuit alleged that an idle transmission tower on the hillside -- one that has not been in use for more than 50 years -- might have sparked the devastating blaze. With more than 9,000 homes lost and 17 people killed, liability is going to be a costly question that could affect how Altadena is rebuilt.

Scientists around the world are gearing up to study the first images taken by the James Webb Space Telescope, which are to be released on July 12. Some astronomers will be running machine-learning algorithms on the data to detect and classify galaxies in deep space at a level of detail never seen before. Brant Robertson, an astrophysics professor at the University of California, Santa Cruz, in the US believes the telescope's snaps will lead to breakthroughs that will help us better understand how the universe formed some 13.7 billion years ago. "The JWST data is exciting because it gives us an unprecedented window on the infrared universe, with a resolution that we've only dreamed about until now," he told The Register. Robertson helped develop Morpheus, a machine-learning model trained to pore over pixels and pick out blurry blob-shaped objects from the deep abyss of space and determine whether these structures are galaxies or not, and if so, of what type.

A prime candidate for secular canonisation – and a personal hero of mine – is Frances Gabe. She was a visionary, a terrible neighbour (she antagonised hers with a succession of snarling great danes and a penchant for nude DIY) and the inventor of the self-cleaning home. Gabe, who died in 2016 at 101, transformed her Oregon bungalow into a "giant dishwasher", with a system of sprinklers, air dryers and drains, plus self-cleaning sinks, bath and toilet. "Housework is a thankless, unending job," Gabe said. I agree with Gabe – and with Lenin, who condemned housework as "barbarously unproductive, petty, nerve-racking, stultifying and crushing drudgery".

The main objective of this research is to increase the quality of AI used in commercial RTS games, which has seen little improvement over the past decade. This objective will be addressed by investigating the use of a learning by observation, case-based reasoning agent, which can be applied to new RTS games with minimal development effort. To be successful, this agent must compare favourably with standard commercial RTS AI techniques: it must be easier to apply, have reasonable resource requirements, and produce a better player. Currently, a prototype implementation has been produced for the game StarCraft, and it has demonstrated the need for processing large sets of input data into a more concise form for use at run-time.

Interactive narratives are branching stories with events that change based on feedback from participants. One method of generating interactive narratives is a plan-based process called mediation. A sub-process within mediation called accommodation creates new story content when a participant deviates from the main storyline. We show that a model of character knowledge allows accommodation to find a novel class of branching stories previously inaccessible by the algorithm.

An open area of research for AI in games is how to provide unique gameplay experiences that present specialized game content to users based on their preferences, in-game actions, or the system's goals. The area of procedural content generation (PCG) focuses on creating or modifying game worlds, assets, and mechanics to generate tailored or personalized game experiences. Similarly, the area of interactive narrative (IN) focuses on creating or modifying story worlds, events, and domains to generate tailored or personalized story experiences. In this paper we describe a game engine that utilizes a PCG pipeline to generate and control a range of gameplay experiences from an underlying IN experience management construct.

Strong story interactive narratives (IN) are stories that branch based on participant actions where all branches conform to a set of predefined constraints. However, participants in these systems may create branches where the constraints no longer hold. Strong story experience management, the process of generating IN trees, can be viewed as a game where the experience management agent wins if the story constraints hold during gameplay and loses if they are broken. In domains where the player has incomplete information of the story world, the experience manager can take action by shifting the player between alternate states that are consistent with the player's observations in order to maximize the probability that constraints will hold. This process is called superposition manipulation. In this paper we present a method of estimating the number of goal states reachable from different states in order to make informed decisions during superposition manipulation.