Transformer architectures have become very popular yet the original implementation requires $O(L^2)$ in serial time and memory as functions of input length $L$. Recent works proposed various linear self-attention mechanisms, scaling only as $O(L)$ for serial computation. We conduct a thorough complexity analysis of Performers, a class which includes most recent linear Transformer mechanisms. We note a remarkable computational flexibility: the gradient computation can be performed with no approximations using sublinear memory as a function of $L$ (in addition to negligible storage for the input sequence), at a cost of greater time complexity in the parallel setting. In the extreme case, a Performer consumes only $O(1)$ memory, and still requires $O(L)$ time. Due to complete backward-compatibility, this discovered time-memory tradeoff can be used for fine-tuning on low-memory devices in a decentralized fashion without any server computations.

A bill focused on how someone's name or likeness can be used after their death was also passed. New York is taking steps to regulate the use of AI in the state's entertainment industry. NY State Governor Kathy Hochul passed two pieces of legislation on Thursday that forces certain productions to disclose the use of AI-generated performers, and defines rules around how someone's likeness can be used after their death. Assembly Bill A8887B, now known as S.8420-A, specifically covers the use of AI performers in advertisements. Per Hochul's announcement, the law requires persons who produce or create an advertisement to identify if it includes AI generated synthetic performers.

A video game at the centre of a debate over artificial intelligence (AI) is in the running for the top prize at next year's Bafta Game Awards. Arc Raiders, from Swedish developer Embark Studios, has been a smash-hit since its October launch, selling more than four million copies. But the multiplayer shooter has been criticised for using text-to-speech tools to create additional lines, based on dialogue previously recorded by the game's actors. It is one of 10 titles longlisted for the prestigious best game award, with a shortlist to be announced in the run-up to April's annual ceremony. Other games up for the top prize include blockbusters Ghost of Yōtei and Death Stranding 2, indie games Hollow Knight: Silksong and Hades II, and indie adventure Blue Prince.

Assessment of medication history-taking has traditionally relied on human observation, limiting scalability and detailed performance data. While Generative AI (GenAI) platforms enable extensive data collection and learning analytics provide powerful methods for analyzing educational traces, these approaches remain largely underexplored in pharmacy clinical training. This study addresses this gap by applying learning analytics to understand how students develop clinical communication competencies with GenAI-powered virtual patients -- a crucial endeavor given the diversity of student cohorts, varying language backgrounds, and the limited opportunities for individualized feedback in traditional training settings. We analyzed 323 students' interaction logs across Australian and Malaysian institutions, comprising 50,871 coded utterances from 1,487 student-GenAI dialogues. Combining Epistemic Network Analysis to model inquiry co-occurrences with Sequential Pattern Mining to capture temporal sequences, we found that high performers demonstrated strategic deployment of information recognition behaviors. Specifically, high performers centered inquiry on recognizing clinically relevant information, integrating rapport-building and structural organization, while low performers remained in routine question-verification loops. Demographic factors including first-language background, prior pharmacy work experience, and institutional context, also shaped distinct inquiry patterns. These findings reveal inquiry patterns that may indicate clinical reasoning development in GenAI-assisted contexts, providing methodological insights for health professions education assessment and informing adaptive GenAI system design that supports diverse learning pathways.

The sparse part covers the noise, while the low-rank part recovers the principle components. However, it still requires too many iterations to be used in each training step. Scatterbrain yields an unbiased estimate of the attention matrix, and we can also understand how its variance changes. On the other hand, Scatterbrain's generality allows it This is similar in spirit to low-rank attention (Linformer) and global tokens, but it is not a low-rank approximation due to the non-linearity between the two attention steps. LSH has been used in estimation problem as well [12, 11].

To better understand this trade-off, we observe that sparse and low-rank approximations excel in different regimes, determined by the softmax temperature in attention, and sparse + low-rank can outperform each individually.

We conduct a thorough complexity analysis of Performers, a class which includes most recent linear Transformer mechanisms.

Interfaces for contemporary large language, generative media, and perception AI models are often engineered for single user interaction. We investigate ritual as a design scaffold for developing collaborative, multi-user human-AI engagement. We consider the specific case of an immersive staging of the musical Xanadu performed at UCLA in Spring 2025. During a two-week run, over five hundred audience members contributed sketches and jazzercise moves that vision language models translated to virtual scenery elements and from choreographic prompts. This paper discusses four facets of interaction-as-ritual within the show: audience input as offerings that AI transforms into components of the ritual; performers as ritual guides, demonstrating how to interact with technology and sorting audience members into cohorts; AI systems as instruments "played" by the humans, in which sensing, generative components, and stagecraft create systems that can be mastered over time; and reciprocity of interaction, in which the show's AI machinery guides human behavior as well as being guided by humans, completing a human-AI feedback loop that visibly reshapes the virtual world. Ritual served as a frame for integrating linear narrative, character identity, music and interaction. The production explored how AI systems can support group creativity and play, addressing a critical gap in prevailing single user AI design paradigms.

Amazon's Used Over 350 AI Shots in Season 2. Its Creator Isn't Sorry The show, which follows David's ascent to King of Israel, used four times as much AI this season, including for many of its battle scenes. A dusty visual overlay partially obscures crowds of men in the desert, sword-fighting in armor and on horseback. With some wardrobe tweaks, this scene could look like something out of or . But showrunner Jon Erwin says he didn't have the budget to bring these scenes to life. Instead, he used AI .