This colossal sci-fi thriller from Gareth Edwards features John David Washington and Gemma Chan in vast mysterious panoramas and vertiginous vistas which deserve to be shown at Imax-plus scale; it also shows that Christopher Nolan isn't the only British director in Hollywood thinking (and acting) big. After a stint making franchise movies such as Godzilla and the enjoyable and underrated Rogue One: A Star Wars Story, Edwards has now crafted this ambitious original picture, co-written with Chris Weitz, which is closer in spirit to his ingenious 2010 debut Monsters. The Creator is an old-fashioned science-fiction actioner with some ideas to match to state-of-the-art digital effects, in the tradition of Ridley Scott's Blade Runner or Neill Blomkamp's District 9, with a creeping colonialist's fear of the unknown to match that in Coppola's Apocalypse Now. And given that Edwards has served some time aboard the Star Wars mother ship, it shouldn't be too surprising to find some holograms in the mix and a certain dustbin-sized droid which whimpers something poignant about what an honour it's been to serve his comrades before lumbering out to face the enemy on a kamikaze mission. Washington shows us some more of that distinctive self-possession and even slight hauteur as a performer, in playing Josh, a US army special forces undercover officer, fighting a strange, dirty war in a postnuclear world upended by the dominance of artificial intelligence.

A strange thing happened last week when you searched for "tank man" on Google. Tap on image results and instead of the usual photos of Tiananmen Square in Beijing, and the iconic image of a brave protester staring down a convoy of tanks that was captured in 1989, the first result was the same historic moment – but from a different point of view. For a time last week, the first result on Google Images for "tank man" was instead an AI-generated image of the same protester, taking a selfie in front of the tank. The image was created by Midjourney, and was at least six months old. First reported by 404 Media, a new tech journalism startup set up by former Vice News staff, the emergence of the tank man selfie – which Google subsequently removed from search results for "tank man" – highlighted one of the main fears that Eddie Perez, Twitter's former head of election integrity, highlighted to me in a recent podcast interview: it's now possible, with the use of AI imagery, to create alternative history. And that has huge ramifications not only on our lives, but also our elections.

SAG-AFTRA members have voted in favor of a strike authorization for performers working in video games, including those who do voice work, motion capture and stunts for the industry. Of the 34,687 members who cast their ballot, 98.32 percent voted "yes." It doesn't automatically mean that the performers are going on strike, only that they could after September 26. The guild explained that it's been negotiating for a new contract with video game companies since 2022, but that they "have refused to agree to critical terms [its] members need." For its next bargaining sessions on September 26, 27 and 28, it believes the strike authorization gives it the added leverage needed to get the companies to agree to its terms.

Visual effects commonly requires both the creation of realistic synthetic humans as well as retargeting actors' performances to humanoid characters such as aliens and monsters. Achieving the expressive performances demanded in entertainment requires manipulating complex models with hundreds of parameters. Full creative control requires the freedom to make edits at any stage of the production, which prohibits the use of a fully automatic ``black box'' solution with uninterpretable parameters. On the other hand, producing realistic animation with these sophisticated models is difficult and laborious. This paper describes FDLS (Facial Deep Learning Solver), which is Weta Digital's solution to these challenges. FDLS adopts a coarse-to-fine and human-in-the-loop strategy, allowing a solved performance to be verified and edited at several stages in the solving process. To train FDLS, we first transform the raw motion-captured data into robust graph features. Secondly, based on the observation that the artists typically finalize the jaw pass animation before proceeding to finer detail, we solve for the jaw motion first and predict fine expressions with region-based networks conditioned on the jaw position. Finally, artists can optionally invoke a non-linear finetuning process on top of the FDLS solution to follow the motion-captured virtual markers as closely as possible. FDLS supports editing if needed to improve the results of the deep learning solution and it can handle small daily changes in the actor's face shape. FDLS permits reliable and production-quality performance solving with minimal training and little or no manual effort in many cases, while also allowing the solve to be guided and edited in unusual and difficult cases. The system has been under development for several years and has been used in major movies.

Text-guided diffusion models such as DALLE-2, Imagen, eDiff-I, and Stable Diffusion are able to generate an effectively endless variety of images given only a short text prompt describing the desired image content. In many cases the images are of very high quality. However, these models often struggle to compose scenes containing several key objects such as characters in specified positional relationships. The missing capability to ``direct'' the placement of characters and objects both within and across images is crucial in storytelling, as recognized in the literature on film and animation theory. In this work, we take a particularly straightforward approach to providing the needed direction. Drawing on the observation that the cross-attention maps for prompt words reflect the spatial layout of objects denoted by those words, we introduce an optimization objective that produces ``activation'' at desired positions in these cross-attention maps. The resulting approach is a step toward generalizing the applicability of text-guided diffusion models beyond single images to collections of related images, as in storybooks. Directed Diffusion provides easy high-level positional control over multiple objects, while making use of an existing pre-trained model and maintaining a coherent blend between the positioned objects and the background. Moreover, it requires only a few lines to implement.

We investigate the internal behavior of Transformer-based Large Language Models (LLMs) when they generate factually incorrect text. We propose modeling factual queries as Constraint Satisfaction Problems and use this framework to investigate how the model interacts internally with factual constraints. Specifically, we discover a strong positive relation between the model's attention to constraint tokens and the factual accuracy of its responses. In our curated suite of 11 datasets with over 40,000 prompts, we study the task of predicting factual errors with the Llama-2 family across all scales (7B, 13B, 70B). We propose SAT Probe, a method probing self-attention patterns, that can predict constraint satisfaction and factual errors, and allows early error identification. The approach and findings demonstrate how using the mechanistic understanding of factuality in LLMs can enhance reliability.

Recent years have witnessed remarkable progress made in large language models (LLMs). Such advancements, while garnering significant attention, have concurrently elicited various concerns. The potential of these models is undeniably vast; however, they may yield texts that are imprecise, misleading, or even detrimental. Consequently, it becomes paramount to employ alignment techniques to ensure these models to exhibit behaviors consistent with human values. This survey endeavors to furnish an extensive exploration of alignment methodologies designed for LLMs, in conjunction with the extant capability research in this domain. Adopting the lens of AI alignment, we categorize the prevailing methods and emergent proposals for the alignment of LLMs into outer and inner alignment. We also probe into salient issues including the models' interpretability, and potential vulnerabilities to adversarial attacks. To assess LLM alignment, we present a wide variety of benchmarks and evaluation methodologies. After discussing the state of alignment research for LLMs, we finally cast a vision toward the future, contemplating the promising avenues of research that lie ahead. Our aspiration for this survey extends beyond merely spurring research interests in this realm. We also envision bridging the gap between the AI alignment research community and the researchers engrossed in the capability exploration of LLMs for both capable and safe LLMs.

Despite significant advancements in deep learning for vision and natural language, unsupervised domain adaptation in audio remains relatively unexplored. We, in part, attribute this to the lack of an appropriate benchmark dataset. To address this gap, we present Synthia's melody, a novel audio data generation framework capable of simulating an infinite variety of 4-second melodies with user-specified confounding structures characterised by musical keys, timbre, and loudness. Unlike existing datasets collected under observational settings, Synthia's melody is free of unobserved biases, ensuring the reproducibility and comparability of experiments. To showcase its utility, we generate two types of distribution shifts-domain shift and sample selection bias-and evaluate the performance of acoustic deep learning models under these shifts. Our evaluations reveal that Synthia's melody provides a robust testbed for examining the susceptibility of these models to varying levels of distribution shift.

The rise of social media has enabled the widespread propagation of fake news, text that is published with an intent to spread misinformation and sway beliefs. Rapidly detecting fake news, especially as new events arise, is important to prevent misinformation. While prior works have tackled this problem using supervised learning systems, automatedly modeling the complexities of the social media landscape that enables the spread of fake news is challenging. On the contrary, having humans fact check all news is not scalable. Thus, in this paper, we propose to approach this problem interactively, where humans can interact to help an automated system learn a better social media representation quality. On real world events, our experiments show performance improvements in detecting factuality of news sources, even after few human interactions.

Knowledge graph completion is a task that revolves around filling in missing triples based on the information available in a knowledge graph. Among the current studies, text-based methods complete the task by utilizing textual descriptions of triples. However, this modeling approach may encounter limitations, particularly when the description fails to accurately and adequately express the intended meaning. To overcome these challenges, we propose the augmentation of data through two additional mechanisms. Firstly, we employ ChatGPT as an external knowledge base to generate coherent descriptions to bridge the semantic gap between the queries and answers. Secondly, we leverage inverse relations to create a symmetric graph, thereby creating extra labeling and providing supplementary information for link prediction. This approach offers additional insights into the relationships between entities. Through these efforts, we have observed significant improvements in knowledge graph completion, as these mechanisms enhance the richness and diversity of the available data, leading to more accurate results.