Generative Retrieval introduces a new approach to Information Retrieval by reframing it as a constrained generation task, leveraging recent advancements in Autoregressive (AR) language models. However, AR-based Generative Retrieval methods suffer from high inference latency and cost compared to traditional dense retrieval techniques, limiting their practical applicability. This paper investigates fully Non-autoregressive (NAR) language models as a more efficient alternative for generative retrieval. While standard NAR models alleviate latency and cost concerns, they exhibit a significant drop in retrieval performance (compared to AR models) due to their inability to capture dependencies between target tokens. To address this, we question the conventional choice of limiting the target token space to solely words or sub-words. We propose PIXAR, a novel approach that expands the target vocabulary of NAR models to include multi-word entities and common phrases (up to 5 million tokens), thereby reducing token dependencies. PIXAR employs inference optimization strategies to maintain low inference latency despite the significantly larger vocabulary. Our results demonstrate that PIXAR achieves a relative improvement of 31.0% in MRR@10 on MS MARCO and 23.2% in Hits@5 on Natural Questions compared to standard NAR models with similar latency and cost.

It had a great new idea for a movie--Elemental, based on characters from The Good Dinosaur's director Peter Sohn--but actually animating the film's titular elements was proving to be a problem. After all, it's one thing to draw a crumbling mound of sentient dirt, but how do you capture the ethereal nature of fire onscreen, and how would a corporeal body made of water even work? Can you see through it? Do the eyes just float around? While some of those questions could be answered with good old-fashioned suspension of disbelief, Pixar's animators thought the fire issue was a real conundrum, especially considering that one of their movie's leads, Ember, was actually supposed to be made of the stuff. They had tools to make a flame effect from years of previous animations, but when you actually tried to shape it into a character, the results were pretty terrifying, a cross between Studio Ghibli's Calcifer and Nicolas Cage's Ghost Rider, but somehow harsher.

Disney has a seemingly endless selection of Marvel movies and plenty of Star Wars and Pixar fare, too. Problem is, there's so much stuff, it's hard to know where to begin. WIRED is here to help. Below are our picks for the best movies on Disney right now. For more viewing ideas, try our guides to the best movies on Netflix and the best movies on Amazon Prime. This content can also be viewed on the site it originates from. Sam Raimi's sequel to 2016's Doctor Strange isn't the beloved director's first superhero movie, but it is his first foray into the Marvel Cinematic Universe style of making movies, which ultimately proves to be both a blessing and a curse. On the plus side, the movie is probably the closest thing the Marvel franchise has gotten to a straight-up horror film, and it's full of Raimi's signature practical effects (plus the ever-important Bruce Campbell cameo). Yet, because the MCU is such a box office powerhouse, the movie never goes full Raimi--which is understandable, but somewhat disappointing for fans of The Evil Dead maestro.

During Nvidia's Siggraph keynote the tech and GPU innovator dropped a lot of jargon, but I think we just glimpsed the future of the internet – and the metaverse. The likes of Pixar, as well as Adobe, Autodesk, Siemens and, yes, Nvidia, are pioneering new tech that will reshape how we use the internet; what they see as a '3D internet'. The future laid out by Nvidia and its partners is quite something, and even the most metaverse-averse may want to sit up and take note; I mean, when was the last time Pixar got it wrong? If you're asking, 'what is the metaverse?' then read our guide for the detail, but generally we're looking at a new way of interacting with the internet through VR, 3D spaces and connected worlds. To make this new 3D internet a real, working thing for everyone, Nvidia is developing its Avatar Cloud Engine (opens in new tab) (ACE), a new tool that enables everyone to build lifelike, 3D'humans' that can talk and interact with other computers (and you) in the metaverse.

What if technology enabled you to share a screen with Hugh Jackman? And no, we are not talking about a look-alike or virtual replica. Warner Bros.' upcoming movie Reminiscence starring Jackman has incorporated deepfake technology to turn a fan's photograph into a short video sequence with Jackman. Along with digital studio Oblio, the American film production company has partnered with Israeli-based synthetic media startup D-ID to use its'Live Portrait' product and create personalised experiences for movie fans. For the Reminiscence project, Warner Bros. has created an official website where users can enter their first names and upload a picture of themselves -- from the past or current.

As digital animators continue to push the boundaries of technology and creativity, the technical teams that support them are turning to artificial intelligence and machine learning to deliver the tools they need. That's the case at Pixar, where the company has made new machine learning breakthroughs it hopes will both improve quality and reduce costs. Vaibhav Vavilala, technical director at Pixar Animation Studios, has been leading some of those efforts after carefully studying recent scientific literature and monitoring the work at the R&D labs of parent company Disney. Vavilala said those advances have the potential to reduce Pixar's datacenter footprint by half for some stages of production. Vavilala made his remarks during a presentation at VentureBeat's Transform 2020 conference. In recent years, Vavilala has worked on the lighting and rendering optimization team on such Pixar movies as Coco, Incredibles 2, and Toy Story 4. But a couple of years ago, he began focusing on a concept called Deep Learned Super Resolution.

Regardless of what is happening around the world, the AI community are one productive bunch, and they have something interesting to share almost every day. Here's what is new this week: The short history of deep learning indicates the incredible effectiveness of infinitely wide networks. Insights from these infinitely wide networks can be used as a lens to study deep learning. However, implementing infinite-width models in an efficient and scalable way requires significant engineering proficiency. To address these challenges and accelerate theoretical progress in deep learning, Google's AI team released Neural Tangents, a new open-source software library written in JAX.

Regardless of what is happening around the world, the AI community are one productive bunch, and they have something interesting to share almost every day. Here's what is new this week: The short history of deep learning indicates the incredible effectiveness of infinitely wide networks. Insights from these infinitely wide networks can be used as a lens to study deep learning. However, implementing infinite-width models in an efficient and scalable way requires significant engineering proficiency. To address these challenges and accelerate theoretical progress in deep learning, Google's AI team released Neural Tangents, a new open-source software library written in JAX.

Companies will finally gather their data, run it through several algorithms and scenario planning, and leverage the results to invest in a handful of ideas that will demonstrate the power and potential of AI. Some of that is indeed happening, as forward-leaning companies continue to be drawn to creative and intuitive thinking that will make their programs, planning and go-to-market promotions smarter while delivering stronger ROI. Still, I can't help but think that the best year ever for AI was actually 2004. Now I know you're thinking, 'AI was really only a whisper of an idea 14 years ago. What could possibly have transpired to make 2004 a banner year for artificial intelligence?'

Artificial intelligence (AI) has generated enormous hype in recent months, but opinions about the real impact of technologies like machine translation, autonomous vehicles, natural language processing, and computer vision vary widely. Optimists envision a world in which "driver" becomes a superfluous word; pessimists worry about the millions of truckers at risk of unemployment. But while often described with the futuristic fancy of Skynet in the "Terminator" movies, AI has very real near-term applications for the entertainment industry. Here are some of the most exciting examples. AI is poised to touch every facet of entertainment, even that most time-honored of parental traditions: reading stories to children before bed.