MAROKO133 Hot ai: Adobe Research Unlocking Long-Term Memory in Video World Models with Sta

📌 MAROKO133 Breaking ai: Adobe Research Unlocking Long-Term Memory in Video World

Video world models, which predict future frames conditioned on actions, hold immense promise for artificial intelligence, enabling agents to plan and reason in dynamic environments. Recent advancements, particularly with video diffusion models, have shown impressive capabilities in generating realistic future sequences. However, a significant bottleneck remains: maintaining long-term memory. Current models struggle to remember events and states from far in the past due to the high computational cost associated with processing extended sequences using traditional attention layers. This limits their ability to perform complex tasks requiring sustained understanding of a scene.

A new paper, “Long-Context State-Space Video World Models” by researchers from Stanford University, Princeton University, and Adobe Research, proposes an innovative solution to this challenge. They introduce a novel architecture that leverages State-Space Models (SSMs) to extend temporal memory without sacrificing computational efficiency.

The core problem lies in the quadratic computational complexity of attention mechanisms with respect to sequence length. As the video context grows, the resources required for attention layers explode, making long-term memory impractical for real-world applications. This means that after a certain number of frames, the model effectively “forgets” earlier events, hindering its performance on tasks that demand long-range coherence or reasoning over extended periods.

The authors’ key insight is to leverage the inherent strengths of State-Space Models (SSMs) for causal sequence modeling. Unlike previous attempts that retrofitted SSMs for non-causal vision tasks, this work fully exploits their advantages in processing sequences efficiently.

The proposed Long-Context State-Space Video World Model (LSSVWM) incorporates several crucial design choices:

1. Block-wise SSM Scanning Scheme: This is central to their design. Instead of processing the entire video sequence with a single SSM scan, they employ a block-wise scheme. This strategically trades off some spatial consistency (within a block) for significantly extended temporal memory. By breaking down the long sequence into manageable blocks, they can maintain a compressed “state” that carries information across blocks, effectively extending the model’s memory horizon.
2. Dense Local Attention: To compensate for the potential loss of spatial coherence introduced by the block-wise SSM scanning, the model incorporates dense local attention. This ensures that consecutive frames within and across blocks maintain strong relationships, preserving the fine-grained details and consistency necessary for realistic video generation. This dual approach of global (SSM) and local (attention) processing allows them to achieve both long-term memory and local fidelity.

The paper also introduces two key training strategies to further improve long-context performance:

• Diffusion Forcing: This technique encourages the model to generate frames conditioned on a prefix of the input, effectively forcing it to learn to maintain consistency over longer durations. By sometimes not sampling a prefix and keeping all tokens noised, the training becomes equivalent to diffusion forcing, which is highlighted as a special case of long-context training where the prefix length is zero. This pushes the model to generate coherent sequences even from minimal initial context.
• Frame Local Attention: For faster training and sampling, the authors implemented a “frame local attention” mechanism. This utilizes FlexAttention to achieve significant speedups compared to a fully causal mask. By grouping frames into chunks (e.g., chunks of 5 with a frame window size of 10), frames within a chunk maintain bidirectionality while also attending to frames in the previous chunk. This allows for an effective receptive field while optimizing computational load.

The researchers evaluated their LSSVWM on challenging datasets, including Memory Maze and Minecraft, which are specifically designed to test long-term memory capabilities through spatial retrieval and reasoning tasks.

The experiments demonstrate that their approach substantially surpasses baselines in preserving long-range memory. Qualitative results, as shown in supplementary figures (e.g., S1, S2, S3), illustrate that LSSVWM can generate more coherent and accurate sequences over extended periods compared to models relying solely on causal attention or even Mamba2 without frame local attention. For instance, on reasoning tasks for the maze dataset, their model maintains better consistency and accuracy over long horizons. Similarly, for retrieval tasks, LSSVWM shows improved ability to recall and utilize information from distant past frames. Crucially, these improvements are achieved while maintaining practical inference speeds, making the models suitable for interactive applications.

The Paper Long-Context State-Space Video World Models is on arXiv

The post Adobe Research Unlocking Long-Term Memory in Video World Models with State-Space Models first appeared on Synced.

🔗 Sumber: syncedreview.com

📌 MAROKO133 Hot ai: OpenAI Fumbled Its $1 Billion Deal With Disney Wajib Baca

On Tuesday, news emerged that OpenAI was canning its chaotic and controversial text-to-video AI app, Sora.

The app, which was announced to much fanfare late last year, has struggled to retain users, with downloads plummeting. The stream of uninspired AI slop wasn’t just and rife with copyright-infringing material— the costs to keep it running were also enormous, even for a company the size of OpenAI.

In short, we can think of a litany of reasons why OpenAI may have chosen to ditch its distasteful slop generator, especially now that investors are starting to ask questions ahead of the company’s rumored IPO later this year.

Caught up in the multi-billion-dollar misstep is entertainment conglomerate Disney, which signed what was supposed to be a groundbreaking $1 billion deal with OpenAI in December. As a source familiar with the matter told The Hollywood Reporter, the media giant has now pulled out of the licensing agreement, which would’ve allowed users to generate AI videos using more than 200 Disney, Marvel, Star Wars, and Pixar characters.

“As the nascent AI field advances rapidly, we respect OpenAI’s decision to exit the video generation business and to shift its priorities elsewhere,” a Disney spokesperson told the publication. “We appreciate the constructive collaboration between our teams and what we learned from it, and we will continue to engage with AI platforms to find new ways to meet fans where they are while responsibly embracing new technologies that respect IP and the rights of creators.”

The abrupt end to the blockbuster deal certainly raises questions. Did OpenAI’s discontinuation of Sora force Disney out of the deal, or did Disney get cold feet first? Given the tone of its spokesperson’s statement, the former explanation appears to be more likely.

Reading the tea leaves for Disney’s AI future is challenging. Former Disney CEO Bob Iger, who was replaced by the company’s new leader and longtime Disney exec Josh D’Amaro last week, made it clear that the company was looking to shoehorn AI-generated content into its Disney+ experience.

How D’Amaro feels about the initiative remains to be seen. In an introductory memo, obtained by Business Insider, D’Amaro said he was looking to embrace tech to “help us create more immersive, interactive, and personal ways for people to experience Disney.”

While OpenAI isn’t pulling out of AI video entirely, as the Reporter points out, shutting its flashy app a mere five months into its chaotic existence speaks volumes about the state of the industry and its attempts to bring consumer-facing AI tools to market.

“We’re saying goodbye to Sora,” OpenAI’s Sora team wrote in an official announcement on X. “To everyone who created with Sora, shared it, and built community around it: thank you. What you made with Sora mattered, and we know this news is disappointing.”

“We’ll share more soon, including timelines for the app and API and details on preserving your work,” the statement reads.

More on Sora: OpenAI Is Reportedly Killing Its Disastrous Video AI Slop App

The post OpenAI Fumbled Its $1 Billion Deal With Disney appeared first on Futurism.

🔗 Sumber: futurism.com