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

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

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: US firm’s thorium nuclear fuel bundles move to manufacturing f

US-based Clean Core Thorium Energy (CCTE) and Canadian Nuclear Laboratories (CNL) have announced an agreement to manufacture demonstration irradiation bundles of ANEEL fuel. 

These bundles are full-scale prototypes designed to match the physical specifications of fuel that are currently used in commercial reactors for the purpose of interface and irradiation testing.

ANEEL fuel is a technology that combines thorium with high-assay low-enriched uranium (HALEU). This uranium component is enriched to between 5% and 20% of the uranium-235 isotope. In comparison, standard fuel for pressurized heavy-water reactors generally contains less than 0.72% uranium-235. 

The fuel maintains the same external geometry as existing 19-element and 37-element fuel designs, which allows it to be integrated into current CANDU and pressurized heavy-water reactor systems without hardware or core design modifications. 

It is designed to achieve higher burnup and improved fuel utilization in a once-through fuel cycle. 

This design has the potential to reduce the total volume of long-lived spent fuel produced per unit of energy generated and to increase resistance to nuclear proliferation.

Supporting future commercial reactor fleet

The manufacturing process will be managed by CNL at the Chalk River Laboratories in Ontario. CNL will handle the development, qualification, and fabrication of the ANEEL fuel in accordance with Canadian Standard Association requirements. 

The resulting data from the irradiation of these bundles will provide in-reactor information to support the future qualification and deployment of the fuel in the commercial reactor fleet. 

This manufacturing stage follows the completion of a collaborative project under the Canadian Nuclear Research Initiative which validated the computer codes and analysis methods used for design and safety analysis. 

The demonstration irradiation is intended to provide data to the designer, the utility, the manufacturer, and the regulator regarding fuel performance.

Other components of the development program include irradiation testing at the Idaho National Laboratory. In 2025, results from the Advanced Test Reactor indicated that ANEEL fuel pellets maintained their structural integrity after several months of testing.

The program is targeting burnup levels that exceed 60 GWd/t/MTU. CCTE has also completed a Phase 1 pre-licensing vendor design review with the Canadian Nuclear Safety Commission.

To provide necessary performance data

Mehul Shah, CEO of Clean Core Thorium Energy, stated that the fabrication of reactor-representative bundles will provide the performance data necessary for regulatory review and fleet deployment. 

“This milestone advances ANEEL from validated models to an engineered, fleet-compatible fuel design,” said Shah.

Dr. Monica Regalbuto, Vice-President of Science and Technology at CNL, stated that the laboratories are using their fuel research and fabrication capabilities for the manufacturing of this fuel type. 

“We are pleased to be working with CCTE and to apply the deep expertise of the labs to this exciting advanced fuel project,” she concluded.  

These activities are intended to support the commercial readiness of the fuel for use in existing heavy-water reactor designs.

🔗 Sumber: interestingengineering.com