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

📌 MAROKO133 Eksklusif 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: Atlantic shock: US mysteriously test fires 130,000-pound nucle

A nuclear-powered submarine from either the United States or the United Kingdom appears to have test-fired a UGM-133 Trident II D5 submarine-launched ballistic missile (SLBM) from Puerto Rico in the North Atlantic on Sunday night, according to various videos circulating online.

The missile carried an “unusual payload” or hypersonic glide vehicle and was launched deep inside the Atlantic Ocean. 

However, neither the US Navy nor Britain’s Royal Navy, the only operators of the Trident II D5, has publicly confirmed the test.

Trident II D5 nuclear missile

A navigation warning was issued for the region between September 17 and 22, covering the same area where the launch was detected. 

Videos posted on social media showed a bright object streaking across the night sky at about 11:25 pm UTC, with several clips recorded over Puerto Rico. 

The Caribbean Astronomy Society described the event as a “military test” but did not specify which nation was responsible.

According to the US Navy, the SLBM Trident II D5 strategic weapon system (SWS) is the planet’s most lethal, accurate, and reliable sea-based strategic deterrent.

The UGM-133 Trident II D5 is the backbone of the sea-based leg of both the US and UK nuclear deterrents. 

It is deployed on 14 Ohio-class submarines in the US Navy and four Vanguard-class submarines in Britain under the 1963 Polaris Sales Agreement. 

Each Ohio-class boat carries 20 Trident II missiles, while the Vanguard-class carries 16.

With a range of about 4,600 miles (7,360 kilometers), the three-stage, solid-fuel missile can deliver up to 12 independently targetable warheads.

However, arms control agreements limit current loadings to four or five. 

Built by Lockheed Martin, each missile costs about $30.9 million. It is 44 feet (13.4 meters) long, 83 inches (2.1 meters) in diameter, and weighs roughly 130,000 pounds (58,500 kilograms).

Future upgrades

The US Navy has begun modernizing its Trident arsenal to ensure its reliability through the 2080s. 

In January, Lockheed Martin announced a $383 million contract to develop the Trident II D5 Life Extension 2 (D5LE2) program, which will equip the Navy’s new Columbia-class ballistic missile submarines.

The Pentagon’s 2022 Nuclear Posture Review urged accelerated investment in the D5LE2 to maintain the credibility of America’s strategic weapons system as Ohio-class submarines retire beginning in the 2040s. 

The same year, a bipartisan Strategic Posture Commission report highlighted the importance of upgrading the system to deter rival nuclear powers.

Britain’s Dreadnought-class submarines, due to replace the Vanguard fleet in the early 2030s, are likewise expected to carry the upgraded missile.

Officials in both countries say modernization of their undersea deterrents is critical as adversaries develop new anti-ship and missile defense systems. 

The sea-based leg of the nuclear triad provides stealth, mobility, and assured second-strike capability, a cornerstone of allied nuclear strategy for more than four decades.

As of Sunday night, neither Washington nor London had released details about the September 21 launch, leaving speculation over whether it marked the first flight of a modified Trident II with a hypersonic glide vehicle or an unpublicized routine test.

🔗 Sumber: interestingengineering.com