MAROKO133 Update ai: Archaeologists just found the largest and most advanced Medieval carg

📌 MAROKO133 Breaking ai: Archaeologists just found the largest and most advanced M

The Viking Ship Museum in Denmark recently announced an unprecedented discovery in the Øresund Strait: the world’s largest and most advanced medieval cargo ship ever found.

Hailed as “a milestone in maritime archaeology,” the discovery occurred while divers were investigating the seabed in the Sound, in anticipation of Copenhagen’s new Lynetteholm district, and stumbled upon a record-breaking cog buried beneath centuries of sand and silt.

Found approximately 43 feet deep, the precious wreckage escaped destructive forces, resulting in an extraordinary state of preservation that provided archaeologists with a rare, close-up look at never-before-seen details.

“It is extraordinary to have so many parts of the rigging,” noted the researchers.  

Its sheer size and remarkable condition turned the excavation into a massive undertaking that required 289 dives and more than two and a half years to complete, according to Arkeonews. Now, the Viking Ship Museum has come forward, brimming with praise and insights into one of the medieval world’s most impressive vessels.

A groundbreaking cargo ship

Named Svælget 2, this medieval cargo ship would have sailed the seas of Northern Europe with the astounding capacity to carry an immense 300-ton load with only a minimal crew.

Made of Polish and Danish wood, the vessel was described by the Viking Ship Museum in a press release as the “super ship” of the Middle Ages, measuring approximately 91 feet long, 30 feet wide, and 20 feet high.

Dendrochronological analysis revealed that Svælget 2 was built around 1410 using timber originating from Poland and the Netherlands. This indicated to archaeologists that while the frames were cut at the building site, shipbuilders imported other primary materials. So, researchers were impressed to discover that such large quantities of wood were moved across Europe.

In fact, this was the major insight archaeologists gleaned: the cog reflected the existence of a robust, complex trade network that the ship itself helped make possible. In a press release, the Museum called it “the backbone of medieval trade” because it could travel long distances and navigate treacherous waters without a large crew, making it an efficient, low-cost trading vessel.

“It is clear evidence that everyday goods were traded. Shipbuilders went as big as possible to transport bulky cargo – salt, timber, bricks or basic food items,” says Otto Uldum, head archaeologist, in a press release.

The cog was a trailblazer

The ship marked a distinct shift in commerce when the goods exchanged were no longer just luxury items but everyday commodities too. The vessel literally expanded trade by cutting unnecessary costs and carrying heavy loads across Northern Europe.

Due to its excellent state of preservation, archaeologists recovered the ship’s hull, which is a rare find. Along the hull were remnants of the ship’s forecastle and aftcastle, which provided shelter for the crew. Until now, archaeologists could not confirm that these “castles” even existed on such ships.

A major surprise was the discovery of a brick-built galley, the earliest example of its kind ever found in Danish waters. This meant the crew could cook over an open fire in “remarkable comfort.” Besides, they found shoes, a comb, a cooking pot, and a wooden tray, as per Arkeonews.

All these extraordinary finds opened up a unique vantage point into daily life on board a groundbreaking cargo ship that changed the face of trade in Northern Europe. It signaled an economic boom in the region as they had the finances to build a vessel to rule the seas.

🔗 Sumber: interestingengineering.com

📌 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