📌 TOPINDIATOURS Hot ai: Adobe Research Unlocking Long-Term Memory in Video World M

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

📌 TOPINDIATOURS Eksklusif ai: German students build mobile energy trailer using so

Mechanical engineering students in Germany have designed an innovative mobile energy trailer that combines solar power, battery storage, and a hydrogen fuel cell to deliver reliable electricity where the grid doesn’t reach.

The young researchers from the Dortmund University of Applied Sciences and Arts (Fachhochschule Dortmund), recently unveiled the eTrail-Ing as part of their master’s project.

According to the team, the trailer combines solar panels, a hydrogen fuel cell, and battery storage to create a sustainable alternative to noisy diesel generators. It can operate completely independently of the public grid for up to seven days.

The novel solution is designed to tackle the common challenge of ensuring a reliable power supply in places far from existing infrastructure, such as open-air music festivals, crisis zones, and remote scientific expeditions.

“The focus of the eTrail-Ing project is on the students themselves,” Sönke Gößling, PhD, a fuel cells and regenerative energy systems professor at the Faculty of Mechanical Engineering, who supervises the project, stated.

Clean energy on wheels

The system integrates fold-out photovoltaic modules with an output of nearly four kilowatts (kW), a large-capacity battery storage unit, and a 2.5-kilowatt hydrogen fuel cell. This provides reliable electricity for appliances while also maintaining cooling for sensitive materials.

“The trailer supplies power for appliances as well as for the integrated cold storage room,” Niklas Wenderoth, one of the students involved in the project, disclosed. “Depending on the application, drinks can be kept cold there, but blood reserves or medicines can also be cooled.”

The system also includes a 6.76-cubic-meter cooling and heating space with a temperature range of 39 to 68 degrees Fahrenheit (four to 20 degrees Celsius), and offers 230-volt sockets and USB charging ports.

The trailer’s outstanding flexibility makes it useful in both leisure settings and life-saving operations. “The challenge is the complex interaction of different systems,” Finn Floßbach, another member of the student team, revealed.

The trailer’s modular design allows researchers, aid organizations, and even event organizers to adapt the trailer to their needs, regardless of whether for charging equipment, powering lights and communication devices, or running refrigeration systems.

Reliable off-grid energy

The system also features a built-in safety framework to ensure reliable and secure operation. A software system continuously monitors all electrical processes, ensuring that critical applications, such as cooling chains, are not interrupted.

Currently, the trailer’s frame is being assembled at the university’s Sonnenstraße campus. Electrical components, including the hydrogen fuel cell and the battery storage, have already been tested for integration.

“The students plan, design and develop the energy trailer over several semesters – from the dimensioning of the subsystems to safety concepts and the integration of the fuel cell,” Gößling elaborated.

Meanwhile, according to Gößling, the students also manage several partnerships with industrial collaborators. It helps them gain real-world experience in project management and interdisciplinary teamwork.

Each participant is involved for at least two semesters, ensuring continuity and depth of learning. The final prototype is expected to be fully functional by the end of 2026.

“The result is not only a functional, sustainable product, but also a platform for knowledge transfer and applied engineering training – the actual aim of the project,” Gößling concluded in a press release.

🔗 Sumber: interestingengineering.com