MAROKO133 Hot ai: ‘Human washing machine’ that can clean head to toe in 15 mins goes on sa

📌 MAROKO133 Hot ai: ‘Human washing machine’ that can clean head to toe in 15 mins

Science Inc., a Japanese company, has now officially launched what it calls the “Mirai Human Washing Machine.” First unveiled at the World Expo in Osaka earlier this year, Japanese consumers can now officially get their hands on one.

The machine acts like a high-tech automatic spa pod that can wash an occupant head to toe in minutes. Users enter the pod and recline into its comfortable seat.

Once inside, the pod is closed, and the machine washes them using microbubbles. The machine can also rinse off the bubbles and dry the occupant, all while playing relaxing music. 

To this end, the machine not only washes your body but also “washes the soul,” according to company spokeswoman Sachiko Maekura. This pod can also monitor a user’s vitals to enhance safety and comfort.

The human washing machine arrives

The entire process takes about 15 minutes, with zero effort from the user. Interestingly, the concept first appeared at the 1970 Osaka Expo, showcased by Sanyo Electric (now Panasonic Holdings).

The Chairman of Science, Yasuaki Aoyama, saw it as a child and revived the idea with modern tech. It became a massive attraction again at the 2025 Osaka World Expo, so much so that they decided to commercialize it.

The pod itself measures 8.2 feet (2.5 meters) long by 3.28 feet (1 meter) wide by 8.53 feet (2.6 meters) tall, making it big enough for most people to comfortably lie inside. The cleaning is achieved using bubbles so small they can get into pores, removing oil, dirt, and dead skin.

This is a high-end Japanese spa technology already used in baths and salons. Sensors in the pod monitor your vital signs to help prevent fainting, panic, or medical issues.

So what’s the catch? The price, of course. According to reports, the machine will hit the market with a cool 60 Million Yen ($385,000) price tag.

To this end, the product is aimed at the luxury commercial spa market, not necessarily as a household consumer item. For this reason, it will likely be the preserve of high-end hotels, Japanese onsens, theme parks, health spas, and holiday resorts. 

A big price tag

The company plans to produce a limited run of 40 to 50 units, each hand-built. One unnamed Japanese hotel has already committed to buying the first unit produced, with another 5 to 8 units reserved by other potential clients.

Given the limited product run and price, the new machine isn’t really about washing your body. It’s about Japan’s long-running fascination with automation and care robots, especially for an aging population.

A device that washes, dries, and keeps you safe is essentially a prototype for future automated elderly care systems and fully automated bathing facilities. For those hoping to get their hands on one at home, there is some hope, too.

“As the technology advances and mass production becomes more feasible, a more affordable home model may be released,” Science has stated.

🔗 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