📌 TOPINDIATOURS Breaking ai: Video Shows Passenger Bail From Waymo in Path of Onco

A passenger bailed from his Waymo robotaxi as the vehicle drove dangerously close to an oncoming railcar.

Viral footage of the incident, which took place last Wednesday, shows the car blundering down a stretch of light rail tracks in Phoenix, Arizona, as if it were an actual motorist lane it was supposed to be driving in. The passenger, smelling trouble, exits the car during one of its random stops without looking back.

“Oh get out get out get out!” the bystander taking the footage can be heard saying. Moments later, a railcar a few hundred yards directly behind the Waymo can be briefly seen beginning to pull out from the station.

But the Waymo’s antics weren’t over yet. After driving again, it then suddenly stops as another railcar comes lumbering down the opposing pair of tracks, while the one behind it appears unsure of whether to proceed. Piling one bad decision on top of another, the robotaxi decides to start reversing, but doesn’t leave the track before the video ends.

A Valley Metro spokesperson said that Waymo was notified about the incident after an employee noticed the baffling spectacle.

“At approximately 9 a.m. on Wednesday morning, a Valley Metro employee observed an autonomous Waymo vehicle on the northbound light rail tracks near Southern Avenue and Central Avenue in Phoenix,” the spokesperson said, as quoted by local news station KTVK. “Light rail operations staff responded to the scene, and Waymo was contacted. To minimize service impacts, northbound and southbound trains exchanged passengers before reversing direction to continue service.”

Waymo robotaxis, despite touting an impressive safety record, have been known to get into all kinds of bizarre traffic contretemps, many of them dangerous. The cars have been caught driving on the wrong side of the road, getting stuck in roundabouts, blowing through police standoffs, and ignoring stopped school buses. The company also faced intense outrage after one of its self-driving cabs ran over and killed a beloved bodega cat in San Francisco.

Scrutiny into the robotaxis’ safety and capabilities is higher than ever after seemingly its entire fleet went haywire during a power outage in the Bay Area last month, plunging the streets of San Francisco into chaos as the rudderless cabs piled up at intersections and blocked traffic. 

Andrew Maynard, an emerging and transformative technology professor at Arizona State University, dismissed the latest Waymo incident as an edge case.

“I actually felt a little sorry for the car. It obviously made a bad decision and got itself in a difficult place,” Maynard told KTVK. “This is exactly one of those edge cases, what we call them. Something unexpected where the machine drove like a machine rather than a person.”

Waymo has not made a statement about the incident.

More on self-driving vehicles: Driverless Delivery Vans in China Are Rampaging Through Cities Like Grand Theft Auto

The post Video Shows Passenger Bail From Waymo in Path of Oncoming Train appeared first on Futurism.

🔗 Sumber: futurism.com

📌 TOPINDIATOURS Eksklusif ai: Adobe Research Unlocking Long-Term Memory in Video W

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