RetailSMV: exocentric vs egocentric adaptation of world models

Comparison between exocentric and egocentric adaptation of world models in retail. Surprising results with the RetailSMV dataset.

jueves, 2 de julio de 2026 • 2 min read • Q2BSTUDIO Team

Exocentric outperforms egocentric in world model adaptation

Video models as world simulators represent a significant advance in artificial intelligence, with potential applications in robotics, automation, and business environments. However, their generic training on internet videos makes them ineffective for specific domains like retail, where conditions and perspectives are very particular. A recent study on the RetailSMV corpus explores a crucial dilemma: which training perspective —egocentric (from the employee's point of view) or exocentric (fixed cameras in environments)— produces models better adapted to tasks such as restocking, weighing, or scanning in supermarkets? The results reveal that exclusively exocentric adaptation can match or surpass the combination of both, especially in short-term predictions, suggesting that global vision provides more relevant information for certain operations.

This finding has important practical implications for companies seeking to implement artificial intelligence in their processes. By choosing the right data source, computational resources can be optimized and model accuracy improved. To integrate these capabilities into real systems, it is necessary to have custom applications that efficiently channel, process, and deploy these models. Q2BSTUDIO offers AI for businesses that includes everything from AI agents to computer vision solutions, tailored to each client's needs.

Furthermore, technological infrastructure is key to scaling these systems. AWS and Azure cloud services provide the necessary computing power, while cybersecurity ensures the protection of sensitive data. Once in production, results can be visualized using business intelligence tools like Power BI, enabling managers to make data-driven decisions in real time. The combination of custom software, AI models, and cloud services creates robust ecosystems that transform retail operations.

Ultimately, the adaptation of world models depends not only on the amount of data but also on the right perspective. Companies that understand this dynamic will be able to implement more effective and efficient AI agents. Q2BSTUDIO, with its expertise in custom software development and artificial intelligence solutions, is ready to accompany this journey, offering both technical know-how and integration capabilities in cloud and cybersecurity environments. Innovation in retail begins with informed decisions about the data and the technology that processes it.

A BREAK?

Play for a moment before you go

OUR SERVICES

How we can help you

Do you have a project in mind?

Tell us your vision and we'll turn it into a software solution. Whatever the scope, we make your idea real.