Event
Lianjie Innovation Embodied Intelligence Center Holds Embodied Data Workshop in Suzhou
1 reports · 1 sources
On October 23, under the guidance of Suzhou International Science and Technology Park (SISPARK), the Lianjie Innovation Embodied Intelligence Center hosted the Embodied Data "Iron Triangle" Workshop in Suzhou, focusing on three major topics: "Models and Data," "Scenario Implementation," and "Infra Infrastructure," organizing a closed-door dialogue between technology suppliers and scenario demand-side parties. The event was by targeted invitation only, and attendees covered segments including data collection and annotation, embodied large models, bodies and core components, scenario implementation, and chips and computing power infrastructure. The first roundtable discussion addressed how to balance the scale and quality of pre-training data against the backdrop of real-machine data demand leaping from the tens-of-thousands-of-hours level to the hundreds-of-thousands-of-hours level and ego data moving toward the tens-of-millions-of-hours level; the other two sessions respectively reviewed the scenario implementation practices of secondary-development companies and the construction of computing power foundations for data collection, training, and deployment.
Report timeline
- Embodied Data 'Iron Triangle' — Models, Scenarios and Infrastructure Seminar Held in Suzhou
On October 23, an embodied data 'Iron Triangle' seminar, guided by Suzhou International Science Park (SISPARK) and hosted by the Lianjie Innovation Embodied Intelligence Center, was held in Suzhou. It focused on three topics — 'models and data,' 'scenario deployment,' and 'Infra infrastructure' — and was held in closed-door format to foster dialogue between technology suppliers and scenario demanders. The first roundtable discussed balancing the scale and quality of pre-training data as real-machine data demand shifts from the tens of thousands of hours level to the hundreds of thousands of hours level and ego data moves toward the tens of millions of hours level; the other two sessions reviewed secondary-development companies' scenario deployment practices and the computing foundation for data collection, training and deployment.
机器人产业应用 · 微信公众号 ·