Technology NewsTechnology NewsTechnology News
  • Computing
  • AI
  • Robotics
  • Cybersecurity
  • Electric Vehicle
  • Wearables
  • Gaming
  • Space
Reading: ShengShu Trains Robots Using Simulated Worlds Through Vidar Model
Share
Font ResizerAa
Technology NewsTechnology News
Font ResizerAa
Search
  • Computing
  • AI
  • Robotics
  • Cybersecurity
  • Electric Vehicle
  • Wearables
  • Gaming
  • Space
Follow US
  • Cookie Policy (EU)
  • Contact
  • About
© 2025 NEWSLINKER - Powered by LK SOFTWARE
AI

ShengShu Trains Robots Using Simulated Worlds Through Vidar Model

Highlights

  • ShengShu introduces Vidar, a model training robots using simulated environments.

  • Vidar reduces data needs and speeds up adaptation across robotics sectors.

  • Hybrid simulation from Vidar addresses prior model limitations in scalability and realism.

Ethan Moreno
Last updated: 8 August, 2025 - 8:19 pm 8:19 pm
Ethan Moreno 3 hours ago
Share
SHARE

Accelerating robot intelligence development presents several barriers, particularly the need for extensive physical data collection. ShengShu Technology is seeking to mitigate these hurdles with the introduction of Vidar, a multi-view physical AI training model that advances robotics development through simulated environments. This model, powered by the company’s Vidu generative video platform, circumvents the constraints of hardware reliance and time-intensive real-world testing, instead leveraging minimal physical data to create rich, scalable virtual training scenarios. The development arrives as AI-driven robotics grow more prevalent in sectors where safety, cost, and adaptability are critical, raising interest in alternatives to traditional, data-heavy AI training methods.

Contents
How Does Vidar Change Robot Training Methods?What Technical Framework Underpins Vidar’s Approach?Where Could Vidar Impact Robotics Industries?

Recent information about embodied AI models mainly focused on the extensive physical training data required to achieve robust decision-making abilities in robots. Earlier solutions, both from startups and industry giants, emphasized the limitations of either all-simulated or all-physical data approaches, noting each method’s trade-offs in scalability, realism, and cost. Vidar departs from past initiatives by combining the advantages of generative video and selective integration of limited physical training, aiming to bridge the performance gap in real-world deployment. While other contemporary models still face challenges with edge-case variability and data consumption, Vidar’s method positions it differently in this field.

How Does Vidar Change Robot Training Methods?

Vidar introduces a hybrid training process that fuses generative simulation with concise real-world data to build complex, yet cost-effective, AI training environments. Unlike past systems rooted primarily in physical interaction or in exclusively virtual data, Vidar’s model achieves scalable results through rapid scenario generation and multi-angle video prediction. ShengShu Technology highlighted that this approach utilizes only around 20 minutes of data for certain tasks, significantly less compared to models like RDT and π0.5 which require larger datasets. By shifting data collection largely to a virtual space, organizations can adapt and deploy robotic AI systems in dynamic settings more efficiently.

What Technical Framework Underpins Vidar’s Approach?

Based on the U-ViT model, Vidar operates through a two-stage modular learning structure, which separates perceptual understanding from control mechanisms. The initial stage leverages general and embodied video data to train perception, which is then processed into motor commands by a task-agnostic model called AnyPos. ShengShu stated,

“Vidar offers a radically different approach to training embodied AI models.”

This structure enables the seamless transfer of learned behaviors across diverse robot types, reducing the engineering workload required for system adaptation.

Where Could Vidar Impact Robotics Industries?

Vidar’s flexibility allows it to be applied in sectors such as manufacturing, healthcare, eldercare, and home automation where rapid adaptation to new environments is crucial. ShengShu noted that its model’s minimal reliance on hardware and its data pyramid—spanning general, embodied, and robot-specific examples—make it a versatile solution for varied use cases. The company remarked,

“Vidar creates an AI-native path for robotics development that is efficient, scalable, and cost-effective.”

Simulation-based training could speed up the deployment of robots able to handle an expanding array of complex tasks with fewer resources and training cycles.

ShengShu’s progress with Vidu and Vidar aligns with larger industry efforts to generalize AI capabilities across multimodal and physical domains. Compared to previous robotic AI models where vast amounts of either simulated or physical data were required, Vidar stands out for its use of minimal real-world input combined with advanced generative video. For organizations considering AI deployment in practical robotics, reducing dependence on resource-heavy data pipelines may lower costs and open new applications previously seen as impractical. Understanding where traditional and hybrid approaches succeed or fall short is vital for decision-makers in robotics strategy and development planning.

You can follow us on Youtube, Telegram, Facebook, Linkedin, Twitter ( X ), Mastodon and Bluesky

You Might Also Like

AI Leaders Call for Clear Rules to Tackle Trust Crisis

Experts Advocate Human-Centred Approach for Future AI Systems

FORT Robotics Secures $18.9M to Advance Robotic Safety Systems

PrismaX Launches Teleop Platform to Broaden Robotic Arm Use

Experts Warn AI Hype Weakens Crucial Human Skills

Share This Article
Facebook Twitter Copy Link Print
Ethan Moreno
By Ethan Moreno
Ethan Moreno, a 35-year-old California resident, is a media graduate. Recognized for his extensive media knowledge and sharp editing skills, Ethan is a passionate professional dedicated to improving the accuracy and quality of news. Specializing in digital media, Moreno keeps abreast of technology, science and new media trends to shape content strategies.
Previous Article Tesla Equips Model 3 with Front Camera in Latest China Filing
Next Article Bithell Games Cuts Staff After Failing to Secure New Project

Stay Connected

6.2kLike
8kFollow
2.3kSubscribe
1.7kFollow

Latest News

Google Pixel Watch 4 Leak Reveals New Features and Major Changes
Wearables
Bithell Games Cuts Staff After Failing to Secure New Project
Gaming
Tesla Equips Model 3 with Front Camera in Latest China Filing
Electric Vehicle
Microsoft Stresses Fast Recovery Depends on Cyberattack Preparation
Cybersecurity
Tesla Prepares Major FSD Update, Targets Driver Attention Reduction
Electric Vehicle
NEWSLINKER – your premier source for the latest updates in ai, robotics, electric vehicle, gaming, and technology. We are dedicated to bringing you the most accurate, timely, and engaging content from across these dynamic industries. Join us on our journey of discovery and stay informed in this ever-evolving digital age.

ARTIFICAL INTELLIGENCE

  • Can Artificial Intelligence Achieve Consciousness?
  • What is Artificial Intelligence (AI)?
  • How does Artificial Intelligence Work?
  • Will AI Take Over the World?
  • What Is OpenAI?
  • What is Artifical General Intelligence?

ELECTRIC VEHICLE

  • What is Electric Vehicle in Simple Words?
  • How do Electric Cars Work?
  • What is the Advantage and Disadvantage of Electric Cars?
  • Is Electric Car the Future?

RESEARCH

  • Robotics Market Research & Report
  • Everything you need to know about IoT
  • What Is Wearable Technology?
  • What is FANUC Robotics?
  • What is Anthropic AI?
Technology NewsTechnology News
Follow US
About Us   -  Cookie Policy   -   Contact

© 2025 NEWSLINKER. Powered by LK SOFTWARE
Welcome Back!

Sign in to your account

Register Lost your password?