Ant Group's AI research lab has released LingBot-VA 2.0, a new foundational model for embodied AI. This model is designed to predict future actions and states, enabling robots to perform complex tasks with greater foresight. LingBot-VA 2.0 achieves this through a novel semantic visual-action tokenizer, causal pre-training, a sparse Mixture-of-Experts (MoE) architecture for efficient inference, and an asynchronous reasoning mechanism that allows for real-time adjustments. AI
IMPACT This model's predictive capabilities could significantly advance robot autonomy and task execution in dynamic environments.
RANK_REASON New foundational model release from a major AI lab. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
