Ant Group's Lingbo held its first embodied large-model challenge, requiring 15 finalist teams to complete data collection, model tuning and real-robot deployment within 48 hours, with one model handling fruit picking, cosmetics shelving and dual-arm sweeping. The TARS team, averaging 25 years old, discarded the official 210 demonstrations that grasped the dustpan by its handle and re-collected its own data gripping the rear edge, overtaking the runner-up at the last minute with 82 points and an 80,000 RMB prize. The case shows that in real-robot VLA tasks, data collection strategy and contact-point choice can matter more than model tuning.
Robbyant
蚂蚁灵波
Ant Group's embodied AI unit, host of the first embodied large-model challenge and builder of a robotics developer ecosystem.
Curated Intelligence
Robbyant has open-sourced LingBot-Touch under Apache 2.0, converting fingertip tactile sensor readings into signals usable by robot control software and filling the gap where cameras cannot sense grip force or slip. The team trained three same-size encoders (5.46M parameters each) on identical touch data; at 4,000 hours, the JEPA variant that predicts masked regions in latent space reached 49.3% top-5 on tactile-to-video retrieval with MAE under 9%, versus just 3.1% for the VAE. The author attributes JEPA's edge to the sparsity and low signal-to-noise ratio of tactile data.
Ant Group's Robbyant has open-sourced LingBot-Touch, a tactile pretrained foundation model that inserts a unified representation translation layer between raw sensor readings and action models. Experiments show zipper-pulling success rising from 39% with vision-only to 59%, and box handover recovering from 48% to 63%, fixing the degradation caused by sparse, noisy raw tactile signals.