Xiaojiang Peng

OneEmo outperforms larger models in emotional intelligence tasks while using significantly fewer parameters, showcasing the power of unified multimodal reasoning.

EmoScope redefines emotional image editing by enabling users to discover unique, context-specific editing strategies rather than relying on static templates.

VLA models can now automatically generate diverse, valid demonstrations, overcoming the limitations of traditional training methods and significantly boosting generalization capabilities.

Stop guessing the right action chunk size for your robot: this method uses action entropy to adaptively determine chunk length, leading to smoother and more responsive manipulation.