Invited Talk at HKUST (GZ): Reason with Multimodal Minds in Space
机器之心:只用图像也能思考,强化学习造就推理模型新范式!复杂场景规划能力Max
量子位:纯靠“脑补”图像,大模型推理准确率狂飙80%丨剑桥谷歌新研究
The TWIML AI Podcast with Sam Charrington
IEEE Spectrum, ‘‘Thinking" Visually Boosts AI Problem Solving
新智元:直接可视化多模态推理过程
BMVA: Trustworthy Multimodal Learning with Foundation Models