Search papers, labs, and topics across Lattice.
1
0
2
By explicitly modeling a multi-level semantic hierarchy and carefully controlling information exchange between modalities, CLCR achieves state-of-the-art results in multimodal learning tasks ranging from emotion recognition to action recognition.
