Supervised Contrastive Representation Learning for Individual Recognition in the Wild
Date:
Talk title: “Supervised Contrastive Representation Learning for Individual Recognition in the Wild.” On-going reaching into the application of state-of-the-art deep learning techniques for speaker recognition in the wild.
A signaller $E$ at time $t$ produces $f(C_t;\phi)$ individually distinct cues $C_t$ such as visual, acoustic or olfactory that is propagated in a medium $M(C_t)$. A cue can be uni-modal or multi modal. A receiver $E^{'}$ perceives $g(C_j, C_t; \theta)$ cues through their sensory percepts that are matched against a templating system. A cue is compared with known template $\mathcal{T}$ to a determine unique behavioral response directed to the signaller.
