Transfer Extreme Learning Machine with Output Weight Alignment

<div>An illustration of TELM-OWA. (1) A mapping function that transforms the output weight of source ELM into one of target ELM is adopted to align the distribution between domains. (2) The output weight approximation constraint to prevent the negative transfer and realize the efficient transferring of knowledge across domains. (3) The objective function is transformed in standard ELM form to be solved.</div>

Computational Intelligence and Neuroscience

fig1

Figure 1

Figure 1: Transfer Extreme Learning Machine with Output Weight Alignment 