Errata for Trading Collobert et al. 2006 for Scalability
In Algorithm 2 the condition
\[0<\alpha_{l}^{t}<C\]for computing $b^t$ should be
\[-\beta_{l}^{t-1}<\alpha_{l}^{t}<C-\beta_{l}^{t-1}\]Thanks for Haoda Fu for pointing out that error.
In Algorithm 2 the condition
\[0<\alpha_{l}^{t}<C\]for computing $b^t$ should be
\[-\beta_{l}^{t-1}<\alpha_{l}^{t}<C-\beta_{l}^{t-1}\]Thanks for Haoda Fu for pointing out that error.