# Sparse Coding

 Revision as of 13:34, 21 March 2011 (view source)Zhenghao (Talk | contribs) (→Sparse Coding)← Older edit Latest revision as of 04:28, 8 April 2013 (view source)Kandeng (Talk | contribs) Line 88: Line 88: &= \sum_{j=1}^{m} \left|\left| \mathbf{x}^{(j)} - \sum_{i=1}^k a^{(j)}_i \mathbf{\phi}_{i}\right|\right|^{2} + \lambda \sum_{i=1}^{k}S(a^{(j)}_i) &= \sum_{j=1}^{m} \left|\left| \mathbf{x}^{(j)} - \sum_{i=1}^k a^{(j)}_i \mathbf{\phi}_{i}\right|\right|^{2} + \lambda \sum_{i=1}^{k}S(a^{(j)}_i) \end{array}[/itex] \end{array}[/itex] - where $\lambda = 2*\sigma^2*\beta$ and irrelevant constants have been hidden. Since maximizing the log-likelihood is equivalent to minimizing the energy function, we recover the original optimization problem: + where $\lambda = 2\sigma^2\beta$ and irrelevant constants have been hidden. Since maximizing the log-likelihood is equivalent to minimizing the energy function, we recover the original optimization problem: :[itex]\begin{align} :[itex]\begin{align} \mathbf{\phi}^{*},\mathbf{a}^{*}=\text{argmin}_{\mathbf{\phi},\mathbf{a}} \sum_{j=1}^{m} \left|\left| \mathbf{x}^{(j)} - \sum_{i=1}^k a^{(j)}_i \mathbf{\phi}_{i}\right|\right|^{2} + \lambda \sum_{i=1}^{k}S(a^{(j)}_i) \mathbf{\phi}^{*},\mathbf{a}^{*}=\text{argmin}_{\mathbf{\phi},\mathbf{a}} \sum_{j=1}^{m} \left|\left| \mathbf{x}^{(j)} - \sum_{i=1}^k a^{(j)}_i \mathbf{\phi}_{i}\right|\right|^{2} + \lambda \sum_{i=1}^{k}S(a^{(j)}_i) Line 103: Line 103: As described above, a significant limitation of sparse coding is that even after a set of basis vectors have been learnt, in order to "encode" a new data example, optimization must be performed to obtain the required coefficients. This significant "runtime" cost means that sparse coding is computationally expensive to implement even at test time especially compared to typical feedforward architectures. As described above, a significant limitation of sparse coding is that even after a set of basis vectors have been learnt, in order to "encode" a new data example, optimization must be performed to obtain the required coefficients. This significant "runtime" cost means that sparse coding is computationally expensive to implement even at test time especially compared to typical feedforward architectures. + + + + {{Languages|稀疏编码|中文}}