List of loss functions for linear models? List of regularization losses?

Some popular loss functions are:

• absolute loss, |x - y|
• epsilon-insensitive loss, max(0, |x-y| - epsilon)
• squared loss, (x - y)^2
• squared epsilon-insensitive loss, max(0, (x-y)^2 - epsilon)
• 0-1 loss (for y discrete), I{x = y}
• hinge loss (for y = +-1), max(0, 1 - yx)
• squared hinge loss (for y = +-1), max(0, 1-yx)^2
• multiclass or structured generalizations of the hinge or squared hinge loss
• exponential loss (for y = +-1), exp(-yx)
• logistic (or log) loss (for y = +-1), log(1 + exp(-yx))
• the softmax generalization of the logistic loss, exp(f(x,y)) / (sum_i exp(f(x,y_i)))
• ramp loss (for y=+-1) , min(1, max(0, 1-yx))
• Huber loss (follow the link, as it looks ugly in this notation)
• the KL divergence (for x and y positive summing to one), sum_i x_i log(x_i / y_i)
• in general, the Bregman divergence of any convex function

Some regularizers:

• any p-norm (square norm, l1 norm, the maximum, etc)
• the "l_0" norm (number of nonzero components of the vector)
• combinations of the above (square norm + l1 norm = elastic net, compositions of norms over norms lead to structured lasso, etc)
• the entropy, - sum_i x_i log x_i
• again, generally, any Bregman divergence with regard to a fixed base vector

Of course, this list is incomplete, but I hope I've covered most used in practice.