|
I'm learning for interesting machine learning papers - specifically ones that are rather a mixture of case-study and new techniques, or new combinations or realizations of existing techniques. E.g, take a specific problem, try out the standard algorithms, explain their strengths and weaknesses, and make modifications to increase performance. Some examples: link:[Predicting accurate probabilities with a ranking loss by Elkan et al] (icml.cc/2012/papers/372.pdf) link:[Detecting Adversarial Advertisements in the Wild by Sculley et al] (www.eecs.tufts.edu/~dsculley/papers/adversarial-ads.pdf) link:[A Case Study in Machine Learning by JR Quinlan] |
