Optimizing search engines using clickthrough data
Citations
2,515 citations
Cites background or methods from "Optimizing search engines using cli..."
...2 Note that there are some algorithms, such as [68], which were also referred to as ordinal regression based algorithms in the literature....
[...]
...Ranking SVM [63, 68] uses SVM for the task of pairwise classification....
[...]
...• Ground truth mining [3, 68, 105], which targets automatically mining ground truth labels for learning to rank, mainly from click-through logs of search engines....
[...]
...Some work has been done along this direction [3, 68, 105], however, they also have certain limitations....
[...]
...6 This kind of judgment can also be mined from click-through logs of search engines [68, 69, 105]....
[...]
2,511 citations
Cites methods from "Optimizing search engines using cli..."
...MN is due to Taskar, Guestrin, and Koller [2003] and StructSVM was presented by Tsochantaridis, Joachims, Hofmann, and Altun [2005]. An alternative technique for tackling structured prediction as a regression problem was presented and analyzed by Cortes, Mohri, and Weston [2007c]....
[...]
2,460 citations
Cites background from "Optimizing search engines using cli..."
...1, if θ is w0 xi, if θ is wi xi ∑n j=1 vj,fxj − vi,fx2i , if θ is vi,f (4) The sum ∑n j=1 vj,fxj is independent of i and thus can be precomputed (e.g. when computing ŷ(x))....
[...]
...Scoring functions can be learned with pairwise training data [5], where a feature tuple (x(A),x(B)) ∈ D means that x(A) should be ranked higher than x(B)....
[...]
2,292 citations
Cites background or methods from "Optimizing search engines using cli..."
...In contrast, we have proposed an efficient algorithm (Hofmann et al., 2002; Altun et al., 2003; Joachims, 2003) even in the case of very large output spaces, that takes advantage of the sparseness of the maximummargin solution....
[...]
...The same is true also for other ranking algorithms (Cohen et al., 1999; Herbrich et al., 2000; Schapire and Singer, 2000; Crammer and Singer, 2002; Joachims, 2002)....
[...]
2,173 citations
References
40,147 citations
37,861 citations
"Optimizing search engines using cli..." refers background in this paper
...However, just like in classification SVMs [7], it is possible to approximate the solution by introducing (non-negative) slack variables ξi,j,k and minimizing the upper bound ∑ ξi,j,k. Adding SVM regularization for margin maximization to the objective leads to the following optimization problem, which is similar to the ordinal regression approach in [12]....
[...]
...However, just like in classification SVMs [7], it is possible to approximate the solution by introducing (non-negative) slack variables ξi,j,k and minimizing the upper bound ∑ ξi,j,k....
[...]
26,531 citations
"Optimizing search engines using cli..." refers background or methods in this paper
...This makes it possible to use Kernels [4][25] and extend the Ranking SVM algorithm to non-linear retrieval functions....
[...]
...Therefore, the following algorithm directly addresses (6), taking an empirical risk minimization approach [25]....
[...]
...Note that (6) is (proportional to) a risk functional [25] with −τ as the loss function....
[...]
11,211 citations
9,923 citations