Implicit relevance feedback from a multi-step search process: a use of query-logs

We evaluate the use of clickthrough information as implicit relevance feedback in sessions. We employ records of user interactions with a search system for pictures retrieval: issued queries, clicked images, and purchased content; we investigate whether and how much of the past search history should be used in a feedback loop. We also assess the benefit of using clicked data as positive tokens of relevance to the task of estimating the probability of an image to be purchased.