Most recommender algorithms are designed to suggest relevant items, but suggesting these items does not always result in user satisfaction. Therefore, the efforts in recommender systems recently shifted towards serendipity, but generating serendipitous recommendations is difficult due to the lack of training data. To the best of our knowledge, there are many large datasets containing relevance scores (relevance oriented) and only one publicly available dataset containing a relatively small number of serendipity scores (serendipity oriented). This limits the learning capabilities of serendipity oriented algorithms. Therefore, in the absence of any known deep learning algorithms for recommending serendipitous items and the lack of large serendipity oriented datasets, we introduce SerRec our novel transfer learning method to recommend serendipitous items. SerRec uses transfer learning to firstly train a deep neural network for relevance scores using a large dataset and then tunes it for serendipity scores using a smaller dataset. Our method shows benefits of transfer learning for recommending serendipitous items as well as performance gains over the state-of-the-art serendipity oriented algorithms

We introduce CitySearcher, a vertical search engine that searches for cities when queried for an interest. Generally in search engines, utilization of semantics between words is favorable for performance improvement. Even though ambiguous query words have multiple semantic meanings, search engines can return diversified results to satisfy different users' information needs. But for CitySearcher, mismatched semantic relationships can lead to extremely unsatisfactory results. For example, the city Sale would incorrectly rank high for the interest shopping because of semantic interpretations of the words. Thus in our system, the main challenge is to eliminate the mismatched semantic relationships resulting from the side effect of the semantic models. In the previous case, we aim to ignore the semantics of a city's name which is not indicative of the city's characteristics. In CitySearcher, we use word2vec, a very popular word embedding technique to estimate the semantics of the words and create the initial ranks of the cities. To reduce the effect of the mismatched semantic relationships, we generate a set of features for learning based on a novel clustering-based method. With the generated features, we then utilize learning to rank algorithms to rerank the cities for return. We use the English version of Wikivoyage dataset for evaluation of our system, where we sample a very small dataset for training. Experimental results demonstrate the performance gain of our system over various standard retrieval techniques.