Anisio Lacerda

Learning to advertise

Content-targeted advertising, the task of automatically associating ads to a Web page, constitutes a key Web monetization strategy nowadays. Further, it introduces new challenging technical problems and raises interesting questions. For instance, how to design ranking functions able to satisfy conflicting goals such as selecting advertisements (ads) that are relevant to the users and suitable and profitable to the publishers and advertisers? In this paper we propose a new framework for associating ads with web pages based on Genetic Programming (GP). Our GP method aims at learning functions that select the most appropriate ads, given the contents of a Web page. These ranking functions are designed to optimize overall precision and minimize the number of misplacements. By using a real ad collection and web pages from a newspaper, we obtained a gain over a state-of-the-art baseline method of 61.7% in average precision. Further, by evolving individuals to provide good ranking estimations, GP was able to discover ranking functions that are very effective in placing ads in web pages while avoiding irrelevant ones.

Pareto-efficient hybridization for multi-objective recommender systems

Performing accurate suggestions is an objective of paramount importance for effective recommender systems. Other important and increasingly evident objectives are novelty and diversity, which are achieved by recommender systems that are able to suggest diversified items not easily discovered by the users. Different recommendation algorithms have particular strengths and weaknesses when it comes to each of these objectives, motivating the construction of hybrid approaches. However, most of these approaches only focus on optimizing accuracy, with no regard for novelty and diversity. The problem of combining recommendation algorithms grows significantly harder when multiple objectives are considered simultaneously. For instance, devising multi-objective recommender systems that suggest items that are simultaneously accurate, novel and diversified may lead to a conflicting-objective problem, where the attempt to improve an objective further may result in worsening other competing objectives. In this paper we propose a hybrid recommendation approach that combines existing algorithms which differ in their level of accuracy, novelty and diversity. We employ an evolutionary search for hybrids following the Strength Pareto approach, which isolates hybrids that are not dominated by others (i.e., the so called Pareto frontier). Experimental results on two recommendation scenarios show that: (i) we can combine recommendation algorithms in order to improve an objective without significantly hurting other objectives, and (ii) we allow for adjusting the compromise between accuracy, diversity and novelty, so that the recommendation emphasis can be adjusted dynamically according to the needs of different users.

Demand-driven tag recommendation

Collaborative tagging allows users to assign arbitrary keywords (or tags) describing the content of objects, which facilitates navigation and improves searching without dependence on pre-configured categories. In large-scale tag-based systems, tag recommendation services can assist a user in the assignment of tags to objects and help consolidate the vocabulary of tags across users. A promising approach for tag recommendation is to exploit the co-occurrence of tags. However, these methods are challenged by the huge size of the tag vocabulary, either because (1) the computational complexity may increase exponentially with the number of tags or (2) the score associated with each tag may become distorted since different tags may operate in different scales and the scores are not directly comparable. In this paper we propose a novel method that recommends tags on a demand-driven basis according to an initial set of tags applied to an object. It reduces the space of possible solutions, so that its complexity increases polynomially with the size of the tag vocabulary. Further, the score of each tag is calibrated using an entropy minimization approach which corrects possible distortions and provides more precise recommendations. We conducted a systematic evaluation of the proposed method using three types of media: audio, bookmarks and video. The experimental results show that the proposed method is fast and boosts recommendation quality on different experimental scenarios. For instance, in the case of a popular audio site it provides improvements in precision (p@5) ranging from 6.4% to 46.7% (depending on the number of tags given as input), outperforming a recently proposed co-occurrence based tag recommendation method.

Multiobjective Pareto-Efficient Approaches for Recommender Systems

Recommender systems are quickly becoming ubiquitous in applications such as e-commerce, social media channels, and content providers, among others, acting as an enabling mechanism designed to overcome the information overload problem by improving browsing and consumption experience. A typical task in many recommender systems is to output a ranked list of items, so that items placed higher in the rank are more likely to be interesting to the users. Interestingness measures include how accurate, novel, and diverse are the suggested items, and the objective is usually to produce ranked lists optimizing one of these measures. Suggesting items that are simultaneously accurate, novel, and diverse is much more challenging, since this may lead to a conflicting-objective problem, in which the attempt to improve a measure further may result in worsening other measures. In this article, we propose new approaches for multiobjective recommender systems based on the concept of Pareto efficiency—a state achieved when the system is devised in the most efficient manner in the sense that there is no way to improve one of the objectives without making any other objective worse off. Given that existing multiobjective recommendation algorithms differ in their level of accuracy, diversity, and novelty, we exploit the Pareto-efficiency concept in two distinct manners: (i) the aggregation of ranked lists produced by existing algorithms into a single one, which we call Pareto-efficient ranking, and (ii) the weighted combination of existing algorithms resulting in a hybrid one, which we call Pareto-efficient hybridization. Our evaluation involves two real application scenarios: music recommendation with implicit feedback (i.e., Last.fm) and movie recommendation with explicit feedback (i.e., MovieLens). We show that the proposed Pareto-efficient approaches are effective in suggesting items that are likely to be simultaneously accurate, diverse, and novel. We discuss scenarios where the system achieves high levels of diversity and novelty without compromising its accuracy. Further, comparison against multiobjective baselines reveals improvements in terms of accuracy (from 10.4% to 10.9%), novelty (from 5.7% to 7.5%), and diversity (from 1.6% to 4.2%).

Minimal perfect hashing: A competitive method for indexing internal memory

A perfect hash function (PHF) is an injective function that maps keys from a set S to unique values. Since no collisions occur, each key can be retrieved from a hash table with a single probe. A minimal perfect hash function (MPHF) is a PHF with the smallest possible range, that is, the hash table size is exactly the number of keys in S. MPHFs are widely used for memory efficient storage and fast retrieval of items from static sets. Differently from other hashing schemes, MPHFs completely avoid the problem of wasted space and wasted time to deal with collisions. Until recently, the amount of space to store an MPHF description for practical implementations found in the literature was O(logn) bits per key and therefore similar to the overhead of space of other hashing schemes. Recent results on MPHFs presented in the literature changed this scenario: an MPHF can now be described by approximately 2.6 bits per key. The objective of this paper is to show that MPHFs are, after the new recent results, a good option to index internal memory when static key sets are involved and both successful and unsuccessful searches are allowed. We have shown that MPHFs provide the best tradeoff between space usage and lookup time when compared with other open addressing and chaining hash schemes such as linear hashing, quadratic hashing, double hashing, dense hashing, cuckoo hashing, sparse hashing, hopscotch hashing, chaining with move to front heuristic and exact fit. We considered lookup time for successful and unsuccessful searches in two scenarios: (i) the MPHF description fits in the CPU cache and (ii) the MPHF description does not fit entirely in the CPU cache. Considering lookup time, the minimal perfect hashing outperforms the other hashing schemes in the two scenarios and, in the first scenario, the performance is better even when the compared methods leave more than 80% of the hash table entries free. Considering space overhead (the amount of used space other than the key-value pairs), the minimal perfect hashing is within a factor of O(logn) bits lower than the other hashing schemes for both scenarios.

A general framework to expand short text for topic modeling

A framework to generate pseudo-documents suitable for topic modeling is proposed.An instantiation of the framework based on word vector representations is presented.Results of NPMI and F1 obtained are better than those of state-of-the art methods. Short texts are everywhere in the Web, including messages posted in social media, status messages and blog comments, and uncovering the topics of this type of messages is crucial to a wide range of applications, e.g., context analysis and user characterization. Extracting topics from short text is challenging because of the dependence of conventional methods, such as Latent Dirichlet Allocation, in words co-occurrence, which in short text is rare and make these methods suffer from severe data sparsity. This paper proposes a general framework for topic modeling of short text by creating larger pseudo-document representations from the original documents. In the framework, document components (e.g., words or bigrams) are defined over a metric space, which provides information about the similarity between them. We present two simple, effective and efficient methods that specialize our general framework to create larger pseudo-documents. While the first method considers word co-occurrence to define the metric space, the second relies on distributed word vector representations. The pseudo-documents generated can be given as input to any topic modeling algorithm. Experiments run in seven datasets and compared against state-of-the-art methods for extracting topics by generating pseudo-documents or modifying current topic modeling methods for short text show the methods significantly improve results in terms of normalized pointwise mutual information. A classification task was also used to evaluate the quality of the topics in terms of document representation, where improvements in F1 varied from 1.5 to 15%.

Exploratory and interactive daily deals recommendation

Daily deals sites (DDSs), such as Groupon and LivingSocial, attract millions of customers in the hunt for products and services at significantly reduced prices. A typical approach to increase revenue is to send email messages featuring the deals of the day. Such daily messages, however, are usually not centered on the customers, instead, all registered users typically receive similar messages with almost the same deals. Traditional recommendation algorithms are innocuous in DDSs because: (i) most of the users are sporadic bargain hunters, and thus past preference data is extremely sparse, (ii) deals have a short living period, and thus data is extremely volatile, and (iii) user taste and interest may undergo temporal drifts. In order to address such particularly challenging scenario, we propose new algorithms for daily deals recommendation based on the explore-then-exploit strategy.Users are split into exploration and exploitation sets -- in the exploration set the users receive non-personalized messages and a co-purchase network is updated with user feedback for purchases of the day, while in the exploitation set the updated network is used for recommending personalized messages for the remaining users.A thorough evaluation of our algorithms using real data obtained from a large daily deals website in Brazil in contrast to state-of-the-art recommendation algorithms show gains in precision ranging from 18% to 34%.

Building user profiles to improve user experience in recommender systems

Recommender systems are quickly becoming ubiquitous in many Web applications, including e-commerce, social media channels, content providers, among others. These systems act as an enabling mechanism designed to overcome the information overload problem by improving browsing and consumption experience. Crucial to the performance of a recommender system is the accuracy of the user profiles used to represent the interests of the users. In this proposal, we analyze three different aspects of user profiling: (i) selecting the most informative events from the interaction between users and the system, (ii) combining different recommendation algorithms to (iii) including trust-aware information in user profiles to improve the accuracy of recommender systems.

Multi-Objective Ranked Bandits for Recommender Systems

This paper is interested in recommender systems that work with implicit feedback in dynamic scenarios providing online recommendations, such as news articles and ads recommendation in Web portals. In these dynamic scenarios, user feedback to the system is given through clicks, and feedback needs to be quickly exploited to improve subsequent recommendations. In this scenario, we propose an algorithm named multi-objective ranked bandits, which in contrast with current methods in the literature, is able to recommend lists of items that are accurate, diverse and novel. The algorithm relies on four main components: a scalarization function, a set of recommendation quality metrics, a dynamic prioritization scheme for weighting these metrics and a base multi-armed bandit algorithm. Results show that our algorithm provides improvements of 7.8 and 10.4% in click-through rate in two real-world large-scale datasets when compared to the single-objective state-of-the-art algorithm.

Topic Modeling for Short Texts with Co-occurrence Frequency-Based Expansion

Short texts are everywhere on the Web, including messages in social media, status messages, etc, and extracting semantically meaningful topics from these collections is an important and difficult task. Topic modeling methods, such as Latent Dirichlet Allocation, were designed for this purpose. However, discovering high quality topics in short text collections is a challenging task. This is because most topic modeling methods rely on information coming from the word co-occurrence distribution in the collection to extract topics. As in short text this information is scarce, topic modeling methods have difficulties in this scenario, and different strategies to tackle this problem have been proposed in the literature. In this direction, this paper introduces a method for topic modeling of short texts that creates pseudo-documents representations from the original documents. The method is simple, effective, and considers word co-occurrence to expand documents, which can be given as input to any topic modeling algorithm. Experiments were run in four datasets and compared against state-of-the-art methods for extracting topics from short text. Results of coherence, NPMI and clustering metrics showed to be statistically significantly better than the baselines in the majority of cases.