Javad Azimi

Multimedia features for click prediction of new ads in display advertising

Non-guaranteed display advertising (NGD) is a multi-billion dollar business that has been growing rapidly in recent years. Advertisers in NGD sell a large portion of their ad campaigns using performance dependent pricing models such as cost-per-click (CPC) and cost-per-action (CPA). An accurate prediction of the probability that users click on ads is a crucial task in NGD advertising because this value is required to compute the expected revenue. State-of-the-art prediction algorithms rely heavily on historical information collected for advertisers, users and publishers. Click prediction of new ads in the system is a challenging task due to the lack of such historical data. The objective of this paper is to mitigate this problem by integrating multimedia features extracted from display ads into the click prediction models. Multimedia features can help us capture the attractiveness of the ads with similar contents or aesthetics. In this paper we evaluate the use of numerous multimedia features (in addition to commonly used user, advertiser and publisher features) for the purposes of improving click prediction in ads with no history. We provide analytical results generated over billions of samples and demonstrate that adding multimedia features can significantly improve the accuracy of click prediction for new ads, compared to a state-of-the-art baseline model.

Clustering Ensembles Using Ants Algorithm

Cluster ensembles combine different clustering outputs to obtain a better partition of the data. There are two distinct steps in cluster ensembles, generating a set of initial partitions that are different from one another, and combining the partitions via a consensus functions to generate the final partition. Most of the previous consensus functions require the number of clusters to be specified a priori to obtain a good final partition. In this paper we introduce a new consensus function based on the Ant Colony Algorithms, which can automatically determine the number of clusters and produce highly competitive final clusters. In addition, the proposed method provides a natural way to determine outlier and marginal examples in the data. Experimental results on both synthetic and real-world benchmark data sets are presented to demonstrate the effectiveness of the proposed method in predicting the number of clusters and generating the final partition as well as detecting outlier and marginal examples from data.

The impact of visual appearance on user response in online display advertising

Display advertising has been a significant source of revenue for publishers and ad networks in the online advertising ecosystem. One of the main goals in display advertising is to maximize user response rate for advertising campaigns, such as click through rates (CTR) or conversion rates. Although %in the online advertising industry we believe that the visual appearance of ads (creatives) matters for propensity of user response, there is no published work so far to address this topic via a systematic data-driven approach. In this paper we quantitatively study the relationship between the visual appearance and performance of creatives using large scale data in the worlds largest display ads exchange system, RightMedia. We designed a set of 43 visual features, some of which are novel and some are inspired by related work. We extracted these features from real creatives served on RightMedia. Then, we present recommendations of visual features that have the most important impact on CTR to the professional designers in order to optimize their creative design. We believe that the findings presented in this paper will be very useful for the online advertising industry in designing high-performance creatives. We have also designed and conducted an experiment to evaluate the effectiveness of visual features by themselves for CTR prediction.

Visual appearance of display ads and its effect on click through rate

One of the most important categories of online advertising is display advertising which provides publishers with significant revenue. Similar to other categories, the main goal in display advertising is to maximize user response rate for advertising campaigns, such as click through rates (CTR) or conversion rates. Previous studies have tried to optimize these parameters using objectives such as behavioral targeting. However, there is no published work so far to address the effect of the visual appearance of ads (creatives) on user response rate via a systematic data-driven approach. In this paper, we quantitatively study the relationship between the visual appearance and performance of creatives using large scale data in the worlds largest display ads exchange system, RightMedia. We designed a set of 43 visual features, some of which are novel and others are inspired by related work. We extracted these features from real creatives served on RightMedia. We also designed and conducted a series of experiments to evaluate the effectiveness of visual features for CTR prediction, ranking and performance classification. Based on the evaluation results, we selected a subset of features that have the highest impact on CTR. We believe that the findings presented in this paper will be very useful for the online advertising industry in designing high-performance creatives. It also provides the research community with the first ever data set, initial insights into visual appearances effect on user response propensity, and evaluation benchmarks for further study.

Budgeted optimization with constrained experiments

Motivated by a real-world problem, we study a novel budgeted optimization problem where the goal is to optimize an unknown function f(ċ) given a budget by requesting a sequence of samples from the function. In our setting, however, evaluating the function at precisely specified points is not practically possible due to prohibitive costs. Instead, we can only request constrained experiments. A constrained experiment, denoted by Q, specifies a subset of the input space for the experimenter to sample the function from. The outcome of Q includes a sampled experiment x, and its function output f(x). Importantly, as the constraints of Q become looser, the cost of fulfilling the request decreases, but the uncertainty about the location x increases. Our goal is to manage this trade-off by selecting a set of constrained experiments that best optimize f(ċ) within the budget. We study this problem in two different settings, the non-sequential (or batch) setting where a set of constrained experiments is selected at once, and the sequential setting where experiments are selected one at a time. We evaluate our proposed methods for both settings using synthetic and real functions. The experimental results demonstrate the efficacy of the proposed methods.

Contextual Query Intent Extraction for Paid Search Selection

Paid Search algorithms play an important role in online advertising where a set of related ads is returned based on a searched query. The Paid Search algorithms mostly consist of two main steps. First, a given searched query is converted to different sub-queries or similar phrases which preserve the core intent of the query. Second, the generated sub-queries are matched to the ads bidded keywords in the data set, and a set of ads with highest utility measuring relevance to the original query are returned. The focus of this paper is optimizing the first step by proposing a contextual query intent extraction algorithm to generate sub-queries online which preserve the intent of the original query the best. Experimental results over a very large real-world data set demonstrate the superb performance of proposed approach in optimizing both relevance and monetization metrics compared with one of the existing successful algorithms in our system.

Ads Keyword Rewriting Using Search Engine Results

Paid Search (PS) ads are one of the main revenue sources of online advertising companies where the goal is returning a set of relevant ads for a searched query in search engine websites such as Bing. Typical PS algorithms, return the ads which their Bided Keywords (BKs) are a subset of searched queries or relevant to them. However, there is a huge gap between BKs and searched queries as a considerable amount of BKs are rarely searched by the users. This is mostly due to the rare BKs provided by advertisers. In this paper, we propose an approach to rewrite the rare BKs to more commonly searched keywords, without compromising the original BKs intent, which increases the coverage and depth of PS ads and thus it delivers higher monetization power. In general, we first find the relevant web documents pertaining to the BKs and then extract common keywords using the web doc title and its summary snippets. Experimental results show the effectiveness of proposed algorithm in rewriting rare BKs and consequently providing us with a significant improvement in recall and thereby revenue.

A Lipschitz Exploration-Exploitation Scheme for Bayesian Optimization

The problem of optimizing unknown costly-to-evaluate functions has been studied extensively in the context of Bayesian optimization. Algorithms in this field aim to find the optimizer of the function by requesting only a few function evaluations at carefully selected locations. An ideal algorithm should maintain a perfect balance between exploration (probing unexplored areas) and exploitation (focusing on promising areas) within the given evaluation budget. In this paper, we assume the unknown function is Lipschitz continuous. Leveraging the Lipschitz property, we propose an algorithm with a distinct exploration phase followed by an exploitation phase. The exploration phase aims to select samples that shrink the search space as much as possible, while the exploitation phase focuses on the reduced search space and selects samples closest to the optimizer. We empirically show that the proposed algorithm significantly outperforms the baseline algorithms.