Targeting Optimization for Internet Advertising by Learning from Logged Bandit Feedback

Gasparini, Margherita; Nuara, Alessandro; Trovò, Francesco; Gatti, Nicola; Restelli, Marcello

doi:10.1109/ijcnn.2018.8489092

Cited by 13 publications

(7 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From an application point of view, only a few works are present to deal with regret minimization and best arm identification for specific SPNB scenarios. In the Internet advertising management field, a method to select the most profitable time slot during the day has been presented in [12]. Nonetheless, this method provides suggestions in an offline fashion, exploiting the information provided from historical data, not including any procedure to include a newly discovered piece of information.…”

Section: Related Workmentioning

confidence: 99%

Adapting Bandit Algorithms for Settings with Sequentially Available Arms

Gabrielli¹,

Trovò²,

Antonelli³

2021

Preprint

View full text Add to dashboard Cite

Although the classical version of the Multi-Armed Bandits (MAB) framework has been applied successfully to several practical problems, in many real-world applications, the possible actions are not presented to the learner simultaneously, such as in the Internet campaign management and environmental monitoring settings. Instead, in such applications, a set of options is presented sequentially to the learner within a time span, and this process is repeated throughout a time horizon. At each time, the learner is asked whether to select the proposed option or not. We define this scenario as the Sequential Pull/No-pull Bandit setting, and we propose a meta-algorithm, namely Sequential Pull/No-pull for MAB (Seq), to adapt any classical MAB policy to better suit this setting for both the regret minimization and best-arm identification problems. By allowing the selection of multiple arms within a round, the proposed meta-algorithm gathers more information, especially in the first rounds, characterized by a high uncertainty in the arms estimate value. At the same time, the adapted algorithms provide the same theoretical guarantees as the classical policy employed. The Seq meta-algorithm was extensively tested and compared with classical MAB policies on synthetic and real-world datasets from advertising and environmental monitoring applications, highlighting its good empirical performances.

show abstract

Section: Related Workmentioning

confidence: 99%

Adapting Bandit Algorithms for Settings with Sequentially Available Arms

Gabrielli¹,

Trovò²,

Antonelli³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The campaigns distinguish for the ad, channel, and target. In real-world applications, the set of campaigns and the spending plan can be optimized from data, e.g., by setting up campaigns with specific targets and adopting a different cumulative daily budget for every day [31]. 5 For the sake of presentation, from now on, we set one day as the unitary temporal step of our algorithms.…”

Section: Optimization Problem Formulationmentioning

confidence: 99%

Online Joint Bid/Daily Budget Optimization of Internet Advertising Campaigns

Nuara¹,

Trovò²,

Gatti³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…The derivatives of loss function L(θ as ) w.r.t. parameters θ as can be presented as: ∇ θ as L(θ as ) = y as t − Q(s t , a as t ; θ as ) ∇ θ as Q(s t , a as t ; θ as ) (13) where y as t = r t (s t , a as t ) + γQ * s t +1 , BS Q * (s t +1 , A as t +1 ; θ as T ) ; θ as T . The operation Q(s t , A as t ) looks through the candidate ad set {a ad t +1 } and all locations {a loc t +1 } (including the location of inserting no ad).…”

Section: The Optimization Taskmentioning

confidence: 99%

“…The second group is real-time bidding (RTB), which allows an advertiser to bid each ad impression in a very short time slot. Ad selection task is typically modeled as multi-armed bandit problem supposing that arms are iid, feedback is immediate and environments are stationary [13,19,23,25,31,32]. The problem of online advertising with budget constraints and variable costs is studied in MAB setting [10], where pulling the arms of bandit results in random rewards and spends random costs.…”

Section: Related Workmentioning

confidence: 99%

Jointly Learning to Recommend and Advertise

Zhao,

Zheng,

Yang

et al. 2020

Preprint

View full text Add to dashboard Cite

Online recommendation and advertising are two major income channels for online recommendation platforms (e.g. e-commerce and news feed site). However, most platforms optimize recommending and advertising strategies by different teams separately via different techniques, which may lead to suboptimal overall performances. To this end, in this paper, we propose a novel two-level reinforcement learning framework to jointly optimize the recommending and advertising strategies, where the first level generates a list of recommendations to optimize user experience in the long run; then the second level inserts ads into the recommendation list that can balance the immediate advertising revenue from advertisers and the negative influence of ads on long-term user experience. To be specific, first level tackles high combinatorial action space problem that selects a subset items from the large item space; while the second level determines three internally related tasks, i.e., (i) whether to insert an ad, and if yes, (ii) the optimal ad and (iii) the optimal location to insert. The experimental results based on real-world data demonstrate the effectiveness of the proposed framework.

show abstract

Targeting Optimization for Internet Advertising by Learning from Logged Bandit Feedback

Cited by 13 publications

References 11 publications

Adapting Bandit Algorithms for Settings with Sequentially Available Arms

Adapting Bandit Algorithms for Settings with Sequentially Available Arms

Online Joint Bid/Daily Budget Optimization of Internet Advertising Campaigns

Jointly Learning to Recommend and Advertise

Contact Info

Product

Resources

About