More Like this

1. Context-aware Multi-stakeholder Recommender Systems

   https://doi.org/10.32473/flairs.v35i.130573  

   Arabghalizi, Tahereh ; Labrinidis, Alexandros ( May 2022 , The International FLAIRS Conference Proceedings)

   Traditional recommender systems help users find the most relevant products or services to match their needs and preferences. However, they overlook the preferences of other sides of the market (aka stakeholders) involved in the system. In this paper, we propose to use contextual bandit algorithms in multi-stakeholder platforms where a multi-sided relevance function with adjusting weights is modeled to consider the preferences of all involved stakeholders. This algorithm sequentially recommends the items based on the contextual features of users along with the priority of the stakeholders and their relevance to the items.Our extensive experimental results on a dataset consisting of MovieLens (1m), IMDB (81k+), and a synthetic dataset show that our proposed approach outperforms the baseline methods and provides a good trade-off between the satisfaction of different stakeholders over time. 

   more » « less
2. Fairness and Control of Exposure in Two-sided Markets

   https://doi.org/10.1145/3471158.3472226  

   Joachims, Thorsten ( January 2021 , ACM SIGIR International Conference on Theory of Information Retrieval)

   A large number of two-sided markets are now mediated by search and recommender systems, ranging from online retail and streaming entertainment to employment and romantic-partner matching. I will discuss in this talk how the design decisions that go into these search and recommender systems carry substantial power in shaping markets and allocating opportunity to the participants. This does not only raise legal and fairness questions, but also questions about how these systems shape incentives and the long-term effectiveness of the market. At the core of these questions lies the problem of where to rank each item, and how this affects both sides of the market. While it is well understood how to maximize the utility to the users, this talk focuses on how rankings affect the items that are being ranked. From the items perspective, the ranking system is an arbiter of exposure and thus economic opportunity. I will discuss how machine learning algorithms that follow the conventional Probability Ranking Principle [1] can lead to undesirable and unfair exposure allocation for both exogenous and endogenous reasons. Exogenous reasons often manifest themselves as biases in the training data, which then get reflected in the learned ranking policy. But even when trained with unbiased data, reasons endogenous to the system can lead to unfair or undesirable allocation of opportunity. To overcome these challenges, I will present new machine learning algorithms [2,3,4] that directly address both endogenous and exogenous factors, allowing the designer to tailor the ranking policy to be appropriate for the specific two-sided market. 

   more » « less
3. Leader–follower equilibria to examine investment decisions in grid resilience‐enhancing measures

   https://doi.org/10.1111/risa.13921  

   Lo Prete, Chiara ; Radhakrishnan, Ashish ( February 2023 , Risk Analysis)

   Large‐area, long‐duration power outages are increasingly common in the United States, and cost the economy billions of dollars each year. Building a strategy to enhance grid resilience requires an understanding of the optimal mix of preventive and corrective actions, the inefficiencies that arise when self‐interested parties make resilience investment decisions, and the conditions under which regulators may facilitate the realization of efficient market outcomes. We develop a bi‐level model to examine the mix of preventive and corrective measures that enhances grid resilience to a severe storm. The model represents a Stackelberg game between a regulated utility (leader) that may harden distribution feeders before a long‐duration outage and/or deploy restoration crews after the disruption, and utility customers with varying preferences for reliable power (followers) who may invest in backup generators. We show that the regulator's denial of cost recovery for the utility's preventive expenditures, coupled with the misalignment between private objectives and social welfare maximization, yields significant inefficiencies in the resilience investment mix. Allowing cost recovery for a higher share of the utility's capital expenditures in preventive measures, extending the time horizon associated with damage cost recovery, and adopting a storm restoration compensation mechanism shift the realized market outcome toward the efficient solution. If about one‐fifth of preventive resilience investments is approved by regulators, requiring utilities to pay a compensation of $365 per customer for a 3‐day outage (about seven times the level of compensation currently offered by US utilities) provides significant incentives toward more efficient preventive resilience investments. 

   more » « less
4. Time-sensitive POI Recommendation by Tensor Completion with Side Information

   Hui, Bo ; Yan, Da ; Ku, Wei-Shinn ( January 2022 , Proceedings of the 38th IEEE International Conference on Data Engineering (ICDE))

   Context has been recognized as an important factor to consider in personalized recommender systems. Particularly in location-based services (LBSs), a fundamental task is to recommend to a mobile user where he/she could be interested to visit next at the right time. Additionally, location-based social networks (LBSNs) allow users to share location-embedded information with friends who often co-occur in the same or nearby points-of-interest (POIs) or share similar POI visiting histories, due to the social homophily theory and Tobler’s first law of geography. So, both the time information and LBSN friendship relations should be utilized for POI recommendation. Tensor completion has recently gained some attention in time-aware recommender systems. The problem decomposes a user-item-time tensor into low-rank embedding matrices of users, items and times using its observed entries, so that the underlying low-rank subspace structure can be tracked to fill the missing entries for time-aware recommendation. However, these tensor completion methods ignore the social-spatial context information available in LBSNs, which is important for POI recommendation since people tend to share their preferences with their friends, and near things are more related than distant things. In this paper, we utilize the side information of social networks and POI locations to enhance the tensor completion model paradigm for more effective time-aware POI recommendation. Specifically, we propose a regularization loss head based on a novel social Hausdorff distance function to optimize the reconstructed tensor. We also quantify the popularity of different POIs with location entropy to prevent very popular POIs from being over-represented hence suppressing the appearance of other more diverse POIs. To address the sensitivity of negative sampling, we train the model on the whole data by treating all unlabeled entries in the observed tensor as negative, and rewriting the loss function in a smart way to reduce the computational cost. Through extensive experiments on real datasets, we demonstrate the superiority of our model over state-of-the-art tensor completion methods. 

   more » « less
5. A Ranked Bandit Approach for Multi-stakeholder Recommender Systems

   Arabghalizi, Tahereh ; Labrinidis, Alexandros ( September 2022 , Workshop of Multi-Objective Recommender Systems (MORS’22))

   Recommender systems traditionally find the most relevant products or services for users tailored to their needs or interests but they ignore the interests of the other sides of the market (aka stakeholders). In this paper, we propose to use a Ranked Bandit approach for an online multi-stakeholder recommender system that sequentially selects top 𝑘 items according to the relevance and priority of all the involved stakeholders. We presented three different criteria to consider the priority of each stakeholder when evaluating our approach. Our extensive experimental results on a movie dataset showed that the contextual multi-armed bandits with a relevance function make a higher level of satisfaction for all involved stakeholders in the long term. Keywords: Multi-stakeholder Recommender Systems; Multi-armed Bandits; Ranked Bandit; 

   more » « less