Off-policy learning, referring to the procedure of policy optimization with access only to logged feedback data, has shown importance in various real-world applications, such as search engines, recommender systems, and etc. While the ground-truth logging policy, which generates the logged data, is usually unknown, previous work simply takes its estimated value in off-policy learning, ignoring both high bias and high variance resulted from such an estimator, especially on samples with small and inaccurately estimated logging probabilities. In this work, we explicitly model the uncertainty in the estimated logging policy and propose a Uncertainty-aware Inverse Propensity Score estimator (UIPS) for improved off-policy learning. Experiment results on synthetic and three real-world recommendation datasets demonstrate the advantageous sample efficiency of the proposed UIPS estimator against an extensive list of state-of-the-art baselines.
Meta-reinforcement learning (meta-RL) aims to quickly solve new tasks by leveraging knowledge from prior tasks. However, previous studies often assume a single mode homogeneous task distribution, ignoring possible structured heterogeneity among tasks. Leveraging such structures can better facilitate knowledge sharing among related tasks and thus improve sample efficiency. In this paper, we explore the structured heterogeneity among tasks via clustering to improve meta-RL. We develop a dedicated exploratory policy to discover task structures via divide-and-conquer. The knowledge of the identified clusters helps to narrow the search space of task-specific information, leading to more sample efficient policy adaptation. Experiments on various MuJoCo tasks showed the proposed method can unravel cluster structures effectively in both rewards and state dynamics, proving strong advantages against a set of state-of-the-art baselines.
Recommendation systems aim to predict users' feedback on items not exposed to them. Confounding bias arises due to the presence of unmeasured variables (e.g., the socio-economic status of a user) that can affect both a user's exposure and feedback. Existing methods either (1) make untenable assumptions about these unmeasured variables or (2) directly infer latent confounders from users' exposure. However, they cannot guarantee the identification of counterfactual feedback, which can lead to biased predictions. In this work, we propose a novel method, i.e., identifiable deconfounder (iDCF), which leverages a set of proxy variables (e.g., observed user features) to resolve the aforementioned non-identification issue. The proposed iDCF is a general deconfounded recommendation framework that applies proximal causal inference to infer the unmeasured confounders and identify the counterfactual feedback with theoretical guarantees. Extensive experiments on various real-world and synthetic datasets verify the proposed method's effectiveness and robustness.
Content creators compete for exposure on recommendation platforms, and such strategic behavior leads to a dynamic shift over the content distribution. However, how the creators' competition impacts user welfare and how the relevance-driven recommendation influences the dynamics in the long run are still largely unknown. This work provides theoretical insights into these research questions. We model the creators' competition under the assumptions that: 1) the platform employs an innocuous top-$K$ recommendation policy; 2) user decisions follow the Random Utility model; 3) content creators compete for user engagement and, without knowing their utility function in hindsight, apply arbitrary no-regret learning algorithms to update their strategies. We study the user welfare guarantee through the lens of Price of Anarchy and show that the fraction of user welfare loss due to creator competition is always upper bounded by a small constant depending on $K$ and randomness in user decisions; we also prove the tightness of this bound. Our result discloses an intrinsic merit of the myopic approach to the recommendation, i.e., relevance-driven matching performs reasonably well in the long run, as long as users' decisions involve randomness and the platform provides reasonably many alternatives to its users.
Accuracy and diversity have long been considered to be two conflicting goals for recommendations. We point out, however, that as the diversity is typically measured by certain pre-selected item attributes, e.g., category as the most popularly employed one, improved diversity can be achieved without sacrificing recommendation accuracy, as long as the diversification respects the user's preference about the pre-selected attributes. This calls for a fine-grained understanding of a user's preferences over items, where one needs to recognize the user's choice is driven by the quality of the item itself, or the pre-selected attributes of the item. In this work, we focus on diversity defined on item categories. We propose a general diversification framework agnostic to the choice of recommendation algorithms. Our solution disentangles the learnt user representation in the recommendation module into category-independent and category-dependent components to differentiate a user's preference over items from two orthogonal perspectives. Experimental results on three benchmark datasets and online A/B test demonstrate the effectiveness of our solution in improving both recommendation accuracy and diversity. In-depth analysis suggests that the improvement is due to our improved modeling of users' categorical preferences and refined ranking within item categories.
In recent years, machine learning has achieved impressive results across different application areas. However, machine learning algorithms do not necessarily perform well on a new domain with a different distribution than its training set. Domain Adaptation (DA) is used to mitigate this problem. One approach of existing DA algorithms is to find domain invariant features whose distributions in the source domain are the same as their distribution in the target domain. In this paper, we propose to let the classifier that performs the final classification task on the target domain learn implicitly the invariant features to perform classification. It is achieved via feeding the classifier during training generated fake samples that are similar to samples from both the source and target domains. We call these generated samples domain-agnostic samples. To accomplish this we propose a novel variation of generative adversarial networks (GAN), called the MiddleGAN, that generates fake samples that are similar to samples from both the source and target domains, using two discriminators and one generator. We extend the theory of GAN to show that there exist optimal solutions for the parameters of the two discriminators and one generator in MiddleGAN, and empirically show that the samples generated by the MiddleGAN are similar to both samples from the source domain and samples from the target domain. We conducted extensive evaluations using 24 benchmarks; on the 24 benchmarks, we compare MiddleGAN against various state-of-the-art algorithms and outperform the state-of-the-art by up to 20.1\% on certain benchmarks.
Graph contrastive learning (GCL), as an emerging self-supervised learning technique on graphs, aims to learn representations via instance discrimination. Its performance heavily relies on graph augmentation to reflect invariant patterns that are robust to small perturbations; yet it still remains unclear about what graph invariance GCL should capture. Recent studies mainly perform topology augmentations in a uniformly random manner in the spatial domain, ignoring its influence on the intrinsic structural properties embedded in the spectral domain. In this work, we aim to find a principled way for topology augmentations by exploring the invariance of graphs from the spectral perspective. We develop spectral augmentation which guides topology augmentations by maximizing the spectral change. Extensive experiments on both graph and node classification tasks demonstrate the effectiveness of our method in self-supervised representation learning. The proposed method also brings promising generalization capability in transfer learning, and is equipped with intriguing robustness property under adversarial attacks. Our study sheds light on a general principle for graph topology augmentation.
Conversational recommender systems (CRS) dynamically obtain the user preferences via multi-turn questions and answers. The existing CRS solutions are widely dominated by deep reinforcement learning algorithms. However, deep reinforcement learning methods are often criticised for lacking interpretability and requiring a large amount of training data to perform. In this paper, we explore a simpler alternative and propose a decision tree based solution to CRS. The underlying challenge in CRS is that the same item can be described differently by different users. We show that decision trees are sufficient to characterize the interactions between users and items, and solve the key challenges in multi-turn CRS: namely which questions to ask, how to rank the candidate items, when to recommend, and how to handle negative feedback on the recommendations. Firstly, the training of decision trees enables us to find questions which effectively narrow down the search space. Secondly, by learning embeddings for each item and tree nodes, the candidate items can be ranked based on their similarity to the conversation context encoded by the tree nodes. Thirdly, the diversity of items associated with each tree node allows us to develop an early stopping strategy to decide when to make recommendations. Fourthly, when the user rejects a recommendation, we adaptively choose the next decision tree to improve subsequent questions and recommendations. Extensive experiments on three publicly available benchmark CRS datasets show that our approach provides significant improvement to the state of the art CRS methods.
Bandit algorithms have become a reference solution for interactive recommendation. However, as such algorithms directly interact with users for improved recommendations, serious privacy concerns have been raised regarding its practical use. In this work, we propose a differentially private linear contextual bandit algorithm, via a tree-based mechanism to add Laplace or Gaussian noise to model parameters. Our key insight is that as the model converges during online update, the global sensitivity of its parameters shrinks over time (thus named dynamic global sensitivity). Compared with existing solutions, our dynamic global sensitivity analysis allows us to inject less noise to obtain $(\epsilon, \delta)$-differential privacy with added regret caused by noise injection in $\tilde O(\log{T}\sqrt{T}/\epsilon)$. We provide a rigorous theoretical analysis over the amount of noise added via dynamic global sensitivity and the corresponding upper regret bound of our proposed algorithm. Experimental results on both synthetic and real-world datasets confirmed the algorithm's advantage against existing solutions.
Advertisements (ads) are an innate part of search engine business models. Advertisers are willing to pay search engines to promote their content to a prominent position in the search result page (SERP). This raises concerns about the search engine manipulation effect (SEME): the opinions of users can be influenced by the way search results are presented. In this work, we investigate the connection between SEME and sponsored content in the health domain. We conduct a series of user studies in which participants need to evaluate the effectiveness of different non-prescription natural remedies for various medical conditions. We present participants SERPs with different intentionally created biases towards certain viewpoints, with or without sponsored content, and ask them to evaluate the effectiveness of the treatment only based on the information presented to them. We investigate two types of sponsored content: 1. Direct marketing ads that directly market the product without expressing an opinion about its effectiveness, and 2. Indirect marketing ads that explicitly advocate the product's effectiveness on the condition in the query. Our results reveal a significant difference between the influence on users from these two ad types. Though direct marketing ads are mostly skipped by users, they can tilt users decision making towards more positive viewpoints. Indirect marketing ads affect both the users' examination behaviour and their perception of the treatment's effectiveness. We further discover that the contrast between the indirect marketing ads and the viewpoint presented in the organic search results plays an important role in users' decision-making. When the contrast is high, users exhibit a strong preference towards a negative viewpoint, and when the contrast is low or none, users exhibit preference towards a more positive viewpoint.