Crowdsourcing platforms use various truth discovery algorithms to aggregate annotations from multiple labelers. In an online setting, however, the main challenge is to decide whether to ask for more annotations for each item to efficiently trade off cost (i.e., the number of annotations) for quality of the aggregated annotations. In this paper, we propose a novel approach for general complex annotation (such as bounding boxes and taxonomy paths), that works in an online crowdsourcing setting. We prove that the expected average similarity of a labeler is linear in their accuracy conditional on the reported label. This enables us to infer reported label accuracy in a broad range of scenarios. We conduct extensive evaluations on real-world crowdsourcing data from Meta and show the effectiveness of our proposed online algorithms in improving the cost-quality trade-off.

Online platforms regularly rely on human annotators to make real-time operational decisions for tasks such as content moderation. While crowdsourcing models have been proposed for aggregating noisy labels, they do not generalize well when annotators produce a labels in a large space, e.g., generated from complex review trees. We study a novel crowdsourcing setting with D possible operational decisions or outcomes, but annotators produce labels in a larger space of size L > D which are mapped to decisions through a known mapping function. For content moderation, such labels can correspond to violation reasons (e.g. nudity, violence), while the space of decisions is binary: remove the content or keep it up. In this setting, it is more important to make the right decision rather than estimating the correct underlying label. Existing methods typically separate out the labels to decisions mapping from the modeling of annotators, leading to sub-optimal statistical inference efficiency and excessive computation complexity. We propose a novel confusion matrix model for each annotator that leverages this mapping. Our model is parameterized in a hierarchical manner with both population parameters shared across annotators to model shared confusions and individual parameters to admit heterogeneity among annotators. With extensive numerical experiments, we demonstrate that the proposed model substantially improves accuracy over existing methods and scales well for moderate and large L. In a real-world application on content moderation at Meta, the proposed method offers a 13% improvement in AUC over prior methods, including Meta's existing model in production.

We study a crowdsourcing setting where we need to infer the latent truth about a task given observed labels together with context in the form of a classifier score. We present Theodon, a hierarchical non-parametric Bayesian model, developed and deployed at Meta, that captures both the prevalence of label categories and the accuracy of labelers as functions of the classifier score. Theodon uses Gaussian processes to model the non-uniformity of mistakes over the range of classifier scores. For our experiments, we used data generated from integrity applications at Meta as well as public datasets. We showed that Theodon (1) obtains 1-4% improvement in AUC-PR predictions on items' true labels compared to state-of-the-art baselines for public datasets, (2) is effective as a calibration method, and (3) provides detailed insights on labelers' performances.

Large online services employ thousands of people to label content for applications such as video understanding, natural language processing, and content policy enforcement. While labelers typically reach their decisions by following a well-defined "protocol'', humans may still make mistakes. A common countermeasure is to have multiple people review the same content; however, this process is often time-intensive and requires accurate aggregation of potentially noisy decisions.

Viet-An Nguyen, Jordan Boyd-Graber, Philip Resnik, Kristina Miler. Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2015.

Markov chain Monte Carlo (MCMC) approximates the posterior distribution of latent variable models by generating many samples and averaging over them.In practice, however, it is often more convenient to cut corners, using only a single sample or following a suboptimal averaging strategy.We systematically study different strategies for averaging MCMC samples and show empirically that averaging properly leads to significant improvements in prediction.

Inspired by a two-level theory that unifies agenda setting and ideological framing, we propose supervised hierarchical latent Dirichlet allocation (SHLDA) which jointly captures documents' multi-level topic structure and their polar response variables. Our model extends the nested Chinese restaurant process to discover a tree-structured topic hierarchy and uses both per-topic hierarchical and per-word lexical regression parameters to model the response variables. Experiments in a political domain and on sentiment analysis tasks show that SHLDA improves predictive accuracy while adding a new dimension of insight into how topics under discussion are framed.

This paper aims to detect users generating spam reviews or review spammers. We identify several characteristic behaviors of review spammers and model these behaviors so as to detect the spammers. In particular, we seek to model the following behaviors. First, spammers may target specific products or product groups in order to maximize their impact. Second, they tend to deviate from the other reviewers in their ratings of products. We propose scoring methods to measure the degree of spam for each reviewer and apply them on an Amazon review dataset. We then select a subset of highly suspicious reviewers for further scrutiny by our user evaluators with the help of a web based spammer evaluation software specially developed for user evaluation experiments. Our results show that our proposed ranking and supervised methods are effective in discovering spammers and outperform other baseline method based on helpfulness votes alone. We finally show that the detected spammers have more significant impact on ratings compared with the unhelpful reviewers.

Trust reciprocity, a special form of link reciprocity, exists in many networks of trust among users. In this paper, we seek to determine the extent to which reciprocity exists in a trust network and develop quantitative models for measuring reciprocity and reciprocity related behaviors. We identify several reciprocity behaviors and their respective measures. These behavior measures can be employed for predicting if a trustee will return trust to her trustor given that the latter initiates a trust link earlier. We develop for this reciprocal trust prediction task a number of ranking method and classification methods, and evaluated them on an Epinions trust network data. Our results show that reciprocity related behaviors provide good features for both ranking and classification based methods under different parameter settings.

This paper analyzes the trustor and trustee factors that lead to inter-personal trust using a well studied Trust Antecedent framework in management science. To apply these factors to trust ranking problem in online rating systems, we derive features that correspond to each factor and develop different trust ranking models. The advantage of this approach is that features relevant to trust can be systematically derived so as to achieve good prediction accuracy. Through a series of experiments on real data from Epinions, we show that even a simple model using the derived features yields good accuracy and outperforms MoleTrust, a trust propagation based model. SVM classifiers using these features also show improvements.