Truthful, Transparent and Fair Data Collection Mechanisms

An important prerequisite for developing trustworthy artificial intelligence is high quality data. Crowdsourcing has emerged as a popular method of data collection in the past few years. However, there is always a concern about the quality of the data thus collected. This thesis addresses two major challenges in collecting high quality data from a crowd: 1) how to incentivize crowd workers to report accurate data; 2) how to ensure that the data collection mechanism is transparent and fair.

We first propose two novel peer-consistency mechanisms for crowdsourcing: the Deep Bayesian Trust (DBT) mechanism and the Personalized Peer Truth Serum (PPTS). The DBT mechanism incentivizes workers to report accurate answers for objective questions with discrete ground truth answers. It is useful, for example, in collecting labels for supervised machine learning tasks. The mechanism ensures dominant uniform strategy incentive compatibility and fair rewards to the workers. The PPTS incentivizes workers to truthfully report their personal data (for example, body measurements). Since data is personal in nature, the tasks can not be shared between two workers. We show that when individuals report combinations of multiple personal data attributes, the correlation between them can be exploited to find peers and provide guarantees on the incentive compatibility of the mechanism.

We next address the transparency issue of data collection. Smart contracts often rely on a trusted third party (oracle) to get correct information about real-world events. We show how peer-consistency mechanisms can be used to build decentralized, trustless and transparent data oracles on blockchain. We derive conditions under which a peer-consistency incentive mechanism can be used to acquire truthful information from an untrusted and self-interested crowd, even when the crowd has outside incentives to provide wrong information. We also show how to implement the peer-consistency mechanisms in Ethereum. We discuss various non-trivial issues that arise in implementing peer-consistency mechanisms in Ethereum, suggest several optimizations to reduce gas cost and provide empirical analysis.

Finally, we address the problem of fair data collection from a crowd. Sharing economy platforms such as Airbnb and Uber face a major challenge in the form of peer-to-peer discrimination based on sensitive personal attributes such as race and gender. We show that how a peer-consistency incentive mechanism can be used to encourage users to go against common bias and provide a truthful rating about others, obtained through a more careful and deeper evaluation. In situations where an incentive mechanism canât be implemented, we show that a simple post-processing approach can also be used to correct bias in the reputation scores, while minimizing loss in the useful information provided by the scores. We also address the problem of fair and diverse data collection from a crowd under budget constraints. We propose a novel algorithm which maximizes the expected accuracy of the collected data, while ensuring that the errors satisfy desired notions of fairness w.r.t sensitive attributes.

Graph Representation Learning with Optimal Transport: Analysis and Applications

Effrosyni Simou

In several machine learning settings, the data of interest are well described by graphs. Examples include data pertaining to transportation networks or social networks. Further, biological data, such as proteins or molecules, lend themselves well to graph-structured descriptions. In such settings, it is desirable to design representation learning algorithms that can take into account the information captured by the underlying structure. This thesis focuses on providing new methods to ingrain geometrical information in end-to-end deep learning architectures for graphs through the use of elements from Optimal Transport theory.First, we consider the setting where the data can be described by a fixed graph. In this setting, each datapoint corresponds to a node of the graph and the edges among nodes signify their pairwise relations. We propose an autoencoder architecture that consists of a linear layer in the encoder and a novel Wasserstein barycentric layer at the decoder. Our proposed barycentric layer takes into account the underlying geometry through the diffusion distance of the graph and provides a way to obtain structure-aware, non-linear interpolations, which lead to directly interpretable node embeddings that are stable to perturbations of the graph structure. Further, the embeddings obtained with our proposed autoencoder achieve competitive or superior results compared to state-of-the-art methods in node classification tasks.Second, we consider the setting where the data consist of multiple graphs. In that setting, the graphs can be of varying size and, therefore, a global pooling operation is needed in order to obtain representations of fixed size and enable end-to-end learning with deep networks. We propose a global pooling operation that optimally preserves the statistical properties of the representations. In order to do so, we take into account the geometry of the representation space and introduce a global pooling layer that minimizes the Wasserstein distance between a graph representation and its pooled counterpart, of fixed size. This is achieved by performing a Wasserstein gradient flow with respect to the pooled representation. Our proposed method demonstrates promising results in the task of graph classification.Overall, in this thesis we provide new methods for incorporating geometrical information in end-to-end deep learning architectures for graph structured data. We believe that our proposals will contribute to the development of algorithms that learn meaningful representations and that take fully into account the geometry of the data under consideration.

EPFL2022

Graph-based image representation learning

Renata Khasanova

Though deep learning (DL) algorithms are very powerful for image processing tasks, they generally require a lot of data to reach their full potential. Furthermore, there is no straightforward way to impose various properties, given by the prior knowledge about the target task, on the features extracted by a DL model. Therefore, in this thesis we propose several techniques that rely on the power of graph representations to embed prior knowledge inside the learning process. This allows to reduce the solution space and leads to faster optimization convergence and higher accuracy in the representation learning. In our first work, inspired by the ability of a human to correctly classify rotated, shifted or flipped objects, we propose an algorithm that permits to inherently encode invariance to isometric transformations of objects in an image. Our DL architecture is based on graph representations and consists of three novel layers, which we refer to as graph convolutional, dynamic pooling and statistical layers. Our experiments on the image classification tasks show that our network correctly recognizes isometrically transformed objects even though such types of transformation are not seen by the network at training time. Standard DL techniques are typically not able to succeed in solving such a problem without extensive data augmentation. Then, we propose to exploit the properties of graph-based approaches to efficiently process images with various types of projective geometry. In particular, we are interested in increasingly popular omnidirectional cameras, which have a 360 degree field of view. Despite their effectiveness, such cameras create images with specific geometric properties, which require special techniques for efficient processing. We propose an efficient way of adjusting the weights of the graph edges to adapt the filter responses to the geometric image properties introduced by omnidirectional cameras. Our experiments prove that using the proposed graph with properly adjusted edge weights permits to reach better performance as compared to using regular grid graph with equal weights. Finally, the approach described above relies on the isotropic filters, which work well within our transformation invariant architecture for image classification. However, for other problems (e.g. image compression) or even when used without dynamic pooling and statistical layers that are defined within the proposed architecture, these filters are unable to efficiently encode the information about the object. Thus, we introduce a different technique based on anisotropic filters that adapt their shape and size according to the omnidirectional image geometry. The main advantage of this approach compared to the previous one is the ability to encode the orientation of an image pattern, which is important for various tasks such as image compression. Our experiments show that our approach adapts to different image projective geometries and achieves state-of-the-art performance on image classification and compression tasks. Overall we propose several methods, which combine the power of DL and graph signal processing towards incorporating prior information about the target task inside the optimization procedure. We hope that the research efforts presented in this thesis will help the development of efficient DL algorithms that can use various types of prior knowledge to make them efficient even when the available training data is scarce.

EPFL2019

Crowdsourcing with Fairness, Diversity and Budget Constraints

Boi Faltings, Naman Goel

Recent studies have shown that the labels collected from crowdworkers can be discriminatory with respect to sensitive attributes such as gender and race. This raises questions about the suitability of using crowdsourced data for further use, such as for training machine learning algorithms. In this work, we address the problem of fair and diverse data collection from a crowd under budget constraints. We propose a novel algorithm which maximizes the expected accuracy of the collected data, while ensuring that the errors satisfy desired notions of fairness. We provide guarantees on the performance of our algorithm and show that the algorithm performs well in practice through experiments on a real dataset.