Inference Of Attributes Research Articles

Privacy preservation in sensor-based Human Activity Recognition through autoencoders for low-power IoT devices

Human activity recognition is increasingly recognized as a key task in many applications. However, gathering data from the variety of sensors commonly available on end devices risks compromising user’s privacy when signals are transmitted to more powerful computing units for inference offloading. It is therefore important to design and implement strategies that could prevent privacy breaches without impairing the capability of the system of recognizing activity patterns, and by taking into account the energy constraints of low-power devices. In this work, we propose an energy-aware approach aimed at preserving the privacy of users during inference of human activities. The proposed method is based on a deep learning autoencoder trained to process the signal in order to remove the most sensitive information regarding privacy attributes, without significantly impacting classification accuracy. We also perform a thorough architecture’s parameter tuning of the designed system to enable its implementation on a low-power platform, which we also characterize in terms of energy expenditure. Experimental results show that this system is capable of effectively transforming the signal in order to prevent the inference of sensitive attributes (i.e. weight, height, age, and gender) and it can be conveniently implemented on a constrained embedded system at different levels of the trade-off between accuracy and energy consumption. Indeed, a complete obfuscation of sensitive attributes can be achieved at the cost of a marginal reduction in classification accuracy (5% at most), with an expenditure of around 165 mJ for an execution time of around 30ms needed during the signal transformation step.

Task-Agnostic Privacy-Preserving Representation Learning for Federated Learning against Attribute Inference Attacks

Federated learning (FL) has been widely studied recently due to its property to collaboratively train data from different devices without sharing the raw data. Nevertheless, recent studies show that an adversary can still be possible to infer private information about devices' data, e.g., sensitive attributes such as income, race, and sexual orientation. To mitigate the attribute inference attacks, various existing privacy-preserving FL methods can be adopted/adapted. However, all these existing methods have key limitations: they need to know the FL task in advance, or have intolerable computational overheads or utility losses, or do not have provable privacy guarantees. We address these issues and design a task-agnostic privacy-preserving presentation learning method for FL (TAPPFL) against attribute inference attacks. TAPPFL is formulated via information theory. Specifically, TAPPFL has two mutual information goals, where one goal learns task-agnostic data representations that contain the least information about the private attribute in each device's data, and the other goal ensures the learnt data representations include as much information as possible about the device data to maintain FL utility. We also derive privacy guarantees of TAPPFL against worst-case attribute inference attacks, as well as the inherent tradeoff between utility preservation and privacy protection. Extensive results on multiple datasets and applications validate the effectiveness of TAPPFL to protect data privacy, maintain the FL utility, and be efficient as well. Experimental results also show that TAPPFL outperforms the existing defenses.

Comprehensive Privacy Analysis on Federated Recommender System Against Attribute Inference Attacks

Comprehensive Privacy Analysis on Federated Recommender System Against Attribute Inference Attacks

Multi-Channel Hypergraph Collaborative Filtering with Attribute Inference

In the field of collaborative filtering, attribute information is often integrated to improve recommendations. However, challenges remain unaddressed. Firstly, existing data modeling methods often fall short of appropriately handling attribute information. Secondly, attribute data are often sparse and can potentially impact recommendation performance due to the challenge of incomplete correspondence between the attribute information and the recommendations. To tackle these challenges, we propose a hypergraph collaborative filtering with attribute inference (HCFA) framework, which segregates attribute and user behavior information into distinct channels and leverages hypergraphs to capture high-order correlations among vertices, offering a more natural approach to modeling. Furthermore, we introduce behavior-based attribute confidence (BAC) for assessing the reliability of inferred attributes concerning the corresponding behaviors and update the most credible portions to enhance recommendation quality. Extensive experiments conducted on three public benchmarks demonstrate the superiority of our model. It consistently outperforms other state-of-the-art approaches, with ablation experiments further confirming the effectiveness of our proposed method.

Securing User Privacy in Cloud-Based Whiteboard Services Against Health Attribute Inference Attacks

Securing User Privacy in Cloud-Based Whiteboard Services Against Health Attribute Inference Attacks

Human intuition as a defense against attribute inference

Attribute inference—the process of analyzing publicly available data in order to uncover hidden information—has become a major threat to privacy, given the recent technological leap in machine learning. One way to tackle this threat is to strategically modify one’s publicly available data in order to keep one’s private information hidden from attribute inference. We evaluate people’s ability to perform this task, and compare it against algorithms designed for this purpose. We focus on three attributes: the gender of the author of a piece of text, the country in which a set of photos was taken, and the link missing from a social network. For each of these attributes, we find that people’s effectiveness is inferior to that of AI, especially when it comes to hiding the attribute in question. Moreover, when people are asked to modify the publicly available information in order to hide these attributes, they are less likely to make high-impact modifications compared to AI. This suggests that people are unable to recognize the aspects of the data that are critical to an inference algorithm. Taken together, our findings highlight the limitations of relying on human intuition to protect privacy in the age of AI, and emphasize the need for algorithmic support to protect private information from attribute inference.

Measuring Fairness Under Unawareness of Sensitive Attributes: A Quantification-Based Approach

Algorithms and models are increasingly deployed to inform decisions about people, inevitably affecting their lives. As a consequence, those in charge of developing these models must carefully evaluate their impact on different groups of people and favour group fairness, that is, ensure that groups determined by sensitive demographic attributes, such as race or sex, are not treated unjustly. To achieve this goal, the availability (awareness) of these demographic attributes to those evaluating the impact of these models is fundamental. Unfortunately, collecting and storing these attributes is often in conflict with industry practices and legislation on data minimisation and privacy. For this reason, it can be hard to measure the group fairness of trained models, even from within the companies developing them. In this work, we tackle the problem of measuring group fairness under unawareness of sensitive attributes, by using techniques from quantification, a supervised learning task concerned with directly providing group-level prevalence estimates (rather than individual-level class labels). We show that quantification approaches are particularly suited to tackle the fairness-under-unawareness problem, as they are robust to inevitable distribution shifts while at the same time decoupling the (desirable) objective of measuring group fairness from the (undesirable) side effect of allowing the inference of sensitive attributes of individuals. More in detail, we show that fairness under unawareness can be cast as a quantification problem and solved with proven methods from the quantification literature. We show that these methods outperform previous approaches to measure demographic parity in five experimental protocols, corresponding to important challenges that complicate the estimation of classifier fairness under unawareness.

MLI: A Multi-level Inference Mechanism for User Attributes in Social Networks

In the social network, each user has attributes for self-description called user attributes, which are semantically hierarchical. Attribute inference has become an essential way for social platforms to realize user classifications and targeted recommendations. Most existing approaches mainly focus on the flat inference problem neglecting the semantic hierarchy of user attributes, which will cause serious inconsistency in multi-level tasks. In this article, we propose a multi-level model MLI, where information propagation part collects attribute information by mining the global graph structure, and the attribute correction part realizes the mutual correction between different levels of attributes. Further, we put forward the concept of generalized semantic tree, a way of representing the hierarchical structure of user attributes, whose nodes are allowed to have multiple parent nodes unlike the regular tree. Both regular and generalized semantic trees are commonly used in practice, and can be handled by our model. Besides, by making the inference start from sub-networks with sufficient attribute information, we design a “Ripple” algorithm to improve the efficiency and effectiveness of our model. For evaluation purposes, we conduct extensive verification experiments on DBLP datasets. The experimental results show the superior effect of MLI, compared with the state-of-the-art methods.

Basin-scale subsurface characterization using single-station teleseismic receiver function analysis

Teleseismic receiver function (RF) analysis traditionally has targeted deep earth structures in the context of observational seismology, allowing the detection of converted seismic phases originating from major boundaries of impedance. Recent efforts to adapt this passive-source seismic imaging technique to basin-scale subsurface characterization have retrieved primary basin geometries and key stratigraphic boundaries by relying on densely deployed broadband seismometer networks. While accurate interpretation of RF time series on Earth benefitted from complementary subsurface constraints (e.g., core and well logs), lack of such resources and instrumental limitations must be considered in the context of planetary geophysics, where passive-source imaging may emerge as a key frontier exploration technique. Thus, it is important to establish an analysis framework in which observed RF signals are interpreted by relying on first-order suppositions about the physical properties of the underlying subsurface. In this work, we seek to simulate this approach in single-station scenarios and qualitatively examine the baseline information inferable from the RF time series. Our results suggest that signals observed in the sedimentary interval can be reasonably attributed to major impedance changes expected from the generalized lithostratigraphy of the local subsurface, including the transition to the underlying crust. We also find patterns of anisotropy-related directional variations in high-frequency signals as well as unique frequency-dependent responses likely associated with the depth and vertical dimension of converting boundaries. Together, these seismic observables enable inference of various geologic attributes within the underlying sedimentary unit, proving it to be a critical tool in future efforts of planetary exploration.

Interactions between attributions and beliefs at trial-by-trial level: Evidence from a novel computer game task

Inferring causes of the good and bad events that we experience is part of the process of building models of our own capabilities and of the world around us. Making such inferences can be difficult because of complex reciprocal relationships between attributions of the causes of particular events, and beliefs about the capabilities and skills that influence our role in bringing them about. Abnormal causal attributions have long been studied in connection with psychiatric disorders, notably depression and paranoia; however, the mechanisms behind attributional inferences and the way they can go awry are not fully understood. We administered a novel, challenging, game of skill to a substantial population of healthy online participants, and collected trial-by-trial time series of both their beliefs about skill and attributions about the causes of the success and failure of real experienced outcomes. We found reciprocal relationships that provide empirical confirmation of the attribution-self representation cycle theory. This highlights the dynamic nature of the processes involved in attribution, and validates a framework for developing and testing computational accounts of attribution-belief interactions.

Infer-AVAE: An attribute inference model based on adversarial variational autoencoder

User attributes, such as gender and education, face severe incompleteness in social networks. Attribute inference aims to infer users’ missing attribute labels based on observed data to make this valuable data usable for downstream tasks like user profiling and personalized recommendation. Recently, variational autoencoder (VAE), an end-to-end deep generative model, has shown promising performance by handling the problem in a semi-supervised way. However, VAEs can easily suffer from over-fitting and over-smoothing when applied to attribute inference. Specifically, VAE implemented with multi-layer perceptron (MLP) can only reconstruct input data but fail to infer missing parts. While using the trending graph neural networks (GNNs) as encoder has the problem that GNNs aggregate redundant information from the neighborhood and generate indistinguishable user representations, known as over-smoothing. In this paper, we propose an attribute Inference model based on Adversarial VAE (Infer-AVAE) to cope with these issues. Specifically, to overcome over-smoothing, Infer-AVAE unifies MLP and GNNs in the encoder to learn positive and negative latent representations respectively. Meanwhile, an adversarial network is trained to distinguish the two representations, and GNNs are trained to aggregate less noise for more robust representations through adversarial training. Finally, to relieve over-fitting, mutual information constraint is introduced as a regularizer for the decoder to make better use of auxiliary information in representations and generate outputs not limited by observations. We evaluate our model on four real-world social network datasets, and experimental results demonstrate that our model averagely outperforms baselines by 7.0% in accuracy.

A Comparative Study of Demographic Attribute Inference in Twitter

Social media platforms have become a major gateway to receive and analyze public opinions. Understandingusers can provide invaluable context information of their social media posts and significantly improve traditional opinion analysis models. Demographic attributes,such as ethnicity, gender, age, among others,have been extensively applied to characterize social mediausers. While studies have shown that user groups formed by demographic attributes can have coherent opinions towards political issues, these attributes are often not explicitly coded by users through their profiles.Previous work has demonstrated the effectiveness of different user signals such as users’ posts and names in determining demographic attributes. Yet, these efforts mostly evaluate linguistic signals from users’ postsand train models from artificially balanced datasets. In this paper, we propose a comprehensive list of user signals:self-descriptions and posts aggregated from users’ friends and followers, users’ profile images, and users’ names.We provide a comparative study of these signalsside-by-side in the tasks on inferring three major demographic attributes, namely ethnicity, gender, and age.We utilize a realistic unbalanced datasets that share similar demographic makeups in Twitter for training modelsand evaluation experiments. Our experiments indicate that self-descriptions provide the strongest signal for ethnicity and age inference and clearly improve the overall performance when combined with tweets. Profile images for gender inference have the highest precision score with overall score close to the best result in our setting. This suggests that signals in self descriptions and profile images have potentials to facilitate demographic attribute inferences in Twitter, and are promising for future investigation.

Homophily and Latent Attribute Inference: Inferring Latent Attributes of Twitter Users from Neighbors

In this paper, we extend existing work on latent attribute inference by leveraging the principle of homophily: we evaluate the inference accuracy gained by augmenting the user features with features derived from the Twitter profiles and postings of her friends. We consider three attributes which have varying degrees of assortativity: gender, age, and political affiliation. Our approach yields a significant and robust increase in accuracy for both age and political affiliation, indicating that our approach boosts performance for attributes with moderate to high assortativity. Furthermore, different neighborhood subsets yielded optimal performance for different attributes, suggesting that different subsamples of the user's neighborhood characterize different aspects of the user herself. Finally, inferences using only the features of a user's neighbors outperformed those based on the user's features alone. This suggests that the neighborhood context alone carries substantial information about the user.

How Dispersive Opinions Affect Consumer Decisions: Endowment Effect Guides Attributional Inferences

Electronic word-of-mouth (eWOM) dispersion, reflecting the extent of reviewers’ opinion divergence regarding a product, determines consumer decisions. Drawing upon the endowment effect and attribution literature, this research proposes that the endowment effect mediates the influence of eWOM dispersion on attributional inferences, and the display formats of eWOM dispersion (“horizontal bar chart” vs. “eWOM content”) moderate the mediating influence of endowment effect on attributional inferences of eWOM dispersion. Three complementary studies reveal three key insights in this setting. First, when consumers already having an emotional attachment to a product and subsequently encountering dispersive eWOM, the higher the level of eWOM dispersion, the stronger the endowment effect. Secondly, In such circumstance, the stronger endowment effect promotes more reviewer-related attribution inference of high-dispersion eWOM, and the weaker endowment effect promotes more product-related attribution inference of low-dispersion eWOM. Third, the positive influence of eWOM dispersion on endowment effect is stronger for eWOM dispersion in the complexity or disfluency display format (i.e. “reviewers’ impression of product”) than for eWOM dispersion in the simply of fluency display format (i.e. “horizontal bar chart”), more importantly, the mediating influence of endowment effect on attributional inferences is moderated by the display formats of eWOM dispersion, and this findings persists even when product category (hedonic vs. utilitarian) is taken into consideration. The theoretical implications for the eWOM dispersion and attribution literature and practical implications for online retailers are discussed.

Attribute substitution, earlier-generation economic approaches and behavioural economics

ABSTRACT The present paper examines attribute substitution in terms of both heuristics and attribution theory in social psychology. Alternative ‘old’ approaches in psychology were special because they considered choice in terms that were similar to attributional inference in social psychology and anticipated limitations of the static heuristics and biases approach. Attribute substitution plays an important role in reflective reasoning as an independent entity relative to rationality and produces good decisions, even those are sub-optimal and may be influenced by primitive processes such as intuition and habits. The attribute-substitution research enhances our understanding of intense shifts in economic expectations. It makes possible the conceptual incorporation of substantive theoretical constructions, which are inherent in psychological propensities of alternative, and mainly, heterodox approaches in economics such as Post Keynesian, Institutional and Austrian economics. This research agenda paves the way for possible forms of synthesis of those theories with behavioural economics.

Mapping Temperate Forest Phenology Using Tower, UAV, and Ground-Based Sensors

Phenology is a distinct marker of the impacts of climate change on ecosystems. Accordingly, monitoring the spatiotemporal patterns of vegetation phenology is important to understand the changing Earth system. A wide range of sensors have been used to monitor vegetation phenology, including digital cameras with different viewing geometries mounted on various types of platforms. Sensor perspective, view-angle, and resolution can potentially impact estimates of phenology. We compared three different methods of remotely sensing vegetation phenology—an unoccupied aerial vehicle (UAV)-based, downward-facing RGB camera, a below-canopy, upward-facing hemispherical camera with blue (B), green (G), and near-infrared (NIR) bands, and a tower-based RGB PhenoCam, positioned at an oblique angle to the canopy—to estimate spring phenological transition towards canopy closure in a mixed-species temperate forest in central Virginia, USA. Our study had two objectives: (1) to compare the above- and below-canopy inference of canopy greenness (using green chromatic coordinate and normalized difference vegetation index) and canopy structural attributes (leaf area and gap fraction) by matching below-canopy hemispherical photos with high spatial resolution (0.03 m) UAV imagery, to find the appropriate spatial coverage and resolution for comparison; (2) to compare how UAV, ground-based, and tower-based imagery performed in estimating the timing of the spring phenological transition. We found that a spatial buffer of 20 m radius for UAV imagery is most closely comparable to below-canopy imagery in this system. Sensors and platforms agree within +/− 5 days of when canopy greenness stabilizes from the spring phenophase into the growing season. We show that pairing UAV imagery with tower-based observation platforms and plot-based observations for phenological studies (e.g., long-term monitoring, existing research networks, and permanent plots) has the potential to scale plot-based forest structural measures via UAV imagery, constrain uncertainty estimates around phenophases, and more robustly assess site heterogeneity.

Attribute Substitution, Earlier-Generation Economic Approaches and Behavioural Economics

The present paper examines attribute substitution in terms of both heuristics and attribution theory in social psychology. Alternative “old” approaches in psychology were special because they considered choice in terms that were similar to attributional inference in social psychology and anticipated limitations of the static heuristics and biases approach. Attribute substitution plays an important role in reflective reasoning as an independent entity relative to rationality and produces good decisions, even those are sub-optimal and may be influenced by primitive processes such as intuition and habits. The attribute-substitution research enhances our understanding of intense shifts in economic expectations. It makes possible the conceptual incorporation of substantive theoretical constructions, which are inherent in psychological propensities of alternative, and mainly, heterodox approaches in economics such as Post Keynesian, Institutional and Austrian economics. This research agenda paves the way for possible forms of synthesis of those theories with behavioral economics.

TweetFit: Fusing Multiple Social Media and Sensor Data for Wellness Profile Learning

Wellness is a widely popular concept that is commonly applied to fitness and self-help products or services. Inference of personal wellness-related attributes, such as body mass index or diseases tendency, as well as understanding of global dependencies between wellness attributes and users' behavior is of crucial importance to various applications in personal and public wellness domains. Meanwhile, the emergence of social media platforms and wearable sensors makes it feasible to perform wellness profiling for users from multiple perspectives. However, research efforts on wellness profiling and integration of social media and sensor data are relatively sparse, and this study represents one of the first attempts in this direction. Specifically, to infer personal wellness attributes, we proposed multi-source individual user profile learning framework named "TweetFit". "TweetFit" can handle data incompleteness and perform wellness attributes inference from sensor and social media data simultaneously. Our experimental results show that the integration of the data from sensors and multiple social media sources can substantially boost the wellness profiling performance.

Latent Attribute Inference of Users in Social Media with Very Small Labeled Dataset

Latent Attribute Inference of Users in Social Media with Very Small Labeled Dataset

Inference of demographic attributes based on mobile phone usage patterns and social network topology

Mobile phone usage provides a wealth of information, which can be used to better understand the demographic structure of a population. In this paper, we focus on the population of Mexican mobile phone users. We first present an observational study of mobile phone usage according to gender and age groups. We are able to detect significant differences in phone usage among different subgroups of the population. We then study the performance of different machine learning (ML) methods to predict demographic features (namely, age and gender) of unlabeled users by leveraging individual calling patterns, as well as the structure of the communication graph. We show how a specific implementation of a diffusion model, harnessing the graph structure, has significantly better performance over other node-based standard ML methods. We provide details of the methodology together with an analysis of the robustness of our results to changes in the model parameters. Furthermore, by carefully examining the topological relations of the training nodes (seed nodes) to the rest of the nodes in the network, we find topological metrics which have a direct influence on the performance of the algorithm.