To address bias issues, fair machine learning usually jointly optimizes two (or more) metrics aiming at predictive utility and fairness. However, the inherent under-representation of minorities in the data often makes the disparate impact of subpopulations less noticeable and difficult to deal with during learning. In this paper, we propose a novel adversarial reweighting method to address such disparate impact. To balance the data distribution between the majority and the minority groups, our approach prefers samples from the majority group that are closer to the minority group as evaluated by the Wasserstein distance. Theoretical analysis shows the effectiveness of our adversarial reweighting approach. Experiments demonstrate that our approach mitigates disparate impact without sacrificing classification accuracy, outperforming related state-of-the-art methods on image and tabular benchmark datasets. Code is available at https://github.com/zhaoxuan00707/wasserstein_reweight.

We present GRACIE (Graph Recalibration and Adaptive Counterfactual Inspection and Explanation), a novel approach for generative classification and counterfactual explanations of dynamically changing graph data. We study graph classification problems through the lens of generative classifiers. We propose a dynamic, self-supervised latent variable model that updates by identifying plausible counterfactuals for input graphs and recalibrating decision boundaries through contrastive optimization. Unlike prior work, we do not rely on linear separability between the learned graph representations to find plausible counterfactuals. Moreover, GRACIE eliminates the need for stochastic sampling in latent spaces and graph-matching heuristics. Our work distills the implicit link between generative classification and loss functions in the latent space, a key insight to understanding recent successes with this architecture. We further observe the inherent trade-off between validity and pulling explainee instances towards the central region of the latent space, empirically demonstrating our theoretical findings. In extensive experiments on synthetic and real-world graph data, we attain considerable improvements, reaching ~99% validity when sampling sets of counterfactuals even in the challenging setting of dynamic data landscapes.

The importance of achieving fairness in machine learning models cannot be overstated. Recent research has pointed out that fairness should be examined from a causal perspective, and several fairness notions based on the on Pearl’s causal framework have been proposed. In this paper, we construct a reweighting scheme of datasets to address causal fairness. Our approach aims at mitigating bias by considering the causal relationships among variables and incorporating them into the reweighting process. The proposed method adopts two neural networks, whose structures are intentionally used to reflect the structures of a causal graph and of an interventional graph. The two neural networks can approximate the causal model of the data, and the causal model of interventions. Furthermore, reweighting guided by a discriminator is applied to achieve various fairness notions. Experiments on real-world datasets show that our method can achieve causal fairness on the data while remaining close to the original data for downstream tasks.

As complex machine learning models are increasingly used in sensitive applications like banking, trading or credit scoring, there is a growing demand for reliable explanation mechanisms. Local feature attribution methods have become a popular technique for post-hoc and model-agnostic explanations. However, attribution methods typically assume a stationary environment in which the predictive model has been trained and remains stable. As a result, it is often unclear how local attributions behave in realistic, constantly evolving settings such as streaming and online applications. In this paper, we discuss the impact of temporal change on local feature attributions. In particular, we show that local attributions can become obsolete each time the predictive model is updated or concept drift alters the data generating distribution. Consequently, local feature attributions in data streams provide high explanatory power only when combined with a mechanism that allows us to detect and respond to local changes over time. To this end, we present CDLEEDS, a flexible and model-agnostic framework for detecting local change and concept drift. CDLEEDS serves as an intuitive extension of attribution-based explanation techniques to identify outdated local attributions and enable more targeted recalculations. In experiments, we also show that the proposed framework can reliably detect both local and global concept drift. Accordingly, our work contributes to a more meaningful and robust explainability in online machine learning.

Data streams are ubiquitous in modern business and society. In practice, data streams may evolve over time and cannot be stored indefinitely. Effective and transparent machine learning on data streams is thus often challenging. Hoeffding Trees have emerged as a state-of-the art for online predictive modelling. They are easy to train and provide meaningful convergence guarantees under a stationary process. Yet, at the same time, Hoeffding Trees often require heuristic and costly extensions to adjust to distributional change, which may considerably impair their interpretability. In this work, we revisit Model Trees for machine learning in evolving data streams. Model Trees are able to maintain more flexible and locally robust representations of the active data concept, making them a natural fit for data stream applications. Our novel framework, called Dynamic Model Tree, satisfies desirable consistency and minimality properties. In experiments with synthetic and real-world tabular streaming data sets, we show that the proposed framework can drastically reduce the number of splits required by existing incremental decision trees. At the same time, our framework often outperforms state-of-the-art models in terms of predictive quality - especially when concept drift is involved. Dynamic Model Trees are thus a powerful online learning framework that contributes to more lightweight and interpretable machine learning in data streams.

Local approximations are popular methods to scale Gaussian processes (GPs) to big data. Local approximations reduce time complexity by dividing the original dataset into subsets and training a local expert on each subset. Aggregating the experts' prediction is done assuming either conditional dependence or independence between the experts. Imposing the conditional independence assumption (CI) between the experts renders the aggregation of different expert predictions time efficient at the cost of poor uncertainty quantification. On the other hand, modeling dependent experts can provide precise predictions and uncertainty quantification at the expense of impractically high computational costs. By eliminating weak experts via a theory-guided expert selection step, we substantially reduce the computational cost of aggregating dependent experts while ensuring calibrated uncertainty quantification. We leverage techniques from the literature on undirected graphical models, using sparse precision matrices that encode conditional dependencies between experts to select the most important experts. Moreover, our approach also provides a solution to the poor uncertainty quantification in CI-based models.

Feature selection can be a crucial factor in obtaining robust and accurate predictions. Online feature selection models, however, operate under considerable restrictions; they need to efficiently extract salient input features based on a bounded set of observations, while enabling robust and accurate predictions. In this work, we introduce FIRES, a novel framework for online feature selection. The proposed feature weighting mechanism leverages the importance information inherent in the parameters of a predictive model. By treating model parameters as random variables, we can penalize features with high uncertainty and thus generate more stable feature sets. Our framework is generic in that it leaves the choice of the underlying model to the user. Strikingly, experiments suggest that the model complexity has only a minor effect on the discriminative power and stability of the selected feature sets. In fact, using a simple linear model, FIRES obtains feature sets that compete with state-of-the-art methods, while dramatically reducing computation time. In addition, experiments show that the proposed framework is clearly superior in terms of feature selection stability.

Counterfactual explanations can be obtained by identifying the smallest change made to an input vector to influence a prediction in a positive way from a user’s viewpoint; for example, from ’loan rejected’ to ’awarded’ or from ’high risk of cardiovascular disease’ to ’low risk’. Previous approaches would not ensure that the produced counterfactuals be proximate (i.e., not local outliers) and connected to regions with substantial data density (i.e., close to correctly classified observations), two requirements known as counterfactual faithfulness. Our contribution is twofold. First, drawing ideas from the manifold learning literature, we develop a framework, called C-CHVAE, that generates faithful counterfactuals. Second, we suggest to complement the catalog of counterfactual quality measures using a criterion to quantify the degree of difficulty for a certain counterfactual suggestion. Our real world experiments suggest that faithful counterfactuals come at the cost of higher degrees of difficulty.

We address the problem of latent truth discovery, LTD for short, where the goal is to discover the underlying true values of entity attributes in the presence of noisy, conflicting or incomplete information. Despite a multitude of algorithms addressing the LTD problem, only little is known about their overall performance with respect to effectiveness, efficiency and robustness. The LTD model proposed in this paper is based on Restricted Boltzmann Machines, thus coined LTD-RBM. In extensive experiments on various heterogeneous and publicly available datasets, LTD-RBM is superior to state-of-the-art LTD techniques in terms of an overall consideration of effectiveness, efficiency and robustness.

In recent years artificial neural networks have become the method of choice for many pattern recognition tasks. Despite their overwhelming success, a rigorous and easy to interpret mathematical explanation of the influence of input variables on a output produced by a neural network is still missing.

Duplicates in a dataset are multiple representations of the same real-world entity and constitute a major data quality problem. This paper investigates the problem of estimating the number and sizes of duplicate record clusters in advance and describes a sampling-based method for solving this problem. In extensive experiments, on multiple datasets, we show that the proposed method reliably estimates the number of duplicate clusters, while being highly efficient.

The Internet has enabled the creation of a growing number of large-scale knowledge bases in a variety of domains containing complementary information. Tools for automatically aligning these knowledge bases would make it possible to unify many sources of structured knowledge and answer complex queries. However, the efficient alignment of large-scale knowledge bases still poses a considerable challenge. Here, we present Simple Greedy Matching (SiGMa), a simple algorithm for aligning knowledge bases with millions of entities and facts. SiGMa is an iterative propagation algorithm that leverages both the structural information from the relationship graph and flexible similarity measures between entity properties in a greedy local search, which makes it scalable. Despite its greedy nature, our experiments indicate that SiGMa can efficiently match some of the world's largest knowledge bases with high accuracy. We provide additional experiments on benchmark datasets which demonstrate that SiGMa can outperform state-of-the-art approaches both in accuracy and efficiency.

Large amounts of graph-structured data are emerging from various avenues, ranging from natural and life sciences to social and semantic web communities. We address the problem of discovering subgraphs of entities that reflect latent topics in graph-structured data. These topics are structured meta-information providing further insights into the data. The presented approach effectively detects such topics by exploiting only the structure of the underlying graph, thus avoiding the dependency on textual labels, which are a scarce asset in prevalent graph datasets. The viability of our approach is demonstrated in experiments on real-world datasets.

The prediction accuracy of any learning algorithm highly depends on the quality of the selected features; but often, the task of feature construction and selection is tedious and nonscalable. In recent years, however, there have been numerous projects with the goal of constructing general-purpose or domain-specific knowledge bases with entity-relationship-entity triples extracted from various Web sources or collected from user communities, e.g. YAGO, DBpedia, Freebase, UMLS, etc. This paper advocates the simple and yet far-reaching idea that the structured knowledge contained in such knowledge bases can be exploited to automatically extract features for general learning tasks. We introduce an expressive graph-based language for extracting features from such knowledge bases and a theoretical framework for constructing feature vectors from the extracted features. Our experimental evaluation on different learning scenarios provides evidence that the features derived through our framework can considerably improve the prediction accuracy, especially when the labeled data at hand is sparse.

Our work aims at building probabilistic tools for constructing and maintaining large-scale knowledge bases containing entity-relationship-entity triples (statements) extracted from the Web. In order to mitigate the uncertainty inherent in information extraction and integration we propose leveraging the "wisdom of the crowds" by aggregating truth assessments that users provide about statements. The suggested method, CoBayes, operates on a collection of statements, a set of deduction rules (e.g. transitivity), a set of users, and a set of truth assessments of users about statements. We propose a joint probabilistic model of the truth values of statements and the expertise of users for assessing statements. The truth values of statements are interconnected through derivations based on the deduction rules. The correctness of a user's assessment for a given statement is modeled by linear mappings from user descriptions and statement descriptions into a common latent knowledge space where the inner product between user and statement vectors determines the probability that the user assessment for that statement will be correct. Bayesian inference in this complex graphical model is performed using mixed variational and expectation propagation message passing. We demonstrate the viability of CoBayes in comparison to other approaches, on realworld datasets and user feedback collected from Amazon Mechanical Turk.

The proliferation of knowledge-sharing communities and the advances in information extraction have enabled the construction of large knowledge bases using the RDF data model to represent entities and relationships. However, as the Web and its latently embedded facts evolve, a knowledge base can never be complete and up-to-date. On the other hand, a rapidly increasing suite of Web services provide access to timely and high-quality information, but this is encapsulated by the service interface. We propose to leverage the information that could be dynamically obtained from Web services in order to enrich RDF knowledge bases on the fly whenever the knowledge base does not suffice to answer a user query.

Many modern applications are faced with the task of knowledge discovery in entity-relationship graphs, such as domain-specific knowledge bases or social networks. Mining an "informative" subgraph that can explain the relations between k(>= 2) given entities of interest is a frequent knowledge discovery scenario on such graphs. We present MING, a principled method for extracting an informative subgraph for given query nodes. MING builds on a new notion of informativeness of nodes. This is used in a random-walk-with-restarts process to compute the informativeness of entire subgraphs.

Large graphs and networks are abundant in modern information systems: entity-relationship graphs over relational data or Web-extracted entities, biological networks, social online communities, knowledge bases, and many more. Often such data comes with expressive node and edge labels that allow an interpretation as a semantic graph, and edge weights that reflect the strengths of semantic relations between entities. Finding close relationships between a given set of two, three, or more entities is an important building block for many search, ranking, and analysis tasks. From an algorithmic point of view, this translates into computing the best Steiner trees between the given nodes, a classical NP-hard problem. In this paper, we present a new approximation algorithm, coined STAR, for relationship queries over large relationship graphs. We prove that for n query entities, STAR yields an O(log(n))-approximation of the optimal Steiner tree in pseudopolynomial run-time, and show that in practical cases the results returned by STAR are qualitatively comparable to or even better than the results returned by a classical 2-approximation algorithm. We then describe an extension to our algorithm to return the top-k Steiner trees. Finally, we evaluate our algorithm over both main-memory as well as completely diskresident graphs containing millions of nodes. Our experiments show that in terms of efficiency STAR outperforms the best state-of-the-art database methods by a large margin, and also returns qualitatively better results.

We introduce a multi-label classification model and algorithm for labeling heterogeneous networks, where nodes belong to different types and different types have different sets of classification labels. We present a graph-based approach which models the mutual influence between nodes in the network as a random walk. When viewing class labels as "colors", the random surfer is "spraying" different node types with different color palettes; hence the name Graffiti. We demonstrate the performance gains of our method by comparing it to three state-of-the-art techniques for graph-based classification.

We present ANGIE, a system that can answer user queries by combining knowledge from a local database with knowledge retrieved from Web services. If a user poses a query that cannot be answered by the local database alone, ANGIE calls the appropriate Web services to retrieve the missing information. This information is integrated seamlessly and transparently into the local database, so that the user can query and browse the knowledge base while appropriate Web services are called automatically in the background.

The Web has the potential to become the world's largest knowledge base. In order to unleash this potential, the wealth of information available on the Web needs to be extracted and organized. There is a need for new querying techniques that are simple and yet more expressive than those provided by standard keyword-based search engines. Searching for knowledge rather than Web pages needs to consider inherent semantic structures like entities (person, organization, etc.) and relationships (isA, located In, etc.). In this paper, we propose NAGA, a new semantic search engine. NAGA builds on a knowledge base, which is organized as a graph with typed edges, and consists of millions of entities and relationships extracted from Web-based corpora. A graph-based query language enables the formulation of queries with additional semantic information. We introduce a novel scoring model, based on the principles of generative language models, which formalizes several notions such as confidence, informativeness and compactness and uses them to rank query results. We demonstrate NAGA's superior result quality over state-of-the-art search engines and question answering systems.

The presence of encyclopedic Web sources, such as Wikipedia, the Internet Movie \nDatabase (IMDB), World Factbook, etc. calls for new querying techniques that \nare simple and yet more expressive than those provided by standard \nkeyword-based search engines. Searching for explicit knowledge needs to \nconsider inherent semantic structures involving entities and relationships.\n\nIn this demonstration proposal, we describe a semantic search system named \nNAGA. NAGA operates on a knowledge graph, which contains millions of entities \nand relationships derived from various encyclopedic Web sources, such as the \nones above. NAGA's graph-based query language is geared towards expressing \nqueries with additional semantic information. Its scoring model is based on the \nprinciples of generative language models, and formalizes several desiderata \nsuch as confidence, informativeness and compactness of answers.\n\nWe propose a demonstration of NAGA which will allow users to browse the \nknowledge base through a user interface, enter queries in NAGA's query language \nand tune the ranking parameters to test various ranking aspects.

In a recent paper, Martens et al. introduced a specification mechanism for XML tree languages, based on rules of the form (r,s), wherer, s are regular expressions. Sets of such rules can be interpreted in an existential or a universal fashion. An XML tree is existentially valid with respect to a rule set, if for each node there is a rule such that the root path of the node matches r and the children sequence of the node matchess. It is universally valid if each node matching r also matchess. This paper investigates the complexity of reasoning about such rule sets, in particular the satisfiability and the implication problem. Whereas, in general these reasoning problems are complete for EXPTIME, two important fragments are identified with PSPACE and PTIME complexity, respectively.

Current keyword-oriented search engines for theWorld WideWeb do not allow specifying the semantics of queries. We address this limitation with NAGA1, a new semantic search engine. NAGA builds on a large semantic knowledge base of binary relationships (facts) derived from the Web. NAGA provides a simple, yet expressive query language to query this knowledge base. The results are then ranked with an intuitive scoring mechanism. We show the effectiveness and utility of NAGA by comparing its output with that of Googleon some interesting queries.

We present YAGO, a light-weight and extensible ontology with high coverage and quality. YAGO builds on entities and relations and currently contains more than 1 million entities and 5 million facts. This includes the Is-A hierarchy as well as non-taxonomic relations between entities (such as HASONEPRIZE). The facts have been automatically extracted from Wikipedia and unified with WordNet, using a carefully designed combination of rule-based and heuristic methods described in this paper. The resulting knowledge base is a major step beyond WordNet: in quality by adding knowledge about individuals like persons, organizations, products, etc. with their semantic relationships - and in quantity by increasing the number of facts by more than an order of magnitude. Our empirical evaluation of fact correctness shows an accuracy of about 95%. YAGO is based on a logically clean model, which is decidable, extensible, and compatible with RDFS. Finally, we show how YAGO can be further extended by state-of-the-art information extraction techniques.