Daniel Lowd

Adversarial learning

Many classification tasks, such as spam filtering, intrusion detection, and terrorism detection, are complicated by an adversary who wishes to avoid detection. Previous work on adversarial classification has made the unrealistic assumption that the attacker has perfect knowledge of the classifier [2]. In this paper, we introduce the adversarial classifier reverse engineering (ACRE) learning problem, the task of learning sufficient information about a classifier to construct adversarial attacks. We present efficient algorithms for reverse engineering linear classifiers with either continuous or Boolean features and demonstrate their effectiveness using real data from the domain of spam filtering.

Naive Bayes models for probability estimation

Naive Bayes models have been widely used for clustering and classification. However, they are seldom used for general probabilistic learning and inference (i.e., for estimating and computing arbitrary joint, conditional and marginal distributions). In this paper we show that, for a wide range of benchmark datasets, naive Bayes models learned using EM have accuracy and learning time comparable to Bayesian networks with context-specific independence. Most significantly, naive Bayes inference is orders of magnitude faster than Bayesian network inference using Gibbs sampling and belief propagation. This makes naive Bayes models a very attractive alternative to Bayesian networks for general probability estimation, particularly in large or real-time domains.

Efficient Weight Learning for Markov Logic Networks

Markov logic networks (MLNs) combine Markov networks and first-order logic, and are a powerful and increasingly popular representation for statistical relational learning. The state-of-the-art method for discriminative learning of MLN weights is the voted perceptron algorithm, which is essentially gradient descent with an MPE approximation to the expected sufficient statistics (true clause counts). Unfortunately, these can vary widely between clauses, causing the learning problem to be highly ill-conditioned, and making gradient descent very slow. In this paper, we explore several alternatives, from per-weight learning rates to second-order methods. In particular, we focus on two approaches that avoid computing the partition function: diagonal Newton and scaled conjugate gradient. In experiments on standard SRL datasets, we obtain order-of-magnitude speedups, or more accurate models given comparable learning times.

Unifying Logical and Statistical AI

Intelligent agents must be able to handle the complexity and uncertainty of the real world. Logical AI has focused mainly on the former, and statistical AI on the latter. Markov logic combines the two by attaching weights to first-order formulas and viewing them as templates for features of Markov networks. Inference algorithms for Markov logic draw on ideas from satisfiability, Markov chain Monte Carlo and knowledge-based model construction. Learning algorithms are based on the voted perceptron, pseudo-likelihood and inductive logic programming. Markov logic has been successfully applied to a wide variety of problems in natural language understanding, vision, computational biology, social networks and others, and is the basis of the open-source Alchemy system.

Learning Markov Network Structure with Decision Trees

Traditional Markov network structure learning algorithms perform a search for globally useful features. However, these algorithms are often slow and prone to finding local optima due to the large space of possible structures. Ravikumar et al. recently proposed the alternative idea of applying L1 logistic regression to learn a set of pair wise features for each variable, which are then combined into a global model. This paper presents the DTSL algorithm, which uses probabilistic decision trees as the local model. Our approach has two significant advantages: it is more efficient, and it is able to discover features that capture more complex interactions among the variables. Our approach can also be seen as a method for converting a dependency network into a consistent probabilistic model. In an extensive empirical evaluation on 13 datasets, our algorithm obtains comparable accuracy to three standard structure learning algorithms while running 1-4 orders of magnitude faster.

Just Add Weights: Markov Logic for the Semantic Web

In recent years, it has become increasingly clear that the vision of the Semantic Web requires uncertain reasoning over rich, first-order representations. Markov logic brings the power of probabilistic modeling to first-order logic by attaching weights to logical formulas and viewing them as templates for features of Markov networks. This gives natural probabilistic semantics to uncertain or even inconsistent knowledge bases with minimal engineering effort. Inference algorithms for Markov logic draw on ideas from satisfiability, Markov chain Monte Carlo and knowledge-based model construction. Learning algorithms are based on the conjugate gradient algorithm, pseudo-likelihood and inductive logic programming. Markov logic has been successfully applied to problems in entity resolution, link prediction, information extraction and others, and is the basis of the open-source Alchemy system.

Learning to Refine an Automatically Extracted Knowledge Base Using Markov Logic

A number of text mining and information extraction projects such as Text Runner and NELL seek to automatically build knowledge bases from the rapidly growing amount of information on the web. In order to scale to the size of the web, these projects often employ ad hoc heuristics to reason about uncertain and contradictory information rather than reasoning jointly about all candidate facts. In this paper, we present a Markov logic-based system for cleaning an extracted knowledge base. This allows a scalable system such as NELL to take advantage of joint probabilistic inference, or, conversely, allows Markov logic to be applied to a web scale problem. Our system uses only the ontological constraints and confidence values of the original system, along with human-labeled data if available. The labeled data can be used to calibrate the confidence scores from the original system or learn the effectiveness of individual extraction patterns. To achieve scalability, we introduce a neighborhood grounding method that only instantiates the part of the network most relevant to the given query. This allows us to partition the knowledge cleaning task into tractable pieces that can be solved individually. In experiments on NELLs knowledge base, we evaluate several variants of our approach and find that they improve both F1 and area under the precision-recall curve.

Weakly Supervised Tweet Stance Classification by Relational Bootstrapping.

Supervised stance classification, in such domains as Congressional debates and online forums, has been a topic of interest in the past decade. Approaches have evolved from text classification to structured output prediction, including collective classification and sequence labeling. In this work, we investigate collective classification of stances on Twitter, using hinge-loss Markov random fields (HLMRFs). Given the graph of all posts, users, and their relationships, we constrain the predicted post labels and latent user labels to correspond with the network structure. We focus on a weakly supervised setting, in which only a small set of hashtags or phrases is labeled. Using our relational approach, we are able to go beyond the stance-indicative patterns and harvest more stance-indicative tweets, which can also be used to train any linear text classifier when the network structure is not available or is costly.

On the hardness of evading combinations of linear classifiers

An increasing number of machine learning applications involve detecting the malicious behavior of an attacker who wishes to avoid detection. In such domains, attackers modify their behavior to evade the classifier while accomplishing their goals as efficiently as possible. The attackers typically do not know the exact classifier parameters, but they may be able to evade it by observing the classifiers behavior on test instances that they construct. For example, spammers may learn the most effective ways to modify their spams by sending test emails to accounts they control. This problem setting has been formally analyzed for linear classifiers with discrete features and convex-inducing classifiers with continuous features, but never for non-linear classifiers with discrete features. In this paper, we extend previous ACRE learning results to convex polytopes representing unions or intersections of linear classifiers. We prove that exponentially many queries are required in the worst case, but that when the features used by the component classifiers are disjoint, previous attacks on linear classifiers can be adapted to efficiently attack them. In experiments, we further analyze the cost and number of queries required to attack different types of classifiers. These results move us closer to a comprehensive understanding of the relative vulnerability of different types of classifiers to malicious adversaries.

Ontology Matching with Knowledge Rules

Ontology matching is the process of automatically determining the semantic equivalences between the concepts of two ontologies. Most ontology matching algorithms are based on two types of strategies: terminology-based strategies, which align concepts based on their names or descriptions, and structure-based strategies, which exploit concept hierarchies to find the alignment. In many domains, there is additional information about the relationships of concepts represented in various ways, such as Bayesian networks, decision trees, and association rules. We propose to use the similarities between these relationships to find more accurate alignments. We accomplish this by defining soft constraints that prefer alignments where corresponding concepts have the same local relationships encoded as knowledge rules. We use a probabilistic framework to integrate this new knowledge-based strategy with standard terminology-based and structure-based strategies. Furthermore, our method is particularly effective in identifying correspondences between complex concepts. Our method achieves better F-score than the state-of-the-art on three ontology matching domains.