In this paper, we present an algorithm for learning a generative model of natural language sentences together with their formal meaning representations with hierarchical structures. The model is applied to the task of mapping sentences to hierarchical representations of their underlying meaning. We introduce dynamic programming techniques for efficient training and decoding. In experiments, we demonstrate that the model, when coupled with a discriminative reranking technique, achieves state-of-the-art performance when tested on two publicly available corpora. The generative model degrades robustly when presented with instances that are different from those seen in training. This allows a notable improvement in recall compared to previous models. 1

We present a novel commentator system that learns language from sportscasts of simulated soccer games. The system learns to parse and generate commentaries without any engineered knowledge about the English language. Training is done using only ambiguous supervision in the form of textual human commentaries and simulation states of the soccer games. The system simultaneously tries to establish correspondences between the commentaries and the simulation states as well as build a translation model. We also present a novel algorithm, Iterative Generation Strategy Learning (IGSL), for deciding which events to comment on. Human evaluations of the generated commentaries indicate they are of reasonable quality compared to human commentaries. 1.

Abstract. Extracting knowledge from text has long been a goal of AI. Initial approaches were purely logical and brittle. More recently, the availability of large quantities of text on the Web has led to the development of machine learning approaches. However, to date these have mainly extracted ground facts, as opposed to general knowledge. Other learning approaches can extract logical forms, but require supervision and do not scale. In this paper we present an unsupervised approach to extracting semantic networks from large volumes of text. We use the TextRunner system [1] to extract tuples from text, and then induce general concepts and relations from them by jointly clustering the objects and relational strings in the tuples. Our approach is defined in Markov logic using four simple rules. Experiments on a dataset of two million tuples show that it outperforms three other relational clustering approaches, and extracts meaningful semantic networks. 1

We present results for a system designed to perform Open Knowledge Extraction, based on a tradition of compositional language processing, as applied to a large collection of text derived from the Web. Evaluation through manual assessment shows that well-formed propositions of reasonable quality, representing general world knowledge, given in a logical form potentially useable for inference, may be extracted in high volume from arbitrary input sentences. We compare these results with those obtained in recent work on Open Information Extraction, indicating with some examples the quite different kinds of output obtained by the two approaches. Finally, we observe that portions of the extracted knowledge are comparable to results of recent work on class attribute extraction. 1

Compositional question answering begins by mapping questions to logical forms, but training a semantic parser to perform this mapping typically requires the costly annotation of the target logical forms. In this paper, we learn to map questions to answers via latent logical forms, which are induced automatically from question-answer pairs. In tackling this challenging learning problem, we introduce a new semantic representation which highlights a parallel between dependency syntax and efficient evaluation of logical forms. On two standard semantic parsing benchmarks (GEO and JOBS), our system obtains the highest published accuracies, despite requiring no annotated logical forms. 1

This paper addresses the problem of learning to map sentences to logical form, given training data consisting of natural language sentences paired with logical representations of their meaning. Previous approaches have been designed for particular natural languages or specific meaning representations; here we present a more general method. The approach induces a probabilistic CCG grammar that represents the meaning of individual words and defines how these meanings can be combined to analyze complete sentences. We use higher-order unification to define a hypothesis space containing all grammars consistent with the training data, and develop an online learning algorithm that efficiently searches this space while simultaneously estimating the parameters of a log-linear parsing model. Experiments demonstrate high accuracy on benchmark data sets in four languages with two different meaning representations. 1

In this paper, we describe an approach to formally assess whether an organization conforms to a body of regulation. Conformance is cast as a model checking question where the regulation is represented in a logic that is evaluated against an abstract model representing the operations of an organization. Regulatory bases are large and complex, and the long term goal of our work is to be able to use natural language processing (NLP) to assist in the translation of regulation to logic. We argue that the translation of regulation to logic should proceed one sentence at a time. A challenge in taking this approach arises from the fact that sentences is regulation often refer to others. We motivate the need for a formal representation of regulation to accomodate references between statements. We briefy describe a logic in which statements can refer to and reason about others. We then discuss preliminary work on using NLP to assist in the translation of regulatory sentences into logic.

We consider the problem of learning context-dependent mappings from sentences to logical form. The training examples are sequences of sentences annotated with lambda-calculus meaning representations. We develop an algorithm that maintains explicit, lambda-calculus representations of salient discourse entities and uses a context-dependent analysis pipeline to recover logical forms. The method uses a hidden-variable variant of the perception algorithm to learn a linear model used to select the best analysis. Experiments on context-dependent utterances from the ATIS corpus show that the method recovers fully correct logical forms with 83.7% accuracy. 1

Current approaches for semantic parsing take a supervised approach requiring a considerable amount of training data which is expensive and difficult to obtain. This supervision bottleneck is one of the major difficulties in scaling up semantic parsing. We argue that a semantic parser can be trained effectively without annotated data, and introduce an unsupervised learning algorithm. The algorithm takes a self training approach driven by confidence estimation. Evaluated over Geoquery, a standard dataset for this task, our system achieved 66 % accuracy, compared to 80 % of its fully supervised counterpart, demonstrating the promise of unsupervised approaches for this task. 1

Machine learning offers a range of tools for training systems from data, but these methods are only as good as the underlying representation. This paper proposes to acquire representations for machine learning by reading text written to accommodate human learning. We propose a novel form of semantic analysis called reading to learn, where the goal is to obtain a high-level semantic abstract of multiple documents in a representation that facilitates through a generative model that requires no labeled data, instead leveraging repetition across multiple documents. The semantic abstract is converted into a transformed feature space for learning, resulting in improved generalization on a relational learning task. 1