A variety of statistical methods for noun compound anMysis are implemented and compared. The results support two main conclusions. First, the use of conceptual association not only enables a broad coverage, but also improves the accuracy. Second, an analysis model based on dependency grammar is substantially more accurate than one based on deepest constituents, even though the latter is more prevalent in the literature.
A system making optimal use of available information in incremental language comprehension might be expected to use linguistic knowledge together with current input to revise beliefs about previous input. Under some circumstances, such an error-correction capability might induce comprehenders to adopt grammatical analyses that are inconsistent with the true input.
The process which has resulted in this book began many decades ago when,
as an undergraduate student, I found myself asking the question, ‘What did
the Romans think they were doing when they created the Roman Empire?’
For many years this question lurked in the background of my thoughts
as I worked on Roman history more generally and on Roman Spain in
particular, not least because itwas not clear tome howsuch a questionmight
be answered.What follows is, I hope, if not an answer, at least a contribution
Women’s education and reproductive health have come to be seen in recent years as the
most effective channels for influencing fertility. In Sections 4-5 I provide an outline of the
theoretical and empirical reasons why they are so seen. It is an interesting analytical feature of
education and reproductive health that they can be studiedwithin a frameworkwhere households
make decisions in isolation of other households.
Another key figure in the history of the diamond trade was German-born businessman and financier
Ernest Oppenheimer. After he established the Anglo American Corporation, he bought De Beers shares
whenever they came up for sale. By 1927 he was one of the most significant shareholders of the company;
he was later named chairman. Under his leadership De Beers evolved into a global diamond empire.
We investigate the empirical behavior of ngram discounts within and across domains. When a language model is trained and evaluated on two corpora from exactly the same domain, discounts are roughly constant, matching the assumptions of modiﬁed Kneser-Ney LMs. However, when training and test corpora diverge, the empirical discount grows essentially as a linear function of the n-gram count. We adapt a Kneser-Ney language model to incorporate such growing discounts, resulting in perplexity improvements over modiﬁed Kneser-Ney and Jelinek-Mercer baselines. ...
We have developed a system that generates evaluative arguments that are tailored to the user, properly arranged and concise. We have also developed an evaluation framework in which the effectiveness of evaluative arguments can be measured with real users. This paper presents the results of a formal experiment we have performed in our framework to verify the influence of argument conciseness on argument effectiveness In the remainder of the paper, we first describe a computational framework for generating evaluative arguments at different levels of conciseness. ...
We describe a corpus-based investigation of proposals in dialogue. First, we describe our DR/compliant coding scheme and report our inter-coder reliability results. Next, we test several hypotheses about what constitutes a well-formed proposal. 1 Introduction we report our findings .on tracking agreement. 2 Tracking Agreement Our corpus consists of 24 computer-mediated dialogues 1 in which two participants collaborate on a simple task of buying furniture for the living and dining rooms of a house (a variant of the task in (Walker, 1993)). ...
Our feature films are currently the source of a substantial portion of our revenue. We derive
revenue from our distributors’ worldwide exploitation of our feature films in theaters and in ancillary
markets such as home entertainment, digital and pay and free broadcast television. In addition, we
earn revenue from the licensing and merchandising of our films and characters in markets around the
Fedwire securities are processed individually, in much the same way that
Fedwire funds transfers are processed, and participants initiate securi-
ties transfers in the same manner, using either a computer connection or
the telephone. When the Federal Reserve receives a request to transfer a
security, for example as a result of the sale of securities, it determines that
the security is held in safekeeping for the institution requesting the transfer
and withdraws the security from the institution’s safekeeping account.
In the general field of Environmental Psychology an increasing number of studies propose that subjects’ general well-being
can be significantly increased as a result of contact with environments considered to have high aesthetic value. The present
study has attempted to study the possible effects of the contemplation of everyday landscapes on citizens’ emotional well-
being, identifying some of the main affective responses associated with aesthetic judgements of urban landscapes.
Web search engines today typically show results as a list of titles and short snippets that summarize how the retrieved documents are related to the query. However, recent research suggests that longer summaries can be preferable for certain types of queries. This paper presents empirical evidence that judges can predict appropriate search result summary lengths, and that perceptions of search result quality can be affected by varying these result lengths. These ﬁndings have important implications for search results presentation, especially for natural language queries. ...
This paper proposes to solve the bottleneck of finding training data for word sense disambiguation (WSD) in the domain of web queries, where a complete set of ambiguous word senses are unknown. In this paper, we present a combination of active learning and semi-supervised learning method to treat the case when positive examples, which have an expected word sense in web search result, are only given. The novelty of our approach is to use “pseudo negative examples” with reliable confidence score estimated by a classifier trained with positive and unlabeled examples.
In this paper we present a new approach to controlling the behaviour of a natural language generation system by correlating internal decisions taken during free generation of a wide range of texts with the surface stylistic characteristics of the resulting outputs, and using the correlation to control the generator. This contrasts with the generate-andtest architecture adopted by most previous empirically-based generation approaches, offering a more efficient, generic and holistic method of generator control. ...
This paper describes an empirical study of the “Information Synthesis” task, deﬁned as the process of (given a complex information need) extracting, organizing and inter-relating the pieces of information contained in a set of relevant documents, in order to obtain a comprehensive, non redundant report that satisﬁes the information need.
The link between Önancial asset prices and macro variables has become a popular
Öeld of the economic research over the past decades. Many studies, mostly applied
on the United States, have shown that the term spread, measured as the di§erence
between yields on longer maturity bonds and money market interest rates, has
predicted macro variables more accurately compared with other Önancial asset
classes. Results concerning the ability of stock prices, usually in the form of
broad-based indices, in predicting such variables have been mixed.
Compounded words are a challenge for NLP applications such as machine translation (MT). We introduce methods to learn splitting rules from monolingual and parallel corpora. We evaluate them against a gold standard and measure their impact on performance of statistical MT systems. Results show accuracy of 99.1% and performance gains for MT of 0.039 BLEU on a German-English noun phrase translation task.
We present the results from a series of experiments aimed at uncovering the discourse structure of man-machine communication in natural language (Wizard of Oz experiments). The results suggest the existence of different classes of dialogue situations, requiring computational discourse representations of various complexity. Important factors seem to be the number of different permissible tasks in the system and to what extent the system takes initiative in the dialogue.
The relationship between political democracy and economic growth has been a center of debate in the past fifty years. A corpus of cross-country research has shown that the theoretical divide on the impact of democratic versus authoritarian regimes on growth is matched by ambiguous empirical results, resulting in a consensus of an inconclusive relationship. Through this paper we challenge this consensus.