While many errors in machine translated texts are due to word sense ambiguity, syntactical transfer still poses problems, too. For a particular type of syntactic structure, namely, subordinate clauses, this paper distinguishes five kinds of translation difficulties: crosslanguage differences, syntactic ambiguity, functional ambiguity, stylistic preferences, and grammatical coverage.

Attention mechanism advanced state-of-the-art neural machine translation (NMT) by jointly learning to align and translate. However, attentional NMT ignores past alignment information, which leads to over-translation and under-translation problems. In response to this problem, we maintain a coverage vector to keep track of the attention history. The coverage vector is fed to the attention model ...

Agents in an open system interact with each other based on (typically, published) protocols. An agent may, however, deviate from the protocol because of its internal policies. Such deviations pose certain challenges: (1) the agent might no longer be conformant with the protocol—how do we determine if the agent is conformant? (2) the agent may no longer be able to interoperate with other agents—...

Using data from a two-year randomized pricing experiment in China, we study the impact and design of subsidy policies for weather insurance. Results show that subsidies are effective in boosting demand in the short-run but not in the longer term. Exploring the channels, we show that while subsidies increase the direct and social effects of payout experiences by enlarging coverage, they also dam...

A numerical likelihood approach to the determination of cosmic ray anisotropies is presented which offers many advantages over other approaches. It allows a wide range of statistically meaningful hypotheses to be compared even when full sky coverage is unavailable, can be readily extended in order to include measurement errors, and makes maximum unbiased use of all available information.

Fault-localization techniques that apply statistical analyses to execution data gathered from multiple tests are quite effective when a large test suite is available. However, if no test suite is available, what is the best approach to generate one? This paper investigates the fault-localization effectiveness of test suites generated according to several test-generation techniques based on comb...

While many errors in machine translated texts are due to word sense ambiguity, syntactical transfer still poses problems, too. For a particular type of syntactic structure, namely, subordinate clauses, this paper distinguishes five kinds of translation difficulties: crosslanguage differences, syntactic ambiguity, functional ambiguity, stylistic preferences, and grammatical coverage.

This paper analyzes the performance of a base station layout for a UMTS network in a densely populated city. The study is carried out using snapshot simulations of an actual city (Barcelona, Spain) with specific traffic and propagation profiles. A first layout is proposed in order to guarantee good coverage with a minimum number of base stations. This layout is approached taking into account th...

Rather than creating and storing thousands of paraphrase examples, paraphrase templates have strong representation capacity and can be used to generate many paraphrase examples. This paper describes a new template representation and generalization method. Combing a semantic dictionary, it uses multiple semantic codes to represent a paraphrase template. Using an existing search engine to extend ...

IRIT developed a new strategy in order to detect the relevant sentences that we did not try in a more general context of document retrieval but did try previously and partially in document categorization. In our approach a sentence is considered as relevant if it matches the topic with a certain level of coverage. This level of coverage depends on the category of the terms used in the texts. Th...