Jorge González

Structure and catalytic properties of the most complex intergrown zeolite ITQ-39 determined by electron crystallography

Porous materials such as zeolites contain well-defined pores in molecular dimensions and have important industrial applications in catalysis, sorption and separation. Aluminosilicates with intersecting 10- and 12-ring channels are particularly interesting as selective catalysts. Many porous materials, especially zeolites, form only nanosized powders and some are intergrowths of different structures, making structure determination very challenging. Here, we report the atomic structures of an aluminosilicate zeolite family, ITQ-39, solved from nanocrystals only a few unit cells in size by electron crystallography. ITQ-39 is an intergrowth of three different polymorphs, built from the same layer but with different stacking sequences. ITQ-39 contains stacking faults and twinning with nano-sized domains, being the most complex zeolite ever solved. The unique structure of ITQ-39, with a three-dimensional intersecting pairwise 12-ring and 10-ring pore system, makes it a promising catalyst for converting naphtha into diesel fuel, a process of emerging interest for the petrochemical industry.

INTEGRATED HANDWRITING RECOGNITION AND INTERPRETATION USING FINITE-STATE MODELS

The interpretation of handwritten sentences is carried out using a holistic approach in which both text image recognition and the interpretation itself are tightly integrated. Conventional approaches follow a serial, first-recognition then-interpretation scheme which cannot adequately use semantic–pragmatic knowledge to recover from recognition errors. Stochastic finite-sate transducers are shown to be suitable models for this integration, permitting a full exploitation of the final interpretation constraints. Continuous-density hidden Markov models are embedded in the edges of the transducer to account for lexical and morphological constraints. Robustness with respect to stroke vertical variability is achieved by integrating tangent vectors into the emission densities of these models. Experimental results are reported on a syntax-constrained interpretation task which show the effectiveness of the proposed approaches. These results are also shown to be comparatively better than those achieved with other conventional, N-gram-based techniques which do not take advantage of full integration.

A New Aluminosilicate Molecular Sieve with a System of Pores between Those of ZSM-5 and Beta Zeolite

A new aluminosilicate zeolite (ITQ-39) has been synthesized. This is an extensively faulted structure with very small domains that makes the structure elucidation very difficult. However, a combination of adsorption spectroscopy and reactivity studies with selected probe molecules suggests that the pore structure of ITQ-39 is related to that of Beta zeolite, with a three-directional channel system with large pores (12-MR), but with an effective pore diameter between those of Beta and ZSM-5, or a three-directional channel system with interconnected large (12-MR) and medium pores (10-MR). The pore topology of ITQ-39 is very attractive for catalysis and shows excellent results for the preparation of cumene by alkylation of benzene, while it can be a promising additive for FCC.

Synthesis design and structure of a multipore zeolite with interconnected 12- and 10-MR channels.

A new molecular sieve, ITQ-38, containing interconnected large and medium pores in its structure has been synthesized. The rational combination of dicationic piperidine-derivative molecules as organic structure directing agents (OSDAs) with germanium and boron atoms in alkaline media has allowed the synthesis of ITQ-38 zeolite. High-resolution transmission electron microscopy (HRTEM) has been used to elucidate the framework topology of ITQ-38, revealing the presence of domains of perfect ITQ-38 crystals as well as very small areas containing nanosized ITQ-38/ITQ-22 intergrowths. The structure of ITQ-38 is highly related to ITQ-22 and the recently described polymorph C of ITQ-39 zeolite. It shares a common building layer with ITQ-22 and contains the same building unit as the polymorph C of ITQ-39. All three structures present similar framework density, 16.1 T atoms/1000 Å(3).

A Syntactic Pattern Recognition Approach to Computer Assisted Translation

It is a fact that current methodologies for automatic translation cannot be expected to produce high quality translations. An alternative approach is to use them as an aid to manual translation. We focus on a possible way to help human translators: to interactively provide completions for the parts of the sentences already translated. We explain how finite state transducers can be used for this task and show experiments in which the keystrokes needed to translate printer manuals were reduced to nearly 25% of the original.

Offline Recognition of Syntax-Constrained Cursive Handwritten Text

The problem of continuous handwritten text (CHT) recognition using standard continuous speech recognition technology is considered. Main advantages of this approach are a) system development is completely based on well understood training techniques and b) no segmentation of sentence or line images into characters or words is required, neither in the training nor in the recognition phases. Many recent papers address this problem in a similar way. Our work aims at contributing to this trend in two main aspects: i) We focus on the recognition of individual, isolated characters using the very same technology as for CHT recognition in order to tune essential representation parameters. The results are themselves interesting since they are comparable with state-of-the-art results on the same standard OCR database. And ii) all the work (except for the image processing and feature extraction steps) is strictly based on a well known and widely available standard toolkit for continuous speech recognition.

Learning finite state transducers using bilingual phrases

Statistical Machine Translation is receiving more and more attention every day due to the success that the phrase-based alignment models are obtaining. However, despite their power, state-of-the-art systems using these models present a series of disadvantages that lessen their effectiveness in working environments where temporal or spacial computational resources are limited. A finite-state framework represents an interesting alternative because it constitutes an efficient paradigm where quality and realtime factors are properly integrated in order to build translation devices that may be of help for their potential users. Here, we describe a way to use the bilingual information in a phrase-based model in order to implement a phrase-based ngram model using finite state transducers. It will be worth the trouble due to the notable decrease in computational requirements that finite state transducers present in practice with respect to the use of some well-known stack-decoding algorithms. Results for the French-English EuroParl benchmark corpus from the 2006 Workshop on Machine Translation of the ACL are reported.

SisHiTra : A Hybrid Machine Translation System from Spanish to Catalan

In the current European scenario, characterized by the coexistence of communities writing and speaking a great variety of languages, machine translation has become a technology of capital importance. In areas of Spain and of other countries, coofficiality of several languages implies producing several versions of public information. Machine translation between all the languages of the Iberian Peninsula and from them into English will allow for a better integration of Iberian linguistic communities among them and inside Europe. The purpose of this paper is to show a machine translation system from Spanish to Catalan that deals with text input. In our approach, both deductive (linguistic) and inductive (corpus-based) methodologies are combined in an homogeneous and efficient framework: finite-state transducers. Some preliminary results show the interest of the proposed architecture.

GREAT: A Finite-State Machine Translation Toolkit Implementing a Grammatical Inference Approach for Transducer Inference (GIATI)

GREAT is a finite-state toolkit which is devoted to Machine Translation and that learns structured models from bilingual data. The training procedure is based on grammatical inference techniques to obtain stochastic transducers that model both the structure of the languages and the relationship between them. The inference of grammars from natural language causes the models to become larger when a less restrictive task is involved; even more if a bilingual modelling is being considered. GREAT has been successful to implement the GIATI learning methodology, using different scalability issues to be able to deal with corpora of high volume of data. This is reported with experiments on the EuroParl corpus, which is a state-of-the-art task in Statistical Machine Translation.

Review of Coding Techniques Applied to Remote Sensing

With the aim of obtaining a valid compression method for remote sensing and geographic information systems, and because comparisons among the different available techniques are not always performed in a sufficiently fair manner, we are currently developing a framework for evaluating several still image coding techniques.