Extension Schemes for the Alignment Model of English-Malayalam Statistical Machine Translator

This study adds to the body of knowledge in research of ICTs in organizations by exploring the relevance of alignment between ICT users and managers responsible for the ICT strategy and ICT project processes. Alignment research is usually conducted at an organizational analytical level, but this study explores alignment between individuals by addressing ICT managers and ICT users, considering both the organizational and individual perspectives. Data was collected by interviewing ICT users and the CIO in a Norwegian entrepreneur corporation. Using a grounded theory analytical approach, the findings provide support for an emerging User Management Alignment Model (UMAM) where the outcome is better ICT strategy and project processes.

Download Full-text

Unsupervised Sub-tree Alignment for Tree-to-Tree Translation

Journal of Artificial Intelligence Research ◽

10.1613/jair.4033 ◽

2013 ◽

Vol 48 ◽

pp. 733-782 ◽

Cited By ~ 3

Author(s):

T. Xiao ◽

J. Zhu

Keyword(s):

Machine Translation ◽

State Of The Art ◽

Experimental Results ◽

Alignment Accuracy ◽

Syntactic Structures ◽

Tree Alignment ◽

Matrix Encoding ◽

Word Alignments ◽

Alignment Model ◽

Translation Systems

This article presents a probabilistic sub-tree alignment model and its application to tree-to-tree machine translation. Unlike previous work, we do not resort to surface heuristics or expensive annotated data, but instead derive an unsupervised model to infer the syntactic correspondence between two languages. More importantly, the developed model is syntactically-motivated and does not rely on word alignments. As a by-product, our model outputs a sub-tree alignment matrix encoding a large number of diverse alignments between syntactic structures, from which machine translation systems can efficiently extract translation rules that are often filtered out due to the errors in 1-best alignment. Experimental results show that the proposed approach outperforms three state-of-the-art baseline approaches in both alignment accuracy and grammar quality. When applied to machine translation, our approach yields a +1.0 BLEU improvement and a -0.9 TER reduction on the NIST machine translation evaluation corpora. With tree binarization and fuzzy decoding, it even outperforms a state-of-the-art hierarchical phrase-based system.

Download Full-text

A Systematic Comparison of Various Statistical Alignment Models

Computational Linguistics ◽

10.1162/089120103321337421 ◽

2003 ◽

Vol 29 (1) ◽

pp. 19-51 ◽

Cited By ~ 841

Author(s):

Franz Josef Och ◽

Hermann Ney

Keyword(s):

Training Algorithm ◽

Dice Coefficient ◽

Design Decisions ◽

First Order ◽

Model Yield ◽

Alignment System ◽

Word Alignments ◽

Alignment Model ◽

Statistical Alignment

We present and compare various methods for computing word alignments using statistical or heuristic models. We consider the five alignment models presented in Brown, Della Pietra, Della Pietra, and Mercer (1993), the hidden Markov alignment model, smoothing techniques, and refinements. These statistical models are compared with two heuristic models based on the Dice coefficient. We present different methods for combining word alignments to perform a symmetrization of directed statistical alignment models. As evaluation criterion, we use the quality of the resulting Viterbi alignment compared to a manually produced reference alignment. We evaluate the models on the German-English Verbmobil task and the French-English Hansards task. We perform a detailed analysis of various design decisions of our statistical alignment system and evaluate these on training corpora of various sizes. An important result is that refined alignment models with a first-order dependence and a fertility model yield significantly better results than simple heuristic models. In the Appendix, we present an efficient training algorithm for the alignment models presented.

Download Full-text

Conceptually expanding the curricular alignment model to understand the coherence of the graded summative assessment practices of teachers: issues and perspectives

Mesure et évaluation en éducation ◽

10.7202/1084128ar ◽

2021 ◽

Vol 42 (spécial) ◽

pp. 39-68

Author(s):

Raphaël Pasquini

Keyword(s):

Secondary School Teachers ◽

Collaborative Study ◽

Summative Assessment ◽

Assessment Practices ◽

School Teachers ◽

Learning Tasks ◽

Assessment Approach ◽

Curricular Alignment ◽

Alignment Model ◽

Teaching Learning

Curricular alignment refers to the links of coherence, found in any teaching-learning process, between curriculum objectives, learning tasks and assessment approaches. This model makes it possible to understand the coherence of any assessment approach. By mobilizing data from a collaborative study carried out on eight secondary school teachers of mathematics and French, we will show, however, that its meaning is rather limited when it comes to understanding coherence in graded summative assessment practices and that, consequently, the model needs to be expanded conceptually. To this end, we will draw on an example of a summative test modelled in this way. Our findings demonstrate the relevance of analyzing summative assessment practices with the help of the expanded model, while considering the role that context plays in certain of its aspects.

Download Full-text

DESIGNING LIGHTWEIGHT OPEN INNOVATION: A CONCEPTUALISATION OF HOW LARGE FIRMS ENGAGE WITH SMALL ENTREPRENEURIAL FIRMS

International Journal of Innovation Management ◽

10.1142/s1363919621501036 ◽

2021 ◽

Author(s):

BJÖRN REMNELAND WIKHAMN

Keyword(s):

Open Innovation ◽

Entrepreneurial Firms ◽

Positive Outcomes ◽

Value Capture ◽

Design Elements ◽

A Value ◽

Value Propositions ◽

Equity Ownership ◽

Alignment Model ◽

Important Design

There is increasing scholarly interest in how large corporations engage in open innovation with small entrepreneurial firms, with synergies potentially producing positive outcomes for both the involved parties and the surrounding ecosystem. “Lightweight models” of open innovation (LOIs) have recently been introduced, governed by trust and relationships rather than by equity ownership and transactional control. This paper introduces a design framework and an alignment model for LOIs, based on 19 inductively generated and highly interrelated design elements associated with five design themes. The study uses empirical data from 18 LOI initiatives in Sweden, and the framework explains important differences in their motives, value propositions, innovation localizations, involved participants, and forms of interactions. Applying a value perspective to open innovation highlights two different value logics, suggesting that LOI initiatives can approach value by emphasizing either value creation or value capture. These logics may greatly influence other important design elements of LOIs.

Download Full-text