This is an old revision of the document!
Slide 1 :public:bg19.jpg
Joint Approaches for Sentence Alignment on Multiparallel Texts
Johannes Graën
2016-11-29
↓ Slide 2
→ Slide 3
Sentence Alignment on Multiparallel Texts
↓ Slide 4
Example
de | 1+3 | Es geht nicht um die Großzügigkeit des Präsidenten, es geht um die Zeit, die Sie sich selbst genehmigen; |
2+3 | ich habe Ihnen angezeigt, wann die Minute abgelaufen war. |
en | 3 | Mr Izquierdo Collado, it is not a question of the President’s generosity. |
It is a question of the time you allow yourself, because I informed you when your minute was up. |
fr | 1+3 | Monsieur Izquierdo, il ne s’agit pas de la générosité du président, il s’agit du temps que vous vous attribuez. |
2+3 | Je vous ai fait signe quand vous avez atteint la minute. |
→ Slide 5
Agreement of Bilingual Alignments
↓ Slide 6
Approach
Perform pairwise alignments with hunalign.
Join all these alignments in a graph.
Calculate “connectedness” by counting supporting languages for each edge.
Continue deleting the least supported edge until small consistent clusters emerge.
An alignment hierarchy can be obtained by reversing the deletion process.
↓ Slide 7
Two Languages
↓ Slide 8
Three Languages
↓ Slide 9
Four Languages
↓ Slide 10
Five Languages
↓ Slide 11
Hierarchical Alignments
↓ Slide 12
Problems/Limitations
Short sentences in 1:n pairs do not align in any language.
Decisions not straightforward if more than one pair disagrees.
Dictionaries for hunalign only allow binary entries.
→ Slide 13
Sampling in Discrete Vector Space
Discrete vector space spanned by languages as dimensions.
Alignments are pairs of location and direction vectors with all positive components.
E.g. $\left((1,2)^T,(2,1)^T\right)$ defines the alignment of the second and third sentence of the first language with the third sentence of the second language.
The location vector of the nth alignment equals the sum of the direction vector 1..n-1.
Alignment is obtained by simulated annealing.
↓ Slide 14
Visual Representation
↓ Slide 15
Approach
Set initial aligment to a sequence of vectors approximating a diagonal line.
Calculate local (pairwise) and global alignment scores (and keep results in memory).
Find and evaluate all applicable sampling operations.
Sample by selecting one of those operations – according to their respective evaluation scores.
Lower temperature, i.e. probability of picking an operation that leads to a worse sample.
Repeat from (3) until temperature reaches zero-point.
↓ Slide 16
Sampling Operations
Changing two consecutive vectors $\vec{x}$ and $\vec{y}$ such that $\vec{x}^\prime + \vec{y}^\prime = \vec{x} + \vec{y}$.
Replace two consecutive vectors $\vec{x}$ and $\vec{y}$ by vector $\vec{z}$ such that $\vec{z} = \vec{x} + \vec{y}$.
Split a vector $\vec{z}$ into vectors $\vec{x}$ and $\vec{y}$ such that $\vec{x} + \vec{y} = \vec{z}$.
↓ Slide 17
Problems/Limitations
The sampling often does not converge.
For higher dimensions (more languages), sampling did not inlcude the correct (gold) alignment.
Finding and evaluation all applicable operations is expensive, even if scores are only calculated on language pairs.
→ Slide 18
Agglomerative Hierarchical Clustering
→ Slide 19
Evaluation Metrics
→ Slide 20
Hierarchical Clustering for Word Alignment
Slide 21 :public:bg23.jpg