Ngoc Thang Vu

3 paper cards in Mova's library · 2018–2020

Ngoc Thang Vu wrote 3 of the papers Mova reads, in 2018–2020, most of them on Dependency parsing, Machine translation, Treebanks; often with Jonas Kuhn, Xiang Yu, Agnieszka Falenska.

The profile is the library's match of the name; two people with one name are not always told apart.

Ideas Mova took (1)

From the ideas ledger: an idea tried, kept or queued, or measured with no gain, and what it gave.

measured, no gain — en/src/parse.rs: add a dynamic oracle for arc-standard that charges each action by the number of gold arcs it makes unreachable, and train with it; measure cv LAS on the train folds against the static oracle, two runs each. (Approximate Dynamic Oracle for Dependency Parsing with… · approximate arc-standard dynamic oracle with exploration, greedy parser, EWT dev our tags: 77.67 -> 76.40 (variants 76.04-76.49; 20 epochs 77.50 -> 76.47)…)

Fast and Accurate Non-Projective Dependency Tree Linearization (2020)
The paper treats linearization as a travelling-salesman problem. A biaffine model gives the cost of putting one token before another, and the costs are solved per subtree with a greedy TSP solver. The subtree orders are then combined into a projective tree. A transition system, borrowed from non-projective parsing…
Head-First Linearization with Tree-Structured Representation (2019)
The paper linearises a dependency tree, that is, it finds the word order of an unordered tree. It has two parts. The encoder is a bidirectional Tree-LSTM that passes information first up the tree and then down, so that each token sees the whole tree. The decoder is head-first: it starts at the head and attaches…
Approximate Dynamic Oracle for Dependency Parsing with Reinforcement Learning (2018)
A dynamic oracle tells a transition parser which action is best from a state the parser itself got into by mistake, so that training covers its own errors.