Closed smiyawaki0820 closed 3 years ago
In Semantic Dependency Parsing (SDP), semantic relations form directed acyclic graphs, rather than trees. We propose a new iterative predicate selection (IPS) algorithm for SDP. Our IPS algorithm combines the graph-based and transition-based parsing approaches in order to handle multiple semantic head words. We train the IPS model using a combination of multi-task learning and task-specific policy gradient training. Trained this way, IPS achieves a new state of the art on the SemEval 2015 Task 18 datasets. Furthermore, we observe that policy gradient training learns an easy-first strategy.
task : SDP .... 意味的関係が直接的な acyclic graph を形成
@inproceedings{kurita-sogaard-2019-multi,
author = {Kurita, Shuhei and S{\o}gaard, Anders},
title = {Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies},
booktitle = {Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics},
year = {2019},
url = {https://www.aclweb.org/anthology/P19-1232},
doi = {10.18653/v1/P19-1232},
pages = {2420--2430},
}
task: Semantic Dependency Parsing(SDP)
propose: 反復的述語選択(IPS)アルゴリズム
train IPS model
Result & Analysis
In Semantic Dependency Parsing (SDP), semantic relations form directed acyclic graphs, rather than trees. We propose a new iterative predicate selection (IPS) algorithm for SDP. Our IPS algorithm combines the graph-based and transition-based parsing approaches in order to handle multiple semantic head words. We train the IPS model using a combination of multi-task learning and task-specific policy gradient training. Trained this way, IPS achieves a new state of the art on the SemEval 2015 Task 18 datasets. Furthermore, we observe that policy gradient training learns an easy-first strategy.
@inproceedings{kurita-sogaard-2019-multi,
author = {Kurita, Shuhei and S{\o}gaard, Anders},
title = {Multi-Task Semantic Dependency Parsing with Policy Gradient for Learning Easy-First Strategies},
booktitle = {Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics},
year = {2019},
publisher = {Association for Computational Linguistics},
url = {https://www.aclweb.org/anthology/P19-1232},
doi = {10.18653/v1/P19-1232},
pages = {2420--2430},
}
SDP
parser | scoring | 備考 | 特徴 |
---|---|---|---|
transition-based | state 間の遷移に対して | 段々と dep-graph を構築 | error 伝搬(unsuitable: long) |
graph-based | 全 edge に対して | tree decoding algorithm を採用 |
train 中は gold transition path に従う
gold paths の state 情報しか得られず,error 伝搬は起こりやすい
全ての取りうる arc に対してスコアリングするよう学習
decoding algorithm は尤もらしい dep-path を選択する
error 伝搬がないが,reinforcement learning が適用できない
(提案)new SDP-algorithm based on the head-selection algorithm (Zhang et al., 2017)
w_i : i-th word
τ : time step
T_i : w_i における考えうる全ての transitions
ARC_{i,j} : a transition to create an arc from w_j (pred) to w_i (arg)
partial parse graph y^τ
T_i = { NULL, ARC_{i,ROOT}, ARC_{i,1}, ... , ARC_{i,n} }
自身に戻る ARC_{i,i} は存在しない
新たに考慮される ARC は the set of arcs A^τ に含まれない
T_i^τ(τ での w_i の transitions) = T_i(全ての transitions)- {ARC_{i,i} + A^τ}
how to create semantic dependency arcs
non-deterministic oracle problem : there are several paths, depending on the orders of creating the arcs
In IPS parsing: path ごとに難易度が異なる
※ sequence tagger では,その effectiveness が proven されている
In this paper ...
updates the partial SDP graph y^τ stored in a semantic dep-matirx G^τ
一単語上での各 transition では,fills in one cell in a row(制限)
G=0 で初期化.(i-1)-th word が w_j の arg であれば G[i,j] = 1
G を rank-3 tensor G' ∈ R^{n×(n+1)×p} に変換
(by) token u(w*) の emb (3p-dims)と 以下の g'ij とを入れ替えることによって
g_ij=1 の時はそのまま使う.=0 の時は u(w_NONE) に置き換える?
これによって作成した G' を single-BiLSTM で encode
最終的に,g_NULL の hidden state と concat することで partial SDP graph G^τ を取得
dep-flags: F'
transition score:
For supervised learning ... cross entropy loss
l_i : gold transition label for i-th word
θ : all trainable parameters
(this supervised)not have a principled answer to the non-deterministic oracle problem
and ランダムに transition paths を those consistent with the gold annotations to create transition labels よりサンプリング
also develop a semantic dep-labeling NN
[hi, hj]
: w_i とその述語 w_j における BiLSTM からの hidden-replabel l の score: from pred j to word i
a method for learning to iteratively act according to a dynamic environment in order to optimize future rewards
- the agent ~ NN model predicting the transition probabilities p_i(t_j^τ)
- the environment ~ include the partial SDP graph y^τ
- the rewards ~ computed by comparing the predicted parse graph to the gold parse graph y^g
objective function: the rewards を最大化
the transition policy for the w_i
transition の確率より given: π ~ p_i(t_j^τ|y^τ)
この勾配を計算する際,the expectation E_π を近似する
for any transition sequence w/ policy π からサンプリングされた a single transition path t
How the cross entropy loss and the policy gradient loss are similar ?
No. | 強化学習(policy gradient loss) | 教師あり学習(cross entropy loss) |
---|---|---|
sampling of transitions | allow model to explore transition paths | never follow |
decisions | dependent | independent(θ updated after finishing parsing) |
loss | can be negative | non-negative |
intermediate rewards(r_i^τ): given during parsing, at dfferent τ
(比較実験)
IPS + ML + RL
(推論時)
IPS + ML
: small improvementIPS + RL
: better than IPS
significantly (p < 10^-3)IPS + ML + RL
: the best IPS + ML + RL
: more likely learns an easy-first strategy(prone to connects neighboring words at the 1st transition)
1. どんなもの?
(タスク)
(提案)
(結果)
2. 先行研究と比べてどこがすごい?
IPS algorithm
graph-based
transition-based
(提案手法)graph-based + transition-based
3. 技術や手法のキモはどこ?
IPS algorithm
IPS model 🤔
Rewards of Policy Gradient
multi-task learning
SDP にはいくつかの linguistic formalisms が存在,overlap / synergy 箇所がある
4. どうやって有効だと検証した?
ablation によって multi-task および reinforcement learning の有効性を検証
main results
Arc length distributions for RL
5. 議論はある?
6. 次に読むべき論文は?