| Challenge: | Existing models for constituency parsing are transition based, decomposing production of each parse tree into a sequence of action decisions. |
| Approach: | They propose to use a policy gradient method to reduce exposure bias by allowing exploration during training. |
| Outcome: | The proposed method outperforms static oracle likelihood training on four constituency parsers in three languages on four parser-agnostic languages. |
Similar Papers
A Dynamic Oracle for Linear-Time 2-Planar Dependency Parsing (N18-2)
Copied to clipboard
| Challenge: | Existing dynamic oracles for greedy parsers can handle non-projective syntax, but none are available for these types of training. |
| Approach: | They propose an efficient dynamic oracle for training the 2-Planar transition-based parser with over 99% coverage on non-projective syntactic corpora. |
| Outcome: | The proposed model outperforms the static training strategy in the vast majority of languages tested and scored better on most datasets than the arc-hybrid parser enhanced with the Swap transition. |
Exploiting Dynamic Oracles to Train Projective Dependency Parsers on Non-Projective Trees (N18-2)
Copied to clipboard
| Challenge: | Several strategies have been proposed to overcome the projectivity constraint by introducing transition-based dependency parsers that can build non-projective dependencies. |
| Approach: | They propose a modification of dynamic oracles to allow use of non-projective data . their method consistently outperforms traditional projectivization and pseudo-projectivisation approaches . |
| Outcome: | The proposed method outperforms projectivization and pseudo-projectivisation methods on 73 treebanks and achieves significant gains for non-projective languages. |
Linear-time Constituency Parsing with RNNs and Dynamic Programming (P18-2)
Copied to clipboard
| Challenge: | Existing span-based constituency parsers are too slow for longer sentences and for applications beyond sentence boundaries. |
| Approach: | They propose a linear-time constituency parser with RNNs and dynamic programming using graph-structured stack and beam search. |
| Outcome: | The proposed parser is faster for long sentences and faster for discourse parsing. |
Dynamic Oracles for Top-Down and In-Order Shift-Reduce Constituent Parsing (D18-1)
Copied to clipboard
| Challenge: | Top-down and in-order shift-reduce constituent parsers are the most accurate known shift-reducing algorithms for constituent paring. |
| Approach: | They propose to use dynamic oracles to train two of the most accurate shift-reduce algorithms for constituent parsing. |
| Outcome: | The proposed top-down and in-order shift-reduce parsers improve on the WSJ benchmark. |
Dynamic Head Selection for Neural Lexicalized Constituency Parsing (2025.acl-long)
Copied to clipboard
| Challenge: | Lexicalized parsing has traditionally been neglected in favor of unlexicalized, span-based methods. |
| Approach: | They propose a latent lexicalization framework that dynamically infers lexicals from data without relying on predefined head-finding rules. |
| Outcome: | The proposed model learns lexical dependencies directly from data, offering greater adaptability across languages and datasets. |
Contrastive Policy Gradient: Aligning LLMs on sequence-level scores in a supervised-friendly fashion (2024.emnlp-main)
Copied to clipboard
Yannis Flet-Berliac, Nathan Grinsztajn, Florian Strub, Eugene Choi, Bill Wu, Chris Cremer, Arash Ahmadian, Yash Chandak, Mohammad Azar, Olivier Pietquin, Matthieu Geist
| Challenge: | Reinforcement Learning (RL) is a method used to fine tune Large Language Models (LLMs) using a reward model trained from preference data to better align with human judgment. |
| Approach: | They propose a Reinforcement Learning (RL) algorithm that can estimate the optimal policy even from off-policy data. |
| Outcome: | The proposed algorithm can estimate the optimal policy even from off-policy data. |
Dynamic Meta-Metrics: Source-Sentence Conditioned Weighting for MT Evaluation (2026.acl-srw)
Copied to clipboard
| Challenge: | Rather than relying on a single static ensemble or language-specific weighting, DMM adapts the metric combination based on properties of the source segment. |
| Approach: | They propose a framework for machine translation evaluation that learns source-sentence conditioned combinations of existing metrics. |
| Outcome: | The proposed framework outperforms linear and Gaussian process-based ensembles across multiple language pairs and introducing soft conditioning yields gains over linear models. |
Automatically Selecting the Best Dependency Annotation Design with Dynamic Oracles (N18-2)
Copied to clipboard
| Challenge: | Multiple annotation conventions have been proposed for representing dependency structures. |
| Approach: | They propose to consider a set of syntactic references encoding alternative syntak representations to train a parser with a dynamic oracle. |
| Outcome: | The proposed approach can predict the best syntactic representation among all possible references. |
Bridging SFT and RL: Dynamic Policy Optimization for Robust Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing unified optimization strategies overlook the statistical conflict between these distinct gradient signals. |
| Approach: | They propose a framework to reduce bias-variance trade-offs in Large Language Models . they propose DYPO, which leverages intrinsic group dynamics to significantly reduce RL gradient variance . |
| Outcome: | The proposed framework outperforms traditional pipelines on reasoning benchmarks and out-of-distribution tasks. |
On the Role of Supervision in Unsupervised Constituency Parsing (2020.emnlp-main)
Copied to clipboard
| Challenge: | Recent work on unsupervised constituency parsing uses labeled examples for tuning . a few-shot parser with labeles can outperform other approaches by a significant margin . |
| Approach: | They propose to use as few labeled examples as possible for model development . they propose to train existing models on the same labeles they access . |
| Outcome: | The proposed model outperforms other models on the WSJ development set by a significant margin . the proposed model can be further improved by augmentation and self-training . |