Differences

This shows you the differences between two versions of the page.

--- nlp:semantic_parsing [2021/06/06 18:33] – [Tutorials] jmflanig
+++ nlp:semantic_parsing [2025/01/17 20:38] (current) – [Implicit Roles] jmflanig
@@ Line 9: / Line 9: @@
   * PropBank
     * [[https://arxiv.org/pdf/1804.08199.pdf|Strubell et al 2018 - Linguistically-Informed Self-Attention for Semantic Role Labeling]] (EMNLP 2018 Best Paper)
+==== Implicit Roles ====
+Related: [[abstract_meaning_representation#Multi-sentence AMR]]. Implicit roles (also called implicit arguments or multi-sentence arguments) are semantic roles that exist across sentences.
+  * [[https://aclanthology.org/I17-1010.pdf|Ngoc Thi Do et al 2017 - Improving Implicit Semantic Role Labeling by Predicting Semantic Frame Arguments]]
+  * [[https://arxiv.org/pdf/1911.03766.pdf|Ebner et al 2019 - Multi-Sentence Argument Linking]]
 ===== Supervised Parsing =====
@@ Line 21: / Line 27: @@
   * [[https://arxiv.org/pdf/1607.01432.pdf|Lee et al 2016 - Global Neural CCG Parsing with Optimality Guarantees]]
-===== Indirect Supervision =====
+==== QA Format ====
+  * QASRL paper
+  * QAMR paper
+  * [[https://aclanthology.org/2021.emnlp-main.108.pdf|Pyatkin et al 2021 - Asking It All: Generating Contextualized Questions for any Semantic Role]]]]
+===== Indirect Supervision =====
+==== Papers ====
+    * [[https://aclanthology.org/2022.acl-long.525.pdf|Ou & Liu 2022 - Learning to Generate Programs for Table Fact Verification via Structure-Aware Semantic Parsing]]
 ==== Learning from Feedback ====
   * [[https://arxiv.org/pdf/1606.02447.pdf|Wang et al 2016 - Learning Language Games through Interaction]] (Percy Liang)
   * [[https://arxiv.org/pdf/1704.08760.pdf|Iyar et al 2017 - Learning a Neural Semantic Parser from User Feedback]] Basic idea: parse to an executable representation, excecute it (try it on some examples), have humans look at the output and provide feedback (in few different ways)
   * [[https://arxiv.org/pdf/2010.05190.pdf|Karamcheti et al 2020 - Learning Adaptive Language Interfaces through Decomposition]] (Percy Liang)
+===== Prompting and Language Models =====
+  * [[https://arxiv.org/pdf/2209.15003.pdf|Drozdov 2022 - Compositional Semantic Parsing with Large Language Models]]
+===== Domain Adaptation =====
+  * [[https://arxiv.org/pdf/2110.08458.pdf|Pasupat et al 2021 - Controllable Semantic Parsing via Retrieval Augmentation]] Retrieves similar exemplars, from training data, and uses essentially GPT-style few-shot learning method to produce the parse with a seq2seq model
+===== Data Augmentation =====
+See also [[nlp:Data Augmentation]].
+  * [[https://arxiv.org/pdf/1606.03622.pdf|Jia & Liang 2016 - Data Recombination for Neural Semantic Parsing]]
 ===== Datasets =====
@@ Line 55: / Line 78: @@
   * [[Abstract Meaning Representation]]
   * [[Language to Programs]]
+  * [[Logic in NLP]]
   * [[Semantic Dependencies]]
   * [[Semantic Representations]]