Towards Generalizable Neuro-Symbolic Systems for Commonsense Question Answering

By Kaixin Ma, Jonathan M Francis, Quanyang Lu, Eric Nyberg, Alessandro Oltramari. 2019

94 citations recorded by the source corpus.

Non-extractive commonsense QA remains a challenging AI task, as it requires systems to reason about, synthesize, and gather disparate pieces of information, in order to generate responses to queries. Recent approaches on such tasks show increased performance, only when models are either pre-trained with additional information or when domain-specific heuristics are used, without any special consideration regarding the knowledge resource type. In this paper, we perform a survey of recent commonsense QA methods and we provide a systematic analysis of popular knowledge resources and knowledge-integration methods, across benchmarks from multiple commonsense datasets. Our results and analysis show that attention-based injection seems to be a preferable choice for knowledge integration and that the degree of domain overlap, between knowledge bases and datasets, plays a crucial role in determining model success.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

GraDA: Graph Generative Data Augmentation for Commonsense Reasoning (2022, methodology): “Our approach is orthogonal to HyKAS, KG-Fusion as their instance-level approach retrieves information for each sample while G RA DA augments knowledge on a global level.”
Explanations for CommonsenseQA: New Dataset and Models (2021, background): “The goal behind original CQA task is confined only till answering the questions and hence almost all the submissions (Ma et al., 2019; Khashabi et al., 2020; Zhu et al., 2020; Yang et al., 2020) to the leaderboard of the CQA dataset focus just on answering the question and not generating…”
Think about it! Improving defeasible reasoning by first modeling the question scenario. (2021, background): “…et al., 2019; Guu et al., 2020), or with knowledge generated from a separate source (Shwartz et al., 2020; Bosselut et al., 2019); and injected either as extra sentences fed directly to the model (Clark et al., 2020), via the loss function (Tandon et al., 2018), or via attention (Ma et al., 2019).”
GraDA: Graph Generative Data Augmentation for Commonsense Reasoning (2022, background): “Our approach is orthogonal to HyKAS, KG-Fusion as their instance-level approach retrieves information for each sample while GRADA augments knowledge on a global level.”
Knowledge-Augmented Methods for Natural Language Processing (2023, methodology): “For the multiple-choice QA setting, we introduce a set of knowledge-augmented neuro-symbolic meth-ods: KagNet* (Lin et al., 2019), HyKAS (Ma et al., 2019), MHGRN* (Feng et al., 2020), Hy-bridGN (Yan et al., 2020) and QA-GNN* (Ya-sunaga et al., 2021).”
Generated Knowledge Prompting for Commonsense Reasoning (2021, background): “However, if there is a domain mismatch between the dataset and the knowledge base, such gains tend to diminish (Mitra et al., 2019; Ma et al., 2019).”
Scalable Multi-Hop Relational Reasoning for Knowledge-Aware Question Answering (2020, result): “We do compare our models with those (Ma et al., 2019; Lv et al., 2019; Khashabi et al., 2020) augmented by other text-form external knowledge (e.”