Multi-hop Question Research Articles

Multi-Hop Question Answering over Knowledge Graphs

Multi-Hop Question Answering over Knowledge Graphs (MHQA-KG) plays a pivotal role in various applications, including but not limited to Question Answering, Recommendation Systems, and Semantic Search. Nevertheless, current models for MHQA have limitations in their ability to grasp all the information included in the question, resulting a reduction in accuracy when producing answers. In order to mitigate this limitation, this paper proposes a novel Multi-Hop Question Answering over Knowledge Graphs approach. It mainly utilizes question and path embedding to answer multi-hop questions, significantly improving accuracy. This approach effectively captures auxiliary information that may present in the question. The experimental findings provide evidence that the suggested methodology outperforms the current state-of-the-art models, achieving highly accurate outcomes with improvements.

Adversarial Entity Graph Convolutional Networks for multi-hop inference question answering

Multi-hop reasoning is critical for natural language understanding but poses challenges for current models, requiring models capable of aggregating and reasoning across multiple sources of information. We propose an Adversarial Entity Graph Convolutional Network (AEGCN) to improve multi-hop inference performance. Unlike previous GNNs-based models, AEGCN places a greater emphasis on the construction of rich entity graph which focuses on identifying the related entities from support document and connecting these entities with innovative edge relationships. The entity graph built by AEGCN preserves both the semantic information and structure of the original text. Further, adversarial training is adopted to generate challenging embeddings for the entity graph, increasing the model’s robustness against the interference. The experiments evaluated on WIKIHOP and MEDHOP dataset indicate that AEGCN achieves 6.8 % and 7.7 % accuracy improvement over baseline model respectively, confirming the model’s advanced capability in multi-hop reasoning tasks.

GS-CBR-KBQA: Graph-structured case-based reasoning for knowledge base question answering

Knowledge Base Question Answering (KBQA) task is an important research direction in natural language processing. Due to the flexibility and ambiguity of natural language, users’ questions often have more complex query types and richer semantic information. To address this issue, this paper proposes the GS-CBR-KBQA model, a Case-Based Reasoning model tailored for KBQA to improve the semantic parsing accuracy and interpretability of natural language questions. The model integrates Knowledge-oriented Programming Language (KoPL) reasoning graphs with query information, employing a Graph Auto-Encoder and the RoBERTa pretrained language model for a highly effective case retrieval. This integration leads to a more robust knowledge retrieval and application approach, particularly innovative in capturing the relationships within KoPL graphs. The model addresses explicitly complex questions such as multi-hop reasoning and questions involving intricate entity relationships. Finally, our extensive experiments show that the model performs excellently in accuracy and F1 metrics on benchmark datasets such as WebQSP and ComplexWebQuestions, particularly in complex question-answering. The code of our model is available at https://anonymous.4open.science/r/GS-CBR-KBQA.

Incorporating multi-perspective information into reinforcement learning to address multi-hop knowledge graph question answering

Knowledge graph question answering (KGQA) aims to answer natural language questions from structured knowledge graphs (KGs). Traditional KQGA methods are usually limited to single-hop queries and cannot handle complex questions involving multi-hop reasoning well. To overcome this issue, multi-hop KGQA based on reinforcement learning (RL) has been proposed. However, multi-hop KGQA based on RL still faces some challenges. Firstly, due to the insufficient availability of latent environmental information during the reasoning process, the agent finds it challenging to make coherent and correct decisions. Secondly, the agent only receives rewards from the environment upon reaching the answer entity during the exploration, leading to slow or even obstructed learning. To address these shortcomings, we construct multi-perspective information based on the state of the environment, and integrate multi-perspective information with RL framework, thereby creating the Multi-Perspective Information Fusion Reasoning Network (MPIFRN). MPIFRN achieves the goal via three steps. (1) We construct three different views of information, i.e., expectation embedding, instruction-guided embedding, and path-aware embedding. These environmental cues provide more reliable support for decision-making. (2) We still adopt the method of mapping entities and relations into the knowledge graph embedding space to answer multi-hop questions. At each step of reasoning, we use a scoring function to measure the plausibility of each “triple” ¡topic entity, question, candidate entity¿ in the embedding space. (3) Furthermore, we employ the asynchronous advantage actor-critic (A3C) algorithm to guide the agent in selecting the most promising entities and to expand the reasoning paths in parallel by updating policy and value network parameters, thereby facilitating multi-hop knowledge graph question answering. We conduct extensive experiments on KGQA benchmark datasets, providing substantial evidence to demonstrate the effectiveness of our approach.

Scoring Multi-hop Question Decomposition Using Masked Language Models

Question answering (QA) is a sub-field of Natural Language Processing (NLP) that focuses on developing systems capable of answering natural language queries. Within this domain, multi-hop question answering represents an advanced QA task that requires gathering and reasoning over multiple pieces of information from diverse sources or passages. To handle the complexity of multi-hop questions, question decomposition has been proven to be a valuable approach. This technique involves breaking down complex questions into simpler sub-questions, reducing the complexity of the problem. However, it’s worth noting that existing question decomposition methods often rely on training data, which may not always be readily available for low-resource languages or specialized domains. To address this issue, we propose a novel approach that utilizes pre-trained masked language models to score decomposition candidates in a zero-shot manner. The method involves generating decomposition candidates, scoring them using a pseudo-log likelihood estimation, and ranking them based on their scores. To evaluate the efficacy of the decomposition process, we conducted experiments on two datasets annotated on decomposition in two different languages, Arabic and English. Subsequently, we integrated our approach into a complete QA system and conducted a reading comprehension performance evaluation on the HotpotQA dataset. The obtained results emphasize that while the system exhibited a small drop in performance, it still maintained a significant advance compared to the baseline model. The proposed approach highlights the efficiency of the language model scoring technique in complex reasoning tasks such as multi-hop question decomposition.

Natural Language Reasoning, A Survey

This survey paper proposes a clearer view of natural language reasoning in the field of Natural Language Processing (NLP), both conceptually and practically. Conceptually, we provide a distinct definition for natural language reasoning in NLP, based on both philosophy and NLP scenarios, discuss what types of tasks require reasoning, and introduce a taxonomy of reasoning. Practically, we conduct a comprehensive literature review on natural language reasoning in NLP, mainly covering classical logical reasoning, natural language inference, multi-hop question answering, and commonsense reasoning. The paper also identifies and views backward reasoning, a powerful paradigm for multi-step reasoning, and introduces defeasible reasoning as one of the most important future directions in natural language reasoning research. We focus on single-modality unstructured natural language text, excluding neuro-symbolic research and mathematical reasoning.

Knowledge Graph Multi-Hop Question Answering Based on Dependent Syntactic Semantic Augmented Graph Networks

In the rapidly evolving domain of question answering systems, the ability to integrate machine comprehension with relational reasoning stands paramount. This paper introduces a novel architecture, the Dependent Syntactic Semantic Augmented Graph Network (DSSAGN), designed to address the intricate challenges of multi-hop question answering. By ingeniously leveraging the synergy between syntactic structures and semantic relationships within knowledge graphs, DSSAGN offers a breakthrough in interpretability, scalability, and accuracy. Unlike previous models that either fall short in handling complex relational paths or lack transparency in reasoning, our framework excels by embedding a sophisticated mechanism that meticulously models multi-hop relations and dynamically prioritizes the syntactic–semantic context.

Question Calibration and Multi-Hop Modeling for Temporal Question Answering

Many models that leverage knowledge graphs (KGs) have recently demonstrated remarkable success in question answering (QA) tasks. In the real world, many facts contained in KGs are time-constrained thus temporal KGQA has received increasing attention. Despite the fruitful efforts of previous models in temporal KGQA, they still have several limitations. (I) They adopt pre-trained language models (PLMs) to obtain question representations, while PLMs tend to focus on entity information and ignore entity transfer caused by temporal constraints, and finally fail to learn specific temporal representations of entities. (II) They neither emphasize the graph structure between entities nor explicitly model the multi-hop relationship in the graph, which will make it difficult to solve complex multi-hop question answering. To alleviate this problem, we propose a novel Question Calibration and Multi-Hop Modeling (QC-MHM) network. Specifically, We first calibrate the question representation by fusing the question and the time-constrained concepts in KG. Then, we construct the GNN layer to complete multi-hop message passing. Finally, the question representation is combined with the embedding output by the GNN to generate the final prediction. Empirical results verify that the proposed model achieves better performance than the state-of-the-art models in the benchmark dataset. Notably, the Hits@1 and Hits@10 results of QC-MHM on the CronQuestions dataset's complex questions are absolutely improved by 5.1% and 1.2% compared to the best-performing baseline. Moreover, QC-MHM can generate interpretable and trustworthy predictions.

Tree-of-Reasoning Question Decomposition for Complex Question Answering with Large Language Models

Large language models (LLMs) have recently demonstrated remarkable performance across various Natual Language Processing tasks. In the field of multi-hop reasoning, the Chain-of-thought (CoT) prompt method has emerged as a paradigm, using curated stepwise reasoning demonstrations to enhance LLM's ability to reason and produce coherent rational pathways. To ensure the accuracy, reliability, and traceability of the generated answers, many studies have incorporated information retrieval (IR) to provide LLMs with external knowledge. However, existing CoT with IR methods decomposes questions into sub-questions based on a single compositionality type, which limits their effectiveness for questions involving multiple compositionality types. Additionally, these methods suffer from inefficient retrieval, as complex questions often contain abundant information, leading to the retrieval of irrelevant information inconsistent with the query's intent. In this work, we propose a novel question decomposition framework called TRQA for multi-hop question answering, which addresses these limitations. Our framework introduces a reasoning tree (RT) to represent the structure of complex questions. It consists of four components: the Reasoning Tree Constructor (RTC), the Question Generator (QG), the Retrieval and LLM Interaction Module (RAIL), and the Answer Aggregation Module (AAM). Specifically, the RTC predicts diverse sub-question structures to construct the reasoning tree, allowing a more comprehensive representation of complex questions. The QG generates sub-questions for leaf-node in the reasoning tree, and we explore two methods for QG: prompt-based and T5-based approaches. The IR module retrieves documents aligned with sub-questions, while the LLM formulates answers based on the retrieved information. Finally, the AAM aggregates answers along the reason tree, producing a definitive response from bottom to top.

A Relation Embedding Assistance Networks for Multi-hop Question Answering

Multi-hop Knowledge Graph Question Answering aims at finding an entity to answer natural language questions from knowledge graphs. When humans perform multi-hop reasoning, people tend to focus on specific relations across different hops and confirm the next entity. Therefore, most algorithms choose the wrong specific relation, which makes the system deviate from the correct reasoning path. The specific relation at each hop plays an important role in multi-hop question answering. Existing work mainly relies on the question representation as relation information, which cannot accurately calculate the specific relation distribution. In this article, we propose an interpretable assistance framework that fully utilizes the relation embeddings to assist in calculating relation distributions at each hop. Moreover, we employ the fusion attention mechanism to ensure the integrity of relation information and hence to enrich the relation embeddings. The experimental results on three English datasets and one Chinese dataset demonstrate that our method significantly outperforms all baselines. The source code of REAN will be available at https://github.com/2399240664/REAN

GeneGPT: augmenting large language models with domain tools for improved access to biomedical information.

While large language models (LLMs) have been successfully applied to various tasks, they still face challenges with hallucinations. Augmenting LLMs with domain-specific tools such as database utilities can facilitate easier and more precise access to specialized knowledge. In this article, we present GeneGPT, a novel method for teaching LLMs to use the Web APIs of the National Center for Biotechnology Information (NCBI) for answering genomics questions. Specifically, we prompt Codex to solve the GeneTuring tests with NCBI Web APIs by in-context learning and an augmented decoding algorithm that can detect and execute API calls. Experimental results show that GeneGPT achieves state-of-the-art performance on eight tasks in the GeneTuring benchmark with an average score of 0.83, largely surpassing retrieval-augmented LLMs such as the new Bing (0.44), biomedical LLMs such as BioMedLM (0.08) and BioGPT (0.04), as well as GPT-3 (0.16) and ChatGPT (0.12). Our further analyses suggest that: First, API demonstrations have good cross-task generalizability and are more useful than documentations for in-context learning; second, GeneGPT can generalize to longer chains of API calls and answer multi-hop questions in GeneHop, a novel dataset introduced in this work; finally, different types of errors are enriched in different tasks, providing valuable insights for future improvements. The GeneGPT code and data are publicly available at https://github.com/ncbi/GeneGPT.

DSAMR: Dual-Stream Attention Multi-hop Reasoning for knowledge-based visual question answering

Knowledge-based visual question answering aims to associate external knowledge facts for answering questions about images. Most existing methods emphasize high-order associations between knowledge facts and questions, and fail to consider the negative effects of unnecessary knowledge facts in multi-hop reasoning. In this paper, we propose a Dual-Stream Attention Multi-hop Reasoning (DSAMR) architecture that constructs two different attention streams to mitigate unnecessary knowledge facts. This dual-stream mechanism enables the model to reduce the attention weights on unnecessary knowledge while gathering essential knowledge by learning the implicit correlations between knowledge facts and questions. In addition, we designed a hypergraph knowledge extraction module in the architecture to extract optimal knowledge facts by evaluating the relevance of each knowledge fact to the question. The experimental results demonstrate the effectiveness of our method not only on the knowledge-based visual question answering dataset KVQA, but also on the multi-hop question answering dataset PathQuestion.

Leveraging Knowledge Graph Reasoning in a Multihop Question Answering System for Hot Rolling Line Fault Diagnosis

Leveraging Knowledge Graph Reasoning in a Multihop Question Answering System for Hot Rolling Line Fault Diagnosis

Global and item-by-item reasoning fusion-based multi-hop KGQA

Existing embedded multi-hop Question Answering over Knowledge Graph (KGQA) methods attempted to handle Knowledge Graph (KG) sparsity using Knowledge Graph Embedding (KGE) to improve KGQA. However, they almost ignore the intermediate path reasoning process of answer prediction, do not consider the information interaction between the question and the KG, and rarely consider the problem that the triple scoring reasoning mechanism is inadequate in extracting deep features. To address the above issues, this paper proposes Global and Item-by-item Reasoning Fusion-based Multi-hop KGQA (GIRFM-KGQA). In global reasoning, a convolutional attention reasoning mechanism is proposed and fused with the triple scoring reasoning mechanism to jointly implement global reasoning, thus enhancing the long-chain reasoning ability of the global reasoning model. In item-by-item reasoning, the reasoning path is formed by serially predicting relations, and then the answer is predicted, which effectively solves the problem that the embedded multi-hop KGQA method lacks the intermediate path reasoning ability. In addition, we introduce an information interaction method between the question and the KG to improve the accuracy of the answer prediction. Finally, we merge the global reasoning score with the item-by-item reasoning score to jointly predict the answer. Our model, compared to the baseline model (EmbedKGQA), achieves an accuracy improvement of 0.5% and 2.7% on two-hop questions, and 6.2% and 4.6% on three-hop questions for the MetaQA_Full and MetaQA_Half datasets, and 1.7% on the WebQuestionSP dataset, respectively. The experimental results show that the proposed model can effectively improve the accuracy of the multi-hop KGQA model and enhance the interpretability of the model. We have made our model’s source code available at github: https://github.com/feixiongfeixiong/GIRFM.

Multi-hop community question answering based on multi-aspect heterogeneous graph

Community question answering aims to connect queries and answers based on users' community behaviors, find the most relevant solutions for newly raised questions, and improve community activity. Existing research mainly focuses on the single-hop method, which relies on a particular feature for simple question answering. However, multi-hop question answering, which depends on multi-step reasoning to solve complex questions, still faces the destitute adaptability caused by various entity features and complex entity correlations. In this work, we put forward a Multi-Hop Community Question-Answering method called MHCQA, that combines the multi-aspect features of community entities, constrains the answer retrieval process from different perspectives and improves the adaptability of multi-hop answer retrieval. Firstly, we utilize phrases for representing the semantic features of question-answering and fuse them with entity constitution properties through heterogeneous graphs to effectively indicate the diverse relationships among entities in communities. At the same time, we bring a dynamic heterogeneous network construction strategy to improve the efficiency of feature representation. Secondly, we enhance heterogeneous graph attention networks for efficiently computing local entity similarities to find relevant answers to questions based on partial graphs. Finally, we propose a multi-constraint multi-hop community question-answering method, which optimizes answer retrieval from three aspects: the hop count, the number of answers, and the relevance of responses, to improve the scene adaptability of multi-hop answer retrieving. Abundant experiments on realistic corpus (Quora, Math, Super User, and Stack Overflow) illustrate that MHCQA improves the multi-hop question answering error rate by about 2 % and is about 11 % ahead of the baselines in terms of operating efficiency. Given our present literature survey, MHCQA is a promising research idea for automatic answering services in online question-answering communities.

Multi-hop question answering of bridge inspection by adopting knowledge graph embedding technology

The complexity of relationships between bridge inspection presents a significant challenge for the answer system. Sparse knowledge graph (KG) due to limited triples further compounds the issue. To overcome this, this paper proposes a dynamic reasoning strategy for bridge inspection question answering. The framework comprises a teacher-student network and dynamic reasoning strategy. The teacher network, based on the neural state machine (NSM), acquires auxiliary intermediate supervision signals. Its output provides probability distribution and entity embedding as input for the student network. The student network, also NSM-based, provides accurate answers with the aid of intermediate supervision signals. To handle incomplete KG, the dynamic reasoning strategy incorporates knowledge embedding, updating the KG by capturing contextual information of each related entity node in relation to the question. Experiments on the bridge inspection dataset demonstrated the effectiveness of this method, outperforming other approaches.

ViWiQA: Efficient end-to-end Vietnamese Wikipedia-based Open-domain Question-Answering systems for single-hop and multi-hop questions

Open-domain Question-Answering (QA) task requires a QA system to answer a given question using a large knowledge base like Wikipedia. Modern Open-domain QA systems often follow the two-stage framework Retriever-Reader where the retriever greatly impacts the end-to-end performance. Efficient Vietnamese Open-domain QA systems for single and multi-hop questions have yet to be studied. Although resource-rich languages like English witnessed many advancements in Open-domain QA, these methods often suffer from low data situations. This study proposes ViWiQA, an efficient Vietnamese Open-domain QA system over the Wikipedia knowledge base, with two novel retriever methods for single-hop and multi-hop questions. ViWiQA can be effectively trained with low data and significantly outperforms Lucene-BM25 and Dense Passage Retrieval when adapted to Vietnamese datasets. For single-hop QA, the proposed retriever outperforms Lucene-BM25 by 20% in top-1 retrieval accuracy, and the end-to-end system achieves 15% and 17% absolute gain in EM and F1 scores, respectively. For multi-hop QA, the proposed retriever increases the accuracy of retrieving correct passage pairs by 4% compared to Lucene-BM25, and the end-to-end system shows 7% and 17% absolute gains in EM and F1 scores.

Multi-hop question answering using sparse graphs

Multi-hop question answering (QA) across multiple documents requires a deep understanding of relationships between entities in documents, questions, and answer candidates. Graph Neural Networks (GNNs) have emerged as a promising tool for multi-hop QA tasks. These models often suffer from increasing computational and model complexity, which makes them inefficient for real-world applications with limited resources. In this paper, we propose a graph-based approach called Sparse Graph-based Multi-hop Question Answering system (SG-MQA), which provides a throughout examination of the mentioned challenges and presents appropriate measures to address them. We propose a novel approach based on the Relational Graph Convolutional Network (R-GCN) that reduces the model complexity and improves its performance. We have utilized various strategies and conducted multiple experiments to achieve this goal. We show the efficacy of the proposed approach by examining the results of experiments on two QA datasets, namely WikiHop and HotpotQA. The SG-MQA model outperforms all the state-of-the-art (SOTA) methods on WikiHop and increases the accuracy of the best previous approach from 74.4% to 78.3%. Additionally, it achieves acceptable performance on HotpotQA. Although, according to the F1 measure, the performance of SG-MQA is inferior to that of the SOTA model, it is comparable to that of all other approaches. On the other hand, based on the Exact Match (EM) measure, SG-MQA shows comparable performance to that of the SOTA model and outperforms all other approaches.

Prompt-WNQA: A prompt-based complex question answering for wireless network over knowledge graph

With the rapid development of wireless networks, the scale of network devices is constantly expanding. Experts spend vast amounts of time consulting and dealing with various network problems of the same nature during the course of daily network operation and maintenance service, and there is no time to collect statistics and sort their information, which affects the overall operation and maintenance efficiency. This paper investigates complex Knowledge Graph Question Answering (KGQA) in the wireless network domain in order to improve operation and maintenance efficiency. Accordingly, we propose Wireless Network QA over the KG using Prompt learning (Prompt-WNQA), a novel method that deals with both constraint-based and multi-hop complex questions making use of prompt learning. Our method is also helpful in performing complex KGQA over incomplete KGs, which can complete missing relations between unconnected entities. Compared to state-of-the-art approaches, our Prompt-WNQA achieves significant improvement over extensive experiments on both the wireless network QA dataset and two public complex QA datasets.

Query Path Generation via Bidirectional Reasoning for Multihop Question Answering From Knowledge Bases

Multi-hop question answering from knowledge bases (KBQA) is a hot research topic in natural language processing. Recently, graph neural network-based (GNN-based) methods have achieved promising results as the KB can be organized as a knowledge graph (KG). However, they often suffered from the sparsity of the KG which was detrimental to the structure-encoding and reasoning capabilities of GNN. Specifically, a KG is a sparse graph linked by directed relations and previous studies have paid scant attention to the directional characteristic of relations in the KG, limiting the patterns of relation path that GNN-based approaches could resolve. This study proposes a bidirectional recurrent graph neural network (BRGNN) to tackle these difficulties. To model the bidirectional information of relations, all adjacent relations of an entity are grouped by their directions, and they are separately aggregated into the entity representation in outward and inward directions. For the reasoning process, BRGNN simultaneously considers the neighbor relations in both directions to cover more patterns of relation paths and improve the recall of answers. Extensive experiments on three benchmarks: WebQuestionsSP, ComplexWebQuestions and MetaQA, verify that BRGNN can answer more questions by taking into account the directional information, and it is competitive to all state-of-the-art approaches.