External Knowledge Bases Research Articles

Soft Prompt-tuning with Self-Resource Verbalizer for short text streams

Short text streams such as real-time news and search snippets have attained vast amounts of attention and research in recent decades, the characteristics of high generation velocity, feature sparsity, and high ambiguity accentuate both the importance and challenges to language models. However, most of the existing short text stream classification methods can neither automatically select relevant knowledge components for arbitrary samples, nor expand knowledge internally instead of rely on external open knowledge base to address the inherent limitations of short text stream. In this paper, we propose a Soft Prompt-tuning with Self-Resource Verbalizer (SPSV for short) for short text stream classification, the soft prompt with self-resource knowledgeable expansion is conducted for updating label words space to address evolved semantic topics in the data streams. Specifically, the automatic constructed prompt is first generated to instruct the model prediction, which is optimized to address the problem of high velocity and topic drift in short text streams. Then, in each chunk, the projection between category names and label words space, i.e. verbalizer, is updated, which is constructed by internal knowledge expansion from the short text itself. Through comprehensive experiments on four well-known benchmark datasets, we validate the superb performance of our method compared to other short text stream classification and fine-tuning PLMs methods, which achieves up to more than 90% classification accuracy with the counts of data chunk increased.

UniKDD: A Unified Generative model for Knowledge-driven Dialogue

knowledge-driven dialogue (KDD) is to introduce an external knowledge base, generating an informative and fluent response. However, previous works employ different models to conduct the sub-tasks of KDD, ignoring the connection between sub-tasks and resulting in a difficulty of training and inference. To solve those issues above, we propose the UniKDD, a unified generative model for KDD, which models all sub-tasks into a generation task, enhancing the connection between tasks and facilitating the training and inference. Specifically, UniKDD simplifies the complex KDD tasks into three main sub-tasks, i.e., entity prediction, attribute prediction, and dialogue generation. These tasks are transformed into a text generation task and trained by an end-to-end way. In the inference phase, UniKDD first predicts a set of entities used for current turn dialogue according to the dialogue history. Then, for each predicted entity, UniKDD predicts the corresponding attributes by the dialogue history. Finally, UniKDD generates a high-quality and informative response using the dialogue history and predicted knowledge triplets. The experimental results show that our proposed UniKDD can perform KDD task well and outperform the baseline on the evaluation of knowledge selection and response generation. The code is available at https://github.com/qianandfei/UniKDD.git.

Knowledge-injected Prompt Learning for Chinese Biomedical Entity Normalization

The Biomedical Entity Normalization (BEN) task aims to align raw, unstructured medical entities to standard entities, thus promoting data coherence and facilitating better downstream medical applications. Recently, prompt learning methods have shown promising results in the natural language processing field. However, existing research falls short in tackling the more complex Chinese BEN task, especially in the few-shot scenario with limited medical data, and the vast potential of the external medical knowledge base has not yet been fully exploited. To address these challenges, this article proposes a novel Knowledge-injected Prompt Learning (PL-Knowledge) method. Specifically, the approach consists of five stages: candidate entity matching, knowledge extraction, knowledge encoding, knowledge injection, and prediction output. By effectively encoding the knowledge items contained in medical entities and incorporating them into tailor-made knowledge-injected templates, the additional knowledge enhances the model’s ability to capture latent relationships between medical entities, thus achieving a better match with the standard entities. Comprehensive experiments are conducted on a benchmark dataset in both few-shot and full-scale settings. This method outperforms existing baselines, with an average accuracy improvement of 12.96 percentage points in few-shot and 0.94 percentage points in full-data cases, showcasing its excellence in the BEN task.

CRUD-RAG: A Comprehensive Chinese Benchmark for Retrieval-Augmented Generation of Large Language Models

Retrieval-Augmented Generation (RAG) is a technique that enhances the capabilities of large language models (LLMs) by incorporating external knowledge sources. This method addresses common LLM limitations, including outdated information and the tendency to produce inaccurate “hallucinated” content. However, evaluating RAG systems is a challenge. Most benchmarks focus primarily on question answering applications, neglecting other potential scenarios where RAG could be beneficial. Accordingly, in the experiments, these benchmarks often assess only the LLM components of the RAG pipeline or the retriever in knowledge-intensive scenarios, overlooking the impact of external knowledge base construction and the retrieval component on the entire RAG pipeline in non-knowledge-intensive scenarios. To address these issues, this paper constructs a large-scale and more comprehensive benchmark, and evaluates all the components of RAG systems in various RAG application scenarios. Specifically, we refer to the CRUD actions that describe interactions between users and knowledge bases, and also categorize the range of RAG applications into four distinct types–Create, Read, Update, and Delete (CRUD). “Create” refers to scenarios requiring the generation of original, varied content. “Read” involves responding to intricate questions in knowledge-intensive situations. “Update” focuses on revising and rectifying inaccuracies or inconsistencies in pre-existing texts. “Delete” pertains to the task of summarizing extensive texts into more concise forms. For each of these CRUD categories, we have developed different datasets to evaluate the performance of RAG systems. We also analyze the effects of various components of the RAG system, such as the retriever, context length, knowledge base construction, and LLM. Finally, we provide useful insights for optimizing the RAG technology for different scenarios 1 .

QA-RAG: Exploring LLM Reliance on External Knowledge

Large language models (LLMs) can store factual knowledge within their parameters and have achieved superior results in question-answering tasks. However, challenges persist in providing provenance for their decisions and keeping their knowledge up to date. Some approaches aim to address these challenges by combining external knowledge with parametric memory. In contrast, our proposed QA-RAG solution relies solely on the data stored within an external knowledge base, specifically a dense vector index database. In this paper, we compare RAG configurations using two LLMs—Llama 2b and 13b—systematically examining their performance in three key RAG capabilities: noise robustness, knowledge gap detection, and external truth integration. The evaluation reveals that while our approach achieves an accuracy of 83.3%, showcasing its effectiveness across all baselines, the model still struggles significantly in terms of external truth integration. These findings suggest that considerable work is still required to fully leverage RAG in question-answering tasks.

Generative commonsense knowledge subgraph retrieval for open-domain dialogue response generation

Grounding on a commonsense knowledge subgraph can help the model generate more informative and diverse dialogue responses. Prior Traverse-based works explicitly retrieve a subgraph from the external knowledge base (eKB). Notably, the available knowledge is strictly restricted by the eKB. To break this restriction, Generative Retrieval methods externalize knowledge from the language model. However, they always generate boring knowledge due to their one-pass externalization procedure. This work proposes a novel TiLM Traverse in Language Model (TiLM), which uses three ‘Chain-of-Thought’ sub-tasks, i.e., Query Entity Production, Topic Entity Prediction, and Knowledge Subgraph Completion, to build a high-quality knowledge subgraph to ground the next Response Generation without explicitly accessing the eKB in inference. Experimental results on both Chinese and English datasets demonstrate TiLM’s outstanding performance even only with a small scale of parameters.

Knowledge-injected prompt learning for actionable information extraction from crisis-related tweets

In crisis situations, emergency response organizations can derive significant benefits from actionable information available on social media. Such information provides effective guidance and support for rescue operations. Unlike traditional sources of information, social media reflects the latest developments during a crisis, as well as actionable suggestions and the needs of affected individuals. However, existing methods that leverage pre-trained language models for this task require a large amount of annotated data for fine-tuning. This data acquisition process is costly and impractical in some data-scarce scenarios. To address this challenge, we propose a Knowledge-injected Actionable Information Extraction Model (KAIEM) that can extract actionable information from crisis-related tweets in few-shot and zero-shot settings. Based on prompt learning, KAIEM automatically constructs the label word space of verbalizers using word embeddings, eliminating the need for manual design or external knowledge bases. This mechanism enables our model to make accurate predictions without relying on labeled data for fine-tuning. Additionally, KAIEM employs knowledge probing to inject conceptual knowledge into prompt templates, thereby enhancing the semantics of crisis-related tweets. We conducted a series of experiments on a real-world crisis-related dataset. The experimental results demonstrate the effectiveness of KAIEM in few-shot and zero-shot scenarios.

Testing and Validation of a Custom Retrained Large Language Model for the Supportive Care of HN Patients with External Knowledge Base.

This study aimed to develop a retrained large language model (LLM) tailored to the needs of HN cancer patients treated with radiotherapy, with emphasis on symptom management and survivorship care. A comprehensive external database was curated for training ChatGPT-4, integrating expert-identified consensus guidelines on supportive care for HN patients and correspondences from physicians and nurses within our institution's electronic medical records for 90 HN patients. The performance of our model was evaluated using 20 patient post-treatment inquiries that were then assessed by three Board certified radiation oncologists (RadOncs). The rating of the model was assessed on a scale of 1 (strongly disagree) to 5 (strongly agree) based on accuracy, clarity of response, completeness s, and relevance. The average scores for the 20 tested questions were 4.25 for accuracy, 4.35 for clarity, 4.22 for completeness, and 4.32 for relevance, on a 5-point scale. Overall, 91.67% (220 out of 240) of assessments received scores of 3 or higher, and 83.33% (200 out of 240) received scores of 4 or higher. The custom-trained model demonstrates high accuracy in providing support to HN patients offering evidence-based information and guidance on their symptom management and survivorship care.

Candidate-Heuristic In-Context Learning: A new framework for enhancing medical visual question answering with LLMs

Medical Visual Question Answering (MedVQA) is designed to answer natural language questions related to medical images. Existing methods largely adopting the cross-modal pre-training and fine-tuning paradigm, face limitations in accuracy due to data scarcity and insufficient incorporation of extensive medical knowledge. Drawing inspiration from the Knowledge-Based Visual Question Answering (KB-VQA) domain, which leverages Large Language Models (LLMs) and external knowledge bases, we introduce the Candidate-Heuristic In-Context Learning (CH-ICL) framework, a novel approach that leverages LLMs augmented with external knowledge to directly enhance existing MedVQA models. Specifically, we collect a pathology terminology dictionary from a public digital pathology library as an external knowledge base and use it to train a knowledge scope discriminator, which helps identify the knowledge scope required to answer a question. Then, we employ existing MedVQA models to provide reliable answer candidates along with their confidence scores. Finally, the knowledge scope and candidates, combined with retrieved in-context exemplars, are aggregated into prompts for heuristically guiding LLMs in answer generation. Experimental results on the PathVQA, VQA-RAD, and SLAKE public benchmarks show state-of-the-art performance, with improvements of 1.91%, 1.88%, and 2.17% respectively over the baseline. Code and dataset are available at https://github.com/ecoxial2007/CH-ICL.

High‐degree penalty based global statistical network embedding for name disambiguation in anonymized graph

SummaryIn person‐centric applications, the prevalence of shared names significantly hampers document retrieval, web search, and database integration, highlighting the critical need for name disambiguation. Network embedding based unsupervised name disambiguation methods have received much attention due to their wide applicability and superior disambiguation performance. However, existing methods require complex feature engineering, such as extracting biographical and source content features or constructing supplementary features from external knowledge bases, which are unavailable in privacy‐preserving scenarios. Moreover, they may face challenges such as imbalanced node training or overlooking global statistical information during node embedding learning. In this article, we propose a method to tackle the name disambiguation problem based on high‐degree penalty and global statistical information using only relational data. First, we construct a weighted source similarity network based on multi‐hop collaborator relationships. Second, we employ a high‐degree penalty based global statistical network embedding model to learn low‐dimensional node embeddings and preserve the structural features of the network. Finally, we cluster the same sources using a joint clustering algorithm that does not require prior knowledge of the number of clusters. The experiment validates the effectiveness of the proposed method on two real‐life datasets.

Artificial Intelligence Text Processing Using Retrieval-Augmented Generation: Applications in Business and Education Fields

Abstract The article studies the current text processing tools based on Artificial Intelligence. A literature review is done emphasizing the dynamic evolution of AI-powered text analytics, having as its central tool ChatGPT and its capabilities. The focus is centered on the techniques and methods that are using embeddings in order to improve large language models (LLMs). In this paper is analyzed the current situation of the literature in terms of text processing using Retrieval-Augmented Generation and is highlighted the potential of this technology to enhance the interpretability and trust in applications critical, such as those related to education or business. AI has revolutionized natural language processing (NLP), which facilitated the machines to interpret and generate text efficiently and accurately. In addition, large language models with external knowledge bases have been developed. These are used to produce more accurate and contextually relevant text responses. This approach is called Retrieval-Augmented Generation (RAG is one of the most significant recent advancements in this field. Based on our study, two use cases are implemented to show the applicability of our study: one related to education and one related to business IT-related documents. The methodology describes the techniques used. This includes retrieval-augmented generation and embedding stored using vector databases. Our custom models are evaluated by comparing them to the general ones, without embeddings, showing superior performance. The article highlights remarkable progress in Retrieval-Augmented Generation (RAG), which is used for AI text processing with a focus on business and education fields. Further in this paper, many of the most significant highlights are presented, which include a scalable framework for AI applications, a new integration of Retrieval-Augmented Generation and embeddings, practical application demonstrations, bridging gaps in the analysis op AI text, significant development in AI performance and optimizing educational and business processes.

Temporal validity reassessment: commonsense reasoning about information obsoleteness

It is useful for machines to know whether text information remains valid or not for various applications including text comprehension, story understanding, temporal information retrieval, and user state tracking on microblogs as well as via chatbot conversations. This kind of inference is still difficult for current models, including also large language models, as it requires temporal commonsense knowledge and reasoning. We approach in this paper the task of Temporal Validity Reassessment, inspired by traditional natural language reasoning to determine the updates of the temporal validity of text content. The task requires judgment whether actions expressed in a sentence are still ongoing or rather completed, hence, whether the sentence still remains valid or has become obsolete, given the presence of context in the form of a supplementary content such as a follow-up sentence. We first construct our own dataset for this task and train several machine learning models. Then we propose an effective method for learning information from an external knowledge base that gives information regarding temporal commonsense knowledge. Using our prepared dataset, we introduce a machine learning model that incorporates the information from the knowledge base and demonstrate that incorporating external knowledge generally improves the results. We also experiment with different embedding types to represent temporal commonsense knowledge as well as with data augmentation methods to increase the size of our dataset.

Infusing internalized knowledge of language models into hybrid prompts for knowledgeable dialogue generation

Existing knowledge-grounded dialogue (KGD) systems access the knowledge from an external knowledge base, then generate the context-coherent response accordingly. However, the knowledge access capability is constrained to the scale of a knowledge base. On the one hand, a small-scale knowledge base makes a model hard to generalize on unseen topics, while the improper shift of topics may induce an unsmooth conversation flow. On the other hand, a large-scale knowledge base requires a strong retrieval component to accurately index the context-relevant knowledge from many plausible candidates, costing significant amounts of time and resources. To address this, we regard the language model as a virtual knowledge base and propose homogenizing internalized knowledge of different language models into hybrid prompts. The hybrid prompts are a set of continuous vectors learned to represent knowledge inherently encoded in different language models. Furthermore, we devise a two-stage knowledge-grounding manner, in which both the knowledge internalized in language models and the knowledge provided by evidence can be jointly optimized to generate a knowledgeable response. We compare our proposed method with two groups of methods, including methods with explicit knowledge retrieval and those with implicit knowledge access. Experimental results on three knowledge-grounded dialogue corpora demonstrate advantages over these competitive methods.

Incorporating Entity Type-Aware and Word–Word Relation-Aware Attention in Generative Named Entity Recognition

Named entity recognition (NER) is a critical subtask in natural language processing. It is particularly valuable to gain a deeper understanding of entity boundaries and entity types when addressing the NER problem. Most previous sequential labeling models are task-specific, while recent years have witnessed the rise of generative models due to the advantage of tackling NER tasks in the encoder–decoder framework. Despite achieving promising performance, our pilot studies demonstrate that existing generative models are ineffective at detecting entity boundaries and estimating entity types. In this paper, a multiple attention framework is proposed which introduces the attention of entity-type embedding and word–word relation into the named entity recognition task. To improve the accuracy of entity-type mapping, we adopt an external knowledge base to calculate the prior entity-type distributions and then incorporate the information input to the model via the encoder’s self-attention. To enhance the contextual information, we take the entity types as part of the input. Our method obtains the other attention from the hidden states of entity types and utilizes it in self- and cross-attention mechanisms in the decoder. We transform the entity boundary information in the sequence into word–word relations and extract the corresponding embedding into the cross-attention mechanism. Through word–word relation information, the method can learn and understand more entity boundary information, thereby improving its entity recognition accuracy. We performed experiments on extensive NER benchmarks, including four flat and two long entity benchmarks. Our approach significantly improves or performs similarly to the best generative NER models. The experimental results demonstrate that our method can substantially enhance the capabilities of generative NER models.

K-PathVQA: Knowledge-Aware Multimodal Representation for Pathology Visual Question Answering.

Pathology imaging is routinely used to detect the underlying effects and causes of diseases or injuries. Pathology visual question answering (PathVQA) aims to enable computers to answer questions about clinical visual findings from pathology images. Prior work on PathVQA has focused on directly analyzing the image content using conventional pretrained encoders without utilizing relevant external information when the image content is inadequate. In this paper, we present a knowledge-driven PathVQA (K-PathVQA), which uses a medical knowledge graph (KG) from a complementary external structured knowledge base to infer answers for the PathVQA task. K-PathVQA improves the question representation with external medical knowledge and then aggregates vision, language, and knowledge embeddings to learn a joint knowledge-image-question representation. Our experiments using a publicly available PathVQA dataset showed that our K-PathVQA outperformed the best baseline method with an increase of 4.15% in accuracy for the overall task, an increase of 4.40% in open-ended question type and an absolute increase of 1.03% in closed-ended question types. Ablation testing shows the impact of each of the contributions. Generalizability of the method is demonstrated with a separate medical VQA dataset.

Research on Fake News Detection Based on Dual Evidence Perception

With the development of online social networks, the prevalence of fake news on social media is increasing, leading to a heightened focus on detecting fake news. To combat this, researchers have developed various strategies, including text-based and fact-checking approaches. The text-based approach analyses linguistic features within fake news posts, while the fact-checking method verifies statements against an external knowledge base, disregarding the linguistic patterns between posts. Despite their differences, both approaches offer complementary benefits in the fight against fake news, highlighting the importance of integrating these methods. However, there has been relatively limited investigation into merging these two strategies into a unified framework. In this paper, we proposed the FakE News Detection Based on Dual Evidence Perception (DEP-FEND) method. This novel approach merges text-based and fact-checking techniques into a unified framework to enhance fake news detection. DEP-FEND consists of three sub-modules: evidence perception based on the historical news environment, evidence perception from an external knowledge base, and the fake news detector. Initially, two independent modules were developed to gather evidence from the historical news environment and an external knowledge base using the Gaussian kernel soft-counting method, generating two distinct evidence vectors. Subsequently, these two independent perception modules were combined using domain gates, and the posts, along with the combined evidence perception signals, were inputted into the fake news detector. The experimental results on the two Chinese datasets demonstrate that our proposed solution outperformed other fake news detection methods with an average accuracy improvement of 3.6% and a macro-f1 improvement of 2.1%.

From Retrieval to Generation: A Simple and Unified Generative Model for End-to-End Task-Oriented Dialogue

Retrieving appropriate records from the external knowledge base to generate informative responses is the core capability of end-to-end task-oriented dialogue systems (EToDs). Most of the existing methods additionally train the retrieval model or use the memory network to retrieve the knowledge base, which decouples the knowledge retrieval task from the response generation task, making it difficult to jointly optimize and failing to capture the internal relationship between the two tasks. In this paper, we propose a simple and unified generative model for task-oriented dialogue systems, which recasts the EToDs task as a single sequence generation task and uses maximum likelihood training to train the two tasks in a unified manner. To prevent the generation of non-existent records, we design the prefix trie to constrain the model generation, which ensures consistency between the generated records and the existing records in the knowledge base. Experimental results on three public benchmark datasets demonstrate that our method achieves robust performance on generating system responses and outperforms the baseline systems. To facilitate future research in this area, the code is available at https://github.com/dzy1011/Uni-ToD.

A relational extraction approach based on multiple embedding representations and multi-head self-attention

Currently, word segmentation errors and polysemy problems are common in the field of Chinese relationship extraction. Although character-based model input can avoid word segmentation errors, in order to obtain the word information of a sentence, it is often necessary to introduce a dictionary or an external knowledge base to expand the word information, which requires a lot of manpower and time. In response to the above existing problems, this article uses characters as input, uses multiple embedding models to jointly form a character vector sequence, and obtains features containing character information through BiLSTM and attention layers; considering that convolutional neural networks are good at extracting local features, obtain features containing word information through multi-kernel convolutional layers and multi-head self-attention layers, and finally use a gating mechanism to fuse the features. The model was tested on the public SanWen data set and our own cultural-travel data set, and obtained F1 values of 61.22% and 60.26% respectively. Experimental results show that our method can achieve better relationship extraction effects without using word segmentation tools and without building a dictionary or external knowledge base, and the effect is better than most commonly used models currently.

Neurophysiological and psychophysical references for trends in supervised VQA multimodal deep learning: An interdisciplinary meta-analysis

Leading trends in multimodal deep learning for visual-question answering include Multimodal joint-embedding model, multimodal attention-based model, and multimodal external knowledge-based model. Several mechanisms and strategies are used in these models, including representation fusion methods, co-attention mechanisms, and knowledge base retrieval mechanisms. While a variety of works have comprehensively reviewed these strategies, a key gap in research is that there is no interdisciplinary analysis that connects these mechanisms with discoveries on human. As discussions of Neuro-AI continues to thrive, it is important to consider synergies among human level investigations and ANNs, specifically for using AI to reproduce higher order cognitive functions such as multisensory integration. Thus, Present meta-analysis aimed at the reviewing and connecting neurophysiological and psychophysical references to trends in VQA multimodal deep learning, focusing on 1) Providing back-up explanations for why several strategies in VQA MMDL leads to performances that are closer to human level and 2) Using VQA MMDL as an example to demonstrate how interdisciplinary perspective may foster the development of human level AI. The result of the meta-analysis builds connections between several sub-fields: Joint embedding mechanisms and SC neurons, multimodal attention mechanism and the retro-cue effect, and external knowledge base and engram mechanisms.

Hierarchical Understanding in Robotic Manipulation: A Knowledge-Based Framework

In the quest for intelligent robots, it is essential to enable them to understand tasks beyond mere manipulation. Achieving this requires a robust parsing mode that can be used to understand human cognition and semantics. However, the existing methods for task and motion planning lack generalization and interpretability, while robotic knowledge bases primarily focus on static manipulation objects, neglecting the dynamic tasks and skills. To address these limitations, we present a knowledge-based framework for hierarchically understanding various factors and knowledge types in robotic manipulation. Using this framework as a foundation, we collect a knowledge graph dataset describing manipulation tasks from text datasets and an external knowledge base with the assistance of large language models and construct the knowledge base. The reasoning tasks of entity alignment and link prediction are accomplished using a graph embedding method. A robot in real-world environments can infer new task execution plans based on experience and knowledge, thereby achieving manipulation skill transfer.