Educators evaluate student knowledge using knowledge component (KC) models that map assessment questions to KCs. Still, designing KC models for large question banks remains an insurmountable challenge for instructors who need to analyze each question by hand. The growing use of Generative AI in education is expected only to aggravate this chronic deficiency of expert-designed KC models, as course engineers designing KCs struggle to keep up with the pace at which questions are generated. In this work, we propose KCluster, a novel KC discovery algorithm based on identifying clusters of congruent questions according to a new similarity metric induced by a large language model (LLM). We demonstrate in three datasets that an LLM can create an effective metric of question similarity, which a clustering algorithm can use to create KC models from questions with minimal human effort. Combining the strengths of LLM and clustering, KCluster generates descriptive KC labels and discovers KC models that predict student performance better than the best expert-designed models available. In anticipation of future work, we illustrate how KCluster can reveal insights into difficult KCs and suggest improvements to instruction.
Existing work on large language model (LLM) personalization assigned different responding roles to LLM, but overlooked the diversity of questioners. In this work, we propose a new form of questioner-aware LLM personalization, generating different responses even for the same query from different questioners. We design a dual-tower model architecture with a cross-questioner general encoder and a questioner-specific encoder. We further apply contrastive learning with multi-view augmentation, pulling close the dialogue representations of the same questioner, while pulling apart those of different questioners. To mitigate the impact of question diversity on questioner-contrastive learning, we cluster the dialogues based on question similarity and restrict the scope of contrastive learning within each cluster. We also build a multi-questioner dataset from English and Chinese scripts and WeChat records, called MQDialog, containing 173 questioners and 12 responders. Extensive evaluation with different metrics shows a significant improvement in the quality of personalized response generation.
The goal of this work was to compute the semantic similarity among publicly available health survey questions in order to facilitate the standardization of survey-based Person-Generated Health Data (PGHD). We compiled various health survey questions authored in both English and Korean from the NIH CDE Repository, PROMIS, Korean public health agencies, and academic publications. Questions were drawn from various health lifelog domains. A randomized question pairing scheme was used to generate a Semantic Text Similarity (STS) dataset consisting of 1758 question pairs. Similarity scores between each question pair were assigned by two human experts. The tagged dataset was then used to build three classifiers featuring: Bag-of-Words, SBERT with BERT-based embeddings, and SBRET with LaBSE embeddings. The algorithms were evaluated using traditional contingency statistics. Among the three algorithms, SBERT-LaBSE demonstrated the highest performance in assessing question similarity across both languages, achieving an Area Under the Receiver Operating Characteristic (ROC) and Precision-Recall Curves of over 0.99. Additionally, it proved effective in identifying cross-lingual semantic similarities.The SBERT-LaBSE algorithm excelled at aligning semantically equivalent sentences across both languages but encountered challenges in capturing subtle nuances and maintaining computational efficiency. Future research should focus on testing with larger multilingual datasets and on calibrating and normalizing scores across the health lifelog domains to improve consistency. This study introduces the SBERT-LaBSE algorithm for calculating semantic similarity across two languages, showing it outperforms BERT-based models and the Bag of Words approach, highlighting its potential to improve semantic interoperability of survey-based PGHD across language barriers.
CQA services are valuable sources of knowledge that can be used to find answers to users' information needs. In these services, question retrieval aims to help users with their information needs by finding similar questions to theirs. However, finding similar questions is obstructed by the lexical gap that exists between relevant questions. In this work, we target this problem by using query expansion methods. We use word-similarity-based methods, propose a question-similarity-based method and selective expansion of these methods to expand a question that's been submitted and mitigate the lexical gap problem. Our best method achieves a significant relative improvement of 1.8\% compared to the best-performing baseline without query expansion.
In-context learning (ICL) is a powerful paradigm where large language models (LLMs) benefit from task demonstrations added to the prompt. Yet, selecting optimal demonstrations is not trivial, especially for complex or multi-modal tasks where input and output distributions differ. We hypothesize that forming task-specific representations of the input is key. In this paper, we propose a method to align representations of natural language questions and those of SQL queries in a shared embedding space. Our technique, dubbed MARLO - Metadata-Agnostic Representation Learning for Text-tO-SQL - uses query structure to model querying intent without over-indexing on underlying database metadata (i.e. tables, columns, or domain-specific entities of a database referenced in the question or query). This allows MARLO to select examples that are structurally and semantically relevant for the task rather than examples that are spuriously related to a certain domain or question phrasing. When used to retrieve examples based on question similarity, MARLO shows superior performance compared to generic embedding models (on average +2.9\%pt. in execution accuracy) on the Spider benchmark. It also outperforms the next best method that masks metadata information by +0.8\%pt. in execution accuracy on average, while imposing a significantly lower inference latency.
Community Question-Answering (CQA) forums have revolutionized how people seek information, especially those related to their healthcare needs, placing their trust in the collective wisdom of the public. However, there can be several answers in response to a single query, which makes it hard to grasp the key information related to the specific health concern. Typically, CQA forums feature a single top-voted answer as a representative summary for each query. However, a single answer overlooks the alternative solutions and other information frequently offered in other responses. Our research focuses on aspect-based summarization of health answers to address this limitation. Summarization of responses under different aspects such as suggestions, information, personal experiences, and questions can enhance the usability of the platforms. We formalize a multi-stage annotation guideline and contribute a unique dataset comprising aspect-based human-written health answer summaries. We build an automated multi-faceted answer summarization pipeline with this dataset based on task-specific fine-tuning of several state-of-the-art models. The pipeline leverages question similarity to retrieve relevant answer sentences, subsequently classifying them into the appropriate aspect type. Following this, we employ several recent abstractive summarization models to generate aspect-based summaries. Finally, we present a comprehensive human analysis and find that our summaries rank high in capturing relevant content and a wide range of solutions.
In the field of intelligent education, knowledge tracing (KT) has attracted increasing attention, which estimates and traces students' mastery of knowledge concepts to provide high-quality education. In KT, there are natural graph structures among questions and knowledge concepts so some studies explored the application of graph neural networks (GNNs) to improve the performance of the KT models which have not used graph structure. However, most of them ignored both the questions' difficulties and students' attempts at questions. Actually, questions with the same knowledge concepts have different difficulties, and students' different attempts also represent different knowledge mastery. In this paper, we propose a difficulty and attempts boosted graph-based KT (DAGKT), using rich information from students' records. Moreover, a novel method is designed to establish the question similarity relationship inspired by the F1 score. Extensive experiments on three real-world datasets demonstrate the effectiveness of the proposed DAGKT.
To investigate the role of linguistic knowledge in data augmentation (DA) for Natural Language Processing (NLP), we designed two adapted DA programs and applied them to LCQMC (a Large-scale Chinese Question Matching Corpus) for a binary Chinese question matching classification task. The two DA programs produce augmented texts by five simple text editing operations (or DA techniques), largely irrespective of language generation rules, but one is enhanced with a pre-trained n-gram language model to fuse it with prior linguistic knowledge. We then trained four neural network models (BOW, CNN, LSTM, and GRU) and a pre-trained model (ERNIE-Gram) on the LCQMCs train sets of varying size as well as the related augmented train sets produced by the two DA programs. The results show that there are no significant performance differences between the models trained on the two types of augmented train sets, both when the five DA techniques are applied together or separately. Moreover, due to the inability of the five DA techniques to make strictly paraphrastic augmented texts, the results indicate the need of sufficient amounts of training examples for the classification models trained on them to mediate the negative impact of false matching augmented text pairs and improve performance, a limitation of random text editing perturbations used as a DA approach. Similar results were also obtained for English.
Knowledge Base Question Answering (KBQA) aims to answer natural language questions with the help of an external knowledge base. The core idea is to find the link between the internal knowledge behind questions and known triples of the knowledge base. Traditional KBQA task pipelines contain several steps, including entity recognition, entity linking, answering selection, etc. In this kind of pipeline methods, errors in any procedure will inevitably propagate to the final prediction. To address this challenge, this paper proposes a Corpus Generation - Retrieve Method (CGRM) with Pre-training Language Model (PLM) for the KBQA task. The major novelty lies in the design of the new method, wherein our approach, the knowledge enhanced T5 (kT5) model aims to generate natural language QA pairs based on Knowledge Graph triples and directly solve the QA by retrieving the synthetic dataset. The new method can extract more information about the entities from PLM to improve accuracy and simplify the processes. We test our method on NLPCC-ICCPOL 2016 KBQA dataset, and the results show that our method improves the performance of KBQA and the out straight-forward method is competitive with the state-of-the-art.
The ability to generate natural-language questions with controlled complexity levels is highly desirable as it further expands the applicability of question generation. In this paper, we propose an end-to-end neural complexity-controllable question generation model, which incorporates a mixture of experts (MoE) as the selector of soft templates to improve the accuracy of complexity control and the quality of generated questions. The soft templates capture question similarity while avoiding the expensive construction of actual templates. Our method introduces a novel, cross-domain complexity estimator to assess the complexity of a question, taking into account the passage, the question, the answer and their interactions. The experimental results on two benchmark QA datasets demonstrate that our QG model is superior to state-of-the-art methods in both automatic and manual evaluation. Moreover, our complexity estimator is significantly more accurate than the baselines in both in-domain and out-domain settings.
Large pretrained language models (PreLMs) are revolutionizing natural language processing across all benchmarks. However, their sheer size is prohibitive for small laboratories or for deployment on mobile devices. Approaches like pruning and distillation reduce the model size but typically retain the same model architecture. In contrast, we explore distilling PreLMs into a different, more efficient architecture, Continual Multiplication of Words (CMOW), which embeds each word as a matrix and uses matrix multiplication to encode sequences. We extend the CMOW architecture and its CMOW/CBOW-Hybrid variant with a bidirectional component for more expressive power, per-token representations for a general (task-agnostic) distillation during pretraining, and a two-sequence encoding scheme that facilitates downstream tasks on sentence pairs, such as sentence similarity and natural language inference. Our matrix-based bidirectional CMOW/CBOW-Hybrid model is competitive to DistilBERT on question similarity and recognizing textual entailment, but uses only half of the number of parameters and is three times faster in terms of inference speed. We match or exceed the scores of ELMo for all tasks of the GLUE benchmark except for the sentiment analysis task SST-2 and the linguistic acceptability task CoLA. However, compared to previous cross-architecture distillation approaches, we demonstrate a doubling of the scores on detecting linguistic acceptability. This shows that matrix-based embeddings can be used to distill large PreLM into competitive models and motivates further research in this direction.
Search is one of the most common platforms used to seek information. However, users mostly get overloaded with results whenever they use such a platform to resolve their queries. Nowadays, direct answers to queries are being provided as a part of the search experience. The question-answer (QA) retrieval process plays a significant role in enriching the search experience. Most off-the-shelf Semantic Textual Similarity models work fine for well-formed search queries, but their performances degrade when applied to a domain-specific setting having incomplete or grammatically ill-formed search queries in prevalence. In this paper, we discuss a framework for calculating similarities between a given input query and a set of predefined questions to retrieve the question which matches to it the most. We have used it for the financial domain, but the framework is generalized for any domain-specific search engine and can be used in other domains as well. We use Siamese network [6] over Long Short-Term Memory (LSTM) [3] models to train a classifier which generates unnormalized and normalized similarity scores for a given pair of questions. Moreover, for each of these question pairs, we calculate three other similarity scores: cosine similarity between their average word2vec embeddings [15], cosine similarity between their sentence embeddings [7] generated using RoBERTa [17] and their customized fuzzy-match score. Finally, we develop a metaclassifier using Support Vector Machines [19] for combining these five scores to detect if a given pair of questions is similar. We benchmark our model's performance against existing State Of The Art (SOTA) models on Quora Question Pairs (QQP) dataset as well as a dataset specific to the financial domain.
We study the zero-shot transfer capabilities of text matching models on a massive scale, by self-supervised training on 140 source domains from community question answering forums in English. We investigate the model performances on nine benchmarks of answer selection and question similarity tasks, and show that all 140 models transfer surprisingly well, where the large majority of models substantially outperforms common IR baselines. We also demonstrate that considering a broad selection of source domains is crucial for obtaining the best zero-shot transfer performances, which contrasts the standard procedure that merely relies on the largest and most similar domains. In addition, we extensively study how to best combine multiple source domains. We propose to incorporate self-supervised with supervised multi-task learning on all available source domains. Our best zero-shot transfer model considerably outperforms in-domain BERT and the previous state of the art on six benchmarks. Fine-tuning of our model with in-domain data results in additional large gains and achieves the new state of the art on all nine benchmarks.
People increasingly search online for answers to their medical questions but the rate at which medical questions are asked online significantly exceeds the capacity of qualified people to answer them. This leaves many questions unanswered or inadequately answered. Many of these questions are not unique, and reliable identification of similar questions would enable more efficient and effective question answering schema. COVID-19 has only exacerbated this problem. Almost every government agency and healthcare organization has tried to meet the informational need of users by building online FAQs, but there is no way for people to ask their question and know if it is answered on one of these pages. While many research efforts have focused on the problem of general question similarity, these approaches do not generalize well to domains that require expert knowledge to determine semantic similarity, such as the medical domain. In this paper, we show how a double fine-tuning approach of pretraining a neural network on medical question-answer pairs followed by fine-tuning on medical question-question pairs is a particularly useful intermediate task for the ultimate goal of determining medical question similarity. While other pretraining tasks yield an accuracy below 78.7% on this task, our model achieves an accuracy of 82.6% with the same number of training examples, an accuracy of 80.0% with a much smaller training set, and an accuracy of 84.5% when the full corpus of medical question-answer data is used. We also describe a currently live system that uses the trained model to match user questions to COVID-related FAQs.
This paper describes our method for the task of Semantic Question Similarity in Arabic in the workshop on NLP Solutions for Under-Resourced Languages (NSURL). The aim is to build a model that is able to detect similar semantic questions in the Arabic language for the provided dataset. Different methods of determining questions similarity are explored in this work. The proposed models achieved high F1-scores, which range from (88% to 96%). Our official best result is produced from the ensemble model of using a pre-trained multilingual BERT model with different random seeds with 95.924% F1-Score, which ranks the first among nine participants teams.
Hate speech detection is a challenging problem with most of the datasets available in only one language: English. In this paper, we conduct a large scale analysis of multilingual hate speech in 9 languages from 16 different sources. We observe that in low resource setting, simple models such as LASER embedding with logistic regression performs the best, while in high resource setting BERT based models perform better. In case of zero-shot classification, languages such as Italian and Portuguese achieve good results. Our proposed framework could be used as an efficient solution for low-resource languages. These models could also act as good baselines for future multilingual hate speech detection tasks. We have made our code and experimental settings public for other researchers at https://github.com/punyajoy/DE-LIMIT.
In this paper, we describe our team's effort on the semantic text question similarity task of NSURL 2019. Our top performing system utilizes several innovative data augmentation techniques to enlarge the training data. Then, it takes ELMo pre-trained contextual embeddings of the data and feeds them into an ON-LSTM network with self-attention. This results in sequence representation vectors that are used to predict the relation between the question pairs. The model is ranked in the 1st place with 96.499 F1-score (same as the second place F1-score) and the 2nd place with 94.848 F1-score (differs by 1.076 F1-score from the first place) on the public and private leaderboards, respectively.
This paper describes the SemEval--2016 Task 3 on Community Question Answering, which we offered in English and Arabic. For English, we had three subtasks: Question--Comment Similarity (subtask A), Question--Question Similarity (B), and Question--External Comment Similarity (C). For Arabic, we had another subtask: Rerank the correct answers for a new question (D). Eighteen teams participated in the task, submitting a total of 95 runs (38 primary and 57 contrastive) for the four subtasks. A variety of approaches and features were used by the participating systems to address the different subtasks, which are summarized in this paper. The best systems achieved an official score (MAP) of 79.19, 76.70, 55.41, and 45.83 in subtasks A, B, C, and D, respectively. These scores are significantly better than those for the baselines that we provided. For subtask A, the best system improved over the 2015 winner by 3 points absolute in terms of Accuracy.
We describe SemEval-2017 Task 3 on Community Question Answering. This year, we reran the four subtasks from SemEval-2016:(A) Question-Comment Similarity,(B) Question-Question Similarity,(C) Question-External Comment Similarity, and (D) Rerank the correct answers for a new question in Arabic, providing all the data from 2015 and 2016 for training, and fresh data for testing. Additionally, we added a new subtask E in order to enable experimentation with Multi-domain Question Duplicate Detection in a larger-scale scenario, using StackExchange subforums. A total of 23 teams participated in the task, and submitted a total of 85 runs (36 primary and 49 contrastive) for subtasks A-D. Unfortunately, no teams participated in subtask E. A variety of approaches and features were used by the participating systems to address the different subtasks. The best systems achieved an official score (MAP) of 88.43, 47.22, 15.46, and 61.16 in subtasks A, B, C, and D, respectively. These scores are better than the baselines, especially for subtasks A-C.
The rate at which medical questions are asked online far exceeds the capacity of qualified people to answer them, and many of these questions are not unique. Identifying same-question pairs could enable questions to be answered more effectively. While many research efforts have focused on the problem of general question similarity for non-medical applications, these approaches do not generalize well to the medical domain, where medical expertise is often required to determine semantic similarity. In this paper, we show how a semi-supervised approach of pre-training a neural network on medical question-answer pairs is a particularly useful intermediate task for the ultimate goal of determining medical question similarity. While other pre-training tasks yield an accuracy below 78.7% on this task, our model achieves an accuracy of 82.6% with the same number of training examples, and an accuracy of 80.0% with a much smaller training set.