Guilin Qi

AAAI Conference 2026 Conference Paper

C³TG: Conflict-aware, Composite, and Collaborative Controlled Text Generation

• Yu Li
• Zhe Yang
• Yi Huang
• Xin Liu
• Guilin Qi

Recent advancements in large language models (LLMs) have demonstrated remarkable text generation capabilities. However, controlling specific attributes of generated text remains challenging without architectural modifications or extensive fine-tuning. Current methods typically toggle a single, basic attribute but struggle with precise multi-attribute control. In scenarios where attribute requirements conflict, existing methods lack coordination mechanisms, causing interference between desired attributes. Furthermore, these methods fail to incorporate iterative optimization processes in the controlled generation pipeline. To address these limitations, we propose Conflict-aware, Composite, and Collaborative Controlled Text Generation (C³TG), a two-phase framework for fine-grained, multi-dimensional text attribute control. During generation, C³TG selectively pairs the LLM with the required attribute classifiers from the 17 available dimensions and employs weighted KL-divergence to adjust token probabilities. The optimization phase then leverages an energy function combining classifier scores and penalty terms to resolve attribute conflicts through iterative feedback, enabling precise control over multiple dimensions simultaneously while preserving natural text flow. Experiments show that C³TG significantly outperforms baselines across multiple metrics including attribute accuracy, linguistic fluency, and output diversity, while simultaneously reducing toxicity. These results establish C³TG as an effective and flexible solution for multi-dimensional text attribute control that requires no costly model modifications.

AAAI Conference 2026 Conference Paper

Forget What Has Seen: Selective Concept Unlearning in Segmentation Foundation Models

• Miaozeng Du
• Jiaqi Li
• Sirui Pan
• Yi Zhan
• Guilin Qi
• Yuxin Zhang
• Rihui Jin
• Yinjia Shu

Machine unlearning (MU) has emerged as a critical tool for removing sensitive or personal information from machine learning models, empowering individuals with the right to be forgotten. While MU has achieved success in classification and generative tasks, whether this technique can be effectively applied to segmentation foundation models remains uncertain. To address this issue, we propose an efficient method, Selective Concept Unlearning (SCU), to unlearn the segmentation capability of target concepts. SCU consists of several key aspects: (1) The Multi-level Forgetting Module, designed with a hierarchical three-level suppression strategy, including (i) distillation-level: Negative distillation steers model’s output distribution away from teacher’s correct outputs, erasing its learned concept recognition. (ii) attention-level: Attention suppression minimizes model’s attention to target regions. (iii) output-level: Directly erases predictions for the target by relabeling as background. (2) The Preservation Module ensures maintaining segmentation quality for non-target concepts. Additionally, we introduce a set of metrics to evaluate segmentation unlearning methods. Experiments demonstrate that SCU consistently outperforms existing baselines.

AAAI Conference 2026 Conference Paper

TaxReasoning: Benchmarking Knowledge-Intensive Mathematical Reasoning with Evolving Tax Laws

• Nan Hu
• Yike Wu
• Jiaye Li
• HuiKang Hu
• Guilin Qi
• Songlin Zhai
• Yongrui Chen
• Tianxing Wu

Recent studies have explored the capabilities of large language models (LLMs) in solving knowledge-intensive mathematical reasoning problems. However, existing benchmarks predominantly involve static theorems that LLMs have encountered during pretraining, failing to assess dynamic knowledge integration. In this work, we introduce TaxReasoning, a novel benchmark designed to evaluate LLMs’ abilities in real-world tax calculation scenarios. These tasks require not only mathematical reasoning and numerical computation, but also the extraction and application of complex, frequently updated tax regulations. Through extensive experiments with state-of-the-art LLMs using diverse prompting strategies and knowledge augmentation techniques, we uncover substantial limitations in their ability to handle dynamic, knowledge-intensive questions—primarily due to missing domain-specific knowledge and ineffective retrieval. Even the best-performing models fall significantly short of human-level performance. Our analysis points to key avenues for improvement, including enhancing LLMs' reasoning capabilities, developing more effective knowledge summarization techniques, and improving retrieval strategies. TaxReasoning offers a critical testbed for advancing LLMs in dynamic knowledge-intensive domains.

KR Conference 2025 System Paper

Can LLMs Solve ASP Problems? Insights from a Benchmarking Study

• Lin Ren
• Guohui Xiao
• Guilin Qi
• Yishuai Geng
• Haohan Xue

Answer Set Programming (ASP) is a powerful paradigm for non-monotonic reasoning. Recently, large language models (LLMs) have demonstrated promising capabilities in logical reasoning. Despite this potential, current evaluations of LLM capabilities in ASP are often limited. Existing works normally employ overly simplified ASP programs, do not support negation, disjunction, or multiple answer sets. Furthermore, there is a lack of benchmarks that introduce tasks specifically designed for ASP solving. To bridge this gap, we introduce ASPBench, a comprehensive ASP benchmark, including three ASP specific tasks: ASP entailment, answer set verification, and answer set computation. Our extensive evaluations on ASPBench reveal that while 14 state-of-the-art LLMs, including deepseek-r1, o4-mini, and gemini-2. 5-flash-thinking, perform relatively well on the first two simpler tasks, they struggle with answer set computation, which is the core of ASP solving. These findings offer insights into the current limitations of LLMs in ASP solving. This highlights the need for new approaches that integrate symbolic reasoning capabilities more effectively. The code and dataset are available at https: //github. com/HomuraT/ASPBench.

AAAI Conference 2025 Conference Paper

HeGTa: Leveraging Heterogeneous Graph-enhanced Large Language Models for Few-shot Complex Table Understanding

• Rihui Jin
• Yu Li
• Guilin Qi
• Nan Hu
• Yuan-Fang Li
• Jiaoyan Chen
• Jianan Wang
• Yongrui Chen

Table Understanding (TU) has achieved promising advancements, but it faces the challenges of the scarcity of manually labeled tables and the presence of complex table structures. To address these challenges, we propose HeGTa, a heterogeneous graph (HG)-enhanced large language model (LLM) designed for few-shot TU tasks. This framework aligns structural table semantics with the LLM's parametric knowledge through soft prompts and instruction tuning. It also addresses complex tables with a multi-task pre-training scheme, incorporating three novel multi-granularity self-supervised HG pre-text tasks. We empirically demonstrate the effectiveness of HeGTa, showing that it outperforms the SOTA for few-shot complex TU on several benchmarks.

NeurIPS Conference 2025 Conference Paper

K-DeCore: Facilitating Knowledge Transfer in Continual Structured Knowledge Reasoning via Knowledge Decoupling

• Yongrui Chen
• Yi Huang
• Yunchang Liu
• Shenyu Zhang
• Junhao He
• Tongtong Wu
• Guilin Qi
• Tianxing Wu

Continual Structured Knowledge Reasoning (CSKR) focuses on training models to handle sequential tasks, where each task involves translating natural language questions into structured queries grounded in structured knowledge. Existing general continual learning approaches face significant challenges when applied to this task, including poor generalization to heterogeneous structured knowledge and inefficient reasoning due to parameter growth as tasks increase. To address these limitations, we propose a novel CSKR framework, \textsc{K-DeCore}, which operates with a fixed number of tunable parameters. Unlike prior methods, \textsc{K-DeCore} introduces a knowledge decoupling mechanism that disentangles the reasoning process into task-specific and task-agnostic stages, effectively bridging the gaps across diverse tasks. Building on this foundation, \textsc{K-DeCore} integrates a dual-perspective memory consolidation mechanism for distinct stages and introduces a structure-guided pseudo-data synthesis strategy to further enhance the model's generalization capabilities. Extensive experiments on four benchmark datasets demonstrate the superiority of \textsc{K-DeCore} over existing continual learning methods across multiple metrics, leveraging various backbone large language models.

IJCAI Conference 2025 Conference Paper

LLM4VKG: Leveraging Large Language Models for Virtual Knowledge Graph Construction

• Guohui Xiao
• Lin Ren
• Guilin Qi
• Haohan Xue
• Marco Di Panfilo
• Davide Lanti

Virtual Knowledge Graphs (VKGs) provide an effective solution for data integration but typically require significant expertise for their construction. This process, involving ontology development, schema analysis, and mapping creation, is often hindered by naming ambiguities and matching issues, which traditional rule-based methods struggle to address. Large language models (LLMs), with their ability to process and generate contextually relevant text, offer a potential solution. In this work, we introduce LLM4VKG, a novel framework that leverages LLMs to automatize VKG construction. Experimental evaluation on the RODI benchmark demonstrates that LLM4VKG surpasses state-of-the-art methods, achieving an average F1-score improvement of +17% and a peak gain of +39%. Moreover, LLM4VKG proves robust against incomplete ontologies and can handle complex mappings where current methods fail.

ICML Conference 2025 Conference Paper

Peripheral Memory for LLMs: Integration of Sequential Memory Banks with Adaptive Querying

• Songlin Zhai
• Yuan Meng
• Yongrui Chen 0002
• Yiwei Wang
• Guilin Qi

Large Language Models (LLMs) have revolutionized various natural language processing tasks with their remarkable capabilities. However, challenges persist in effectively integrating new knowledge into LLMs without compromising their performance, particularly in the Large Language Models (LLMs) have revolutionized various natural language processing tasks with their remarkable capabilities. However, a challenge persists in effectively processing new information, particularly in the area of long-term knowledge updates without compromising model performance. To address this challenge, this paper introduces a novel memory augmentation framework that conceptualizes memory as a peripheral component (akin to physical RAM), with the LLM serving as the information processor (analogous to a CPU). Drawing inspiration from RAM architecture, we design memory as a sequence of memory banks, each modeled using Kolmogorov-Arnold Network (KAN) to ensure smooth state transitions. Memory read and write operations are dynamically controlled by query signals derived from the LLMs’ internal states, closely mimicking the interaction between a CPU and RAM. Furthermore, a dedicated memory bank is used to generate a mask value that indicates the relevance of the retrieved data, inspired by the sign bit in binary coding schemes. The retrieved memory feature is then integrated as a prefix to enhance the model prediction. Extensive experiments on knowledge-based model editing validate the effectiveness and efficiency of our peripheral memory.

NeurIPS Conference 2025 Conference Paper

Uncertain Knowledge Graph Completion via Semi-Supervised Confidence Distribution Learning

• Tianxing Wu
• Shutong Zhu
• Jingting Wang
• Ning Xu
• Guilin Qi
• Haofen Wang

Uncertain knowledge graphs (UKGs) associate each triple with a confidence score to provide more precise knowledge representations. Recently, since real-world UKGs suffer from the incompleteness, uncertain knowledge graph (UKG) completion attracts more attention, aiming to complete missing triples and confidences. Current studies attempt to learn UKG embeddings to solve this problem, but they neglect the extremely imbalanced distributions of triple confidences. This causes that the learnt embeddings are insufficient to high-quality UKG completion. Thus, in this paper, to address the above issue, we propose a new semi-supervised Confidence Distribution Learning (ssCDL) method for UKG completion, where each triple confidence is transformed into a confidence distribution to introduce more supervision information of different confidences to reinforce the embedding learning process. ssCDL iteratively learns UKG embedding by relational learning on labeled data (i. e. , existing triples with confidences) and unlabeled data with pseudo labels (i. e. , unseen triples with the generated confidences), which are predicted by meta-learning to augment the training data and rebalance the distribution of triple confidences. Experiments on two UKG datasets demonstrate that ssCDL consistently outperforms the state-of-the-art baselines in different evaluation metrics.

NeurIPS Conference 2024 Conference Paper

Single Image Unlearning: Efficient Machine Unlearning in Multimodal Large Language Models

• Jiaqi Li
• Qianshan Wei
• Chuanyi Zhang
• Guilin Qi
• Miaozeng Du
• Yongrui Chen
• Sheng Bi
• Fan Liu

Machine unlearning (MU) empowers individuals with the `right to be forgotten' by removing their private or sensitive information encoded in machine learning models. However, it remains uncertain whether MU can be effectively applied to Multimodal Large Language Models (MLLMs), particularly in scenarios of forgetting the leaked visual data of concepts. To overcome the challenge, we propose an efficient method, Single Image Unlearning (SIU), to unlearn the visual recognition of a concept by fine-tuning a single associated image for few steps. SIU consists of two key aspects: (i) Constructing Multifaceted fine-tuning data. We introduce four targets, based on which we construct fine-tuning data for the concepts to be forgotten; (ii) Joint training loss. To synchronously forget the visual recognition of concepts and preserve the utility of MLLMs, we fine-tune MLLMs through a novel Dual Masked KL-divergence Loss combined with Cross Entropy loss. Alongside our method, we establish MMUBench, a new benchmark for MU in MLLMs and introduce a collection of metrics for its evaluation. Experimental results on MMUBench show that SIU completely surpasses the performance of existing methods. Furthermore, we surprisingly find that SIU can avoid invasive membership inference attacks and jailbreak attacks. To the best of our knowledge, we are the first to explore MU in MLLMs. We will release the code and benchmark in the near future.

AAAI Conference 2023 Conference Paper

Learn from Yesterday: A Semi-supervised Continual Learning Method for Supervision-Limited Text-to-SQL Task Streams

• Yongrui Chen
• Xinnan Guo
• Tongtong Wu
• Guilin Qi
• Yang Li
• Yang Dong

Conventional text-to-SQL studies are limited to a single task with a fixed-size training and test set. When confronted with a stream of tasks common in real-world applications, existing methods struggle with the problems of insufficient supervised data and high retraining costs. The former tends to cause overfitting on unseen databases for the new task, while the latter makes a full review of instances from past tasks impractical for the model, resulting in forgetting of learned SQL structures and database schemas. To address the problems, this paper proposes integrating semi-supervised learning (SSL) and continual learning (CL) in a stream of text-to-SQL tasks and offers two promising solutions in turn. The first solution Vanilla is to perform self-training, augmenting the supervised training data with predicted pseudo-labeled instances of the current task, while replacing the full volume retraining with episodic memory replay to balance the training efficiency with the performance of previous tasks. The improved solution SFNet takes advantage of the intrinsic connection between CL and SSL. It uses in-memory past information to help current SSL, while adding high-quality pseudo instances in memory to improve future replay. The experiments on two datasets shows that SFNet outperforms the widely-used SSL-only and CL-only baselines on multiple metrics.

NeurIPS Conference 2023 Conference Paper

Parameterizing Context: Unleashing the Power of Parameter-Efficient Fine-Tuning and In-Context Tuning for Continual Table Semantic Parsing

• Yongrui Chen
• Shenyu Zhang
• Guilin Qi
• Xinnan Guo

Continual table semantic parsing aims to train a parser on a sequence of tasks, where each task requires the parser to translate natural language into SQL based on task-specific tables but only offers limited training examples. Conventional methods tend to suffer from overfitting with limited supervision, as well as catastrophic forgetting due to parameter updates. Despite recent advancements that partially alleviate these issues through semi-supervised data augmentation and retention of a few past examples, the performance is still limited by the volume of unsupervised data and stored examples. To overcome these challenges, this paper introduces a novel method integrating parameter-efficient fine-tuning (PEFT) and in-context tuning (ICT) for training a continual table semantic parser. Initially, we present a task-adaptive PEFT framework capable of fully circumventing catastrophic forgetting, which is achieved by freezing the pre-trained model backbone and fine-tuning small-scale prompts. Building on this, we propose a teacher-student framework-based solution. The teacher addresses the few-shot problem using ICT, which procures contextual information by demonstrating a few training examples. In turn, the student leverages the proposed PEFT framework to learn from the teacher's output distribution, and subsequently compresses and saves the contextual information to the prompts, eliminating the need to store any training examples. Experimental evaluations on two benchmarks affirm the superiority of our method over prevalent few-shot and continual learning baselines across various metrics.

IJCAI Conference 2022 Conference Paper

Improving Few-Shot Text-to-SQL with Meta Self-Training via Column Specificity

• Xinnan Guo
• Yongrui Chen
• Guilin Qi
• Tianxing Wu
• Hao Xu

The few-shot problem is an urgent challenge for single-table text-to-SQL. Existing methods ignore the potential value of unlabeled data, and merely rely on a coarse-grained Meta-Learning (ML) algorithm that neglects the differences of column contributions to the optimization object. This paper proposes a Meta Self-Training text-to-SQL (MST-SQL) method to solve the problem. Specifically, MST-SQL is based on column-wise HydraNet and adopts self-training as an effective mechanism to learn from readily available unlabeled samples. During each epoch of training, it first predicts pseudo-labels for unlabeled samples and then leverages them to update the parameters. A fine-grained ML algorithm is used in updating, which weighs the contribution of columns by their specificity, in order to further improve the generalizability. Extensive experimental results on both open-domain and domain-specific benchmarks reveal that our MST-SQL has significant advantages in few-shot scenarios, and is also competitive in standard supervised settings.

AILAW Journal 2022 Journal Article

Judicial knowledge-enhanced magnitude-aware reasoning for numerical legal judgment prediction

• Sheng Bi
• Zhiyao Zhou
• Lu Pan
• Guilin Qi

Abstract Legal Judgment Prediction (LJP) is an essential component of legal assistant systems, which aims to automatically predict judgment results from a given criminal fact description. As a vital subtask of LJP, researchers have paid little attention to the numerical LJP, i. e. , the prediction of imprisonment and penalty. Existing methods ignore numerical information in the criminal facts, making their performances far from satisfactory. For instance, the amount of theft varies, as do the prison terms and penalties. The major challenge is how the model can obtain the ability of numerical comparison and magnitude perception, e. g. , 400 < 500 < 800, 500 is closer to 400 than to 800. To this end, we propose a judicial knowledge-enhanced magnitude-aware reasoning architecture, called NumLJP, for the numerical LJP task. Specifically, we first implement a contrastive learning-based judicial knowledge selector to distinguish confusing criminal cases efficiently. Unlike previous approaches that employ the law article as external knowledge, judicial knowledge is a quantitative guideline in real scenarios. It contains many numerals (called anchors) that can construct a reference frame. Then we design a masked numeral prediction task to help the model remember these anchors to acquire legal numerical commonsense from the selected judicial knowledge. We construct a scale-based numerical graph using the anchors and numerals in facts to perform magnitude-aware numerical reasoning. Finally, the representations of fact description, judicial knowledge, and numerals are fused to make decisions. We conduct extensive experiments on three real-world datasets and select several competitive baselines. The results demonstrate that the macro-F1 of NumLJP improves by at least 9. 53% and 11. 57% on the prediction of penalty and imprisonment, respectively.

ICLR Conference 2022 Conference Paper

Pretrained Language Model in Continual Learning: A Comparative Study

• Tongtong Wu
• Massimo Caccia
• Zhuang Li 0001
• Yuan-Fang Li
• Guilin Qi
• Gholamreza Haffari

Continual learning (CL) is a setting in which a model learns from a stream of incoming data while avoiding to forget previously learned knowledge. Pre-trained language models (PLMs) have been successfully employed in continual learning of different natural language problems. With the rapid development of many continual learning methods and PLMs, understanding and disentangling their interactions become essential for continued improvement of continual learning performance. In this paper, we thoroughly compare the continual learning performance over the combination of 5 PLMs and 4 CL approaches on 3 benchmarks in 2 typical incremental settings. Our extensive experimental analyses reveal interesting performance differences across PLMs and across CL methods. Furthermore, our representativeness probing analyses dissect PLMs’ performance characteristics in a layer-wise and task-wise manner, uncovering the extent to which their inner layers suffer from forgetting, and the effect of different CL approaches on each layer. Finally, our observations and analyses open up a number of important research questions that will inform and guide the design of effective continual learning techniques.

AAAI Conference 2021 Conference Paper

Curriculum-Meta Learning for Order-Robust Continual Relation Extraction

• Tongtong Wu
• Xuekai Li
• Yuan-Fang Li
• Gholamreza Haffari
• Guilin Qi
• Yujin Zhu
• Guoqiang Xu

Continual relation extraction is an important task that focuses on extracting new facts incrementally from unstructured text. Given the sequential arrival order of the relations, this task is prone to two serious challenges, namely catastrophic forgetting and order-sensitivity. We propose a novel curriculummeta learning method to tackle the above two challenges in continual relation extraction. We combine meta learning and curriculum learning to quickly adapt model parameters to a new task and to reduce interference of previously seen tasks on the current task. We design a novel relation representation learning method through the distribution of domain and range types of relations. Such representations are utilized to quantify the difficulty of tasks for the construction of curricula. Moreover, we also present novel difficultybased metrics to quantitatively measure the extent of ordersensitivity of a given model, suggesting new ways to evaluate model robustness. Our comprehensive experiments on three benchmark datasets show that our proposed method outperforms the state-of-the-art techniques. The code is available at https: //github. com/wutong8023/AAAI-CML.

AAAI Conference 2021 Conference Paper

Leveraging Table Content for Zero-shot Text-to-SQL with Meta-Learning

• Yongrui Chen
• Xinnan Guo
• Chaojie Wang
• Jian Qiu
• Guilin Qi
• Meng Wang
• Huiying Li

Single-table text-to-SQL aims to transform a natural language question into a SQL query according to one single table. Recent work has made promising progress on this task by pretrained language models and a multi-submodule framework. However, zero-shot table, that is, the invisible table in the training set, is currently the most critical bottleneck restricting the application of existing approaches to real-world scenarios. Although some work has utilized auxiliary tasks to help handle zero-shot tables, expensive extra manual annotation limits their practicality. In this paper, we propose a new approach for the zero-shot text-to-SQL task which does not rely on any additional manual annotations. Our approach consists of two parts. First, we propose a new model that leverages the abundant information of table content to help establish the mapping between questions and zero-shot tables. Further, we propose a simple but efficient meta-learning strategy to train our model. The strategy utilizes the two-step gradient update to force the model to learn a generalization ability towards zero-shot tables. We conduct extensive experiments on a public open-domain text-to-SQL dataset WikiSQL and a domain-specific dataset ESQL. Compared to existing approaches using the same pre-trained model, our approach achieves significant improvements on both datasets. Compared to the larger pre-trained model and the tabular-specific pre-trained model, our approach is still competitive. More importantly, on the zero-shot subsets of both the datasets, our approach further increases the improvements.

AAAI Conference 2021 Conference Paper

Towards Balanced Defect Prediction with Better Information Propagation

• Xianda Zheng
• Yuan-Fang Li
• Huan Gao
• Yuncheng Hua
• Guilin Qi

Defect prediction, the task of predicting the presence of defects in source code artifacts, has broad application in software development. Defect prediction faces two major challenges, label scarcity, where only a small percentage of code artifacts are labeled, and data imbalance, where the majority of labeled artifacts are non-defective. Moreover, current defect prediction methods ignore the impact of information propagation among code artifacts, and this negligence leads to performance degradation. In this paper, we propose DPCAG, a novel model to address the above three issues. We treat code artifacts as nodes in a graph, and learn to propagate influence among neighboring nodes iteratively in an EM framework. DPCAG dynamically adjusts the contributions of each node and selects high-confidence nodes for data augmentation. Experimental results on real-world benchmark datasets show that DPCAG improves performance compare to the state-of-the-art models. In particular, DPCAG achieves substantial performance superiority when measured by Matthews Correlation Coefficient (MCC), a metric that is widely acknowledged to be the most suitable for imbalanced data.

ECAI Conference 2020 Conference Paper

A Hash Learning Framework for Search-Oriented Knowledge Graph Embedding

• Meng Wang 0009
• Tongtong Wu
• Guilin Qi

Knowledge graph representation learning, also called knowledge graph embedding, is the task of mapping entities and relations into a low-dimensional, continuous vector space, and, as a result, can support various machine learning models to perform knowledge completion tasks with good performance and robustness. However, most of existing embedding models focus on improving the link prediction accuracy while ignoring the time-efficiency in search-intensive applications over large-scale knowledge graphs. To tackle this problem, in this paper, we encode knowledge graph into Hamming space and introduce a novel HAsh Learning Framework (HALF) for search-oriented knowledge graph embedding. The proposed method can be applied to recent various knowledge graph embedding models for accelerating the computation of searching embeddings by utilizing the bitwise operations (XNOR and Bitcount). Experimental results on benchmark datasets demonstrate the effectiveness of our proposed method, which gets a bonus of speed-up in the searching embeddings while the accuracy and scalability of the original model are basically maintained.

IJCAI Conference 2020 Conference Paper

Formal Query Building with Query Structure Prediction for Complex Question Answering over Knowledge Base

• Yongrui Chen
• Huiying Li
• Yuncheng Hua
• Guilin Qi

Formal query building is an important part of complex question answering over knowledge bases. It aims to build correct executable queries for questions. Recent methods try to rank candidate queries generated by a state-transition strategy. However, this candidate generation strategy ignores the structure of queries, resulting in a considerable number of noisy queries. In this paper, we propose a new formal query building approach that consists of two stages. In the first stage, we predict the query structure of the question and leverage the structure to constrain the generation of the candidate queries. We propose a novel graph generation framework to handle the structure prediction task and design an encoder-decoder model to predict the argument of the predetermined operation in each generative step. In the second stage, we follow the previous methods to rank the candidate queries. The experimental results show that our formal query building approach outperforms existing methods on complex questions while staying competitive on simple questions.

AAAI Conference 2020 Conference Paper

One-Shot Learning for Long-Tail Visual Relation Detection

• Weitao Wang
• Meng Wang
• Sen Wang
• Guodong Long
• Lina Yao
• Guilin Qi
• Yang Chen

The aim of visual relation detection is to provide a comprehensive understanding of an image by describing all the objects within the scene, and how they relate to each other, in form; for example, . This ability is vital for image captioning, visual question answering, and many other applications. However, visual relationships have long-tailed distributions and, thus, the limited availability of training samples is hampering the practicability of conventional detection approaches. With this in mind, we designed a novel model for visual relation detection that works in one-shot settings. The embeddings of objects and predicates are extracted through a network that includes a feature-level attention mechanism. Attention alleviates some of the problems with feature sparsity, and the resulting representations capture more discriminative latent features. The core of our model is a dual graph neural network that passes and aggregates the context information of predicates and objects in an episodic training scheme to improve recognition of the one-shot predicates and then generate the triplets. To the best of our knowledge, we are the first to center on the viability of one-shot learning for visual relation detection. Extensive experiments on two newly-constructed datasets show that our model significantly improved the performance of two tasks PredCls and SGCls from 2. 8% to 12. 2% compared with state-of-the-art baselines.

IJCAI Conference 2020 Conference Paper

Retrieve, Program, Repeat: Complex Knowledge Base Question Answering via Alternate Meta-learning

• Yuncheng Hua
• Yuan-Fang Li
• Gholamreza Haffari
• Guilin Qi
• Wei Wu

A compelling approach to complex question answering is to convert the question to a sequence of actions, which can then be executed on the knowledge base to yield the answer, aka the programmer-interpreter approach. Use similar training questions to the test question, meta-learning enables the programmer to adapt to unseen questions to tackle potential distributional biases quickly. However, this comes at the cost of manually labeling similar questions to learn a retrieval model, which is tedious and expensive. In this paper, we present a novel method that automatically learns a retrieval model alternately with the programmer from weak supervision, i. e. , the system’s performance with respect to the produced answers. To the best of our knowledge, this is the first attempt to train the retrieval model with the programmer jointly. Our system leads to state-of-the-art performance on a large-scale task for complex question answering over knowledge bases. We have released our code at https: //github. com/DevinJake/MARL.

KER Journal 2018 Journal Article

A survey of large-scale reasoning on the Web of data

• Grigoris Antoniou
• Sotiris Batsakis
• Raghava Mutharaju
• Jeff Z. Pan
• Guilin Qi
• Ilias Tachmazidis
• Jacopo Urbani
• Zhangquan Zhou

Abstract As more and more data is being generated by sensor networks, social media and organizations, the Web interlinking this wealth of information becomes more complex. This is particularly true for the so-called Web of Data, in which data is semantically enriched and interlinked using ontologies. In this large and uncoordinated environment, reasoning can be used to check the consistency of the data and of associated ontologies, or to infer logical consequences which, in turn, can be used to obtain new insights from the data. However, reasoning approaches need to be scalable in order to enable reasoning over the entire Web of Data. To address this problem, several high-performance reasoning systems, which mainly implement distributed or parallel algorithms, have been proposed in the last few years. These systems differ significantly; for instance in terms of reasoning expressivity, computational properties such as completeness, or reasoning objectives. In order to provide a first complete overview of the field, this paper reports a systematic review of such scalable reasoning approaches over various ontological languages, reporting details about the methods and over the conducted experiments. We highlight the shortcomings of these approaches and discuss some of the open problems related to performing scalable reasoning.

IJCAI Conference 2018 Conference Paper

Fine-grained Image Classification by Visual-Semantic Embedding

• Huapeng Xu
• Guilin Qi
• Jingjing Li
• Meng Wang
• Kang Xu
• Huan Gao

This paper investigates a challenging problem, which is known as fine-grained image classification(FGIC). Different from conventional computer visionproblems, FGIC suffers from the large intraclassdiversities and subtle inter-class differences. Existing FGIC approaches are limited to exploreonly the visual information embedded in the images. In this paper, we present a novel approachwhich can use handy prior knowledge from eitherstructured knowledge bases or unstructured text tofacilitate FGIC. Specifically, we propose a visual-semanticembedding model which explores semanticembedding from knowledge bases and text, andfurther trains a novel end-to-end CNN frameworkto linearly map image features to a rich semanticembedding space. Experimental results on a challenginglarge-scale UCSD Bird-200-2011 datasetverify that our approach outperforms several state-of-the-art methods with significant advances.

ECAI Conference 2016 Conference Paper

A Joint Model for Sentiment-Aware Topic Detection on Social Media

• Kang Xu 0001
• Guilin Qi
• Junheng Huang
• Tianxing Wu 0001

Joint sentiment/topic models are widely applied in detecting sentiment-aware topics on the lengthy review data and they are achieved with Latent Dirichlet Allocation (LDA) based model. Nowadays plenty of user-generated posts, e. g. , tweets and E-commerce short reviews, are published on the social media and the posts imply the public's sentiments (i. e. , positive and negative) towards various topics. However, the existing sentiment/topic models are not applicable to detect sentiment-aware topics on the posts, i. e. , short texts, because applying the models to the short texts directly will suffer from the context sparsity problem. In this paper, we propose a Time-User Sentiment/Topic Latent Dirichlet Allocation (TUS-LDA) which aggregates posts in the same timeslice or user as a pseudo-document to alleviate the context sparsity problem. Moreover, we design approaches for parameter inference and incorporating prior knowledge into TUS-LDA. Experiments on the Sentiment140 and tweets of electronic products from Twitter7 show that TUS-LDA outperforms previous models in the tasks of sentiment classification and sentiment-aware topic extraction. Finally, we visualize the sentiment-aware topics discovered by TUS-LDA.

AAAI Conference 2016 Conference Paper

Cross-Lingual Taxonomy Alignment with Bilingual Biterm Topic Model

• Tianxing Wu
• Guilin Qi
• Haofen Wang
• Kang Xu
• Xuan Cui

As more and more multilingual knowledge becomes available on the Web, knowledge sharing across languages has become an important task to benefit many applications. One of the most crucial kinds of knowledge on the Web is taxonomy, which is used to organize and classify the Web data. To facilitate knowledge sharing across languages, we need to deal with the problem of cross-lingual taxonomy alignment, which discovers the most relevant category in the target taxonomy of one language for each category in the source taxonomy of another language. Current approaches for aligning crosslingual taxonomies strongly rely on domain-specific information and the features based on string similarities. In this paper, we present a new approach to deal with the problem of cross-lingual taxonomy alignment without using any domain-specific information. We first identify the candidate matched categories in the target taxonomy for each category in the source taxonomy using the crosslingual string similarity. We then propose a novel bilingual topic model, called Bilingual Biterm Topic Model (BiBTM), to perform exact matching. BiBTM is trained by the textual contexts extracted from the Web. We conduct experiments on two kinds of real world datasets. The experimental results show that our approach significantly outperforms the designed state-of-the-art comparison methods.

JAIR Journal 2016 Journal Article

DL-Lite Contraction and Revision

• Zhiqiang Zhuang
• Zhe Wang
• Kewen Wang
• Guilin Qi

Two essential tasks in managing description logic knowledge bases are eliminating problematic axioms and incorporating newly formed ones. Such elimination and incorporation are formalised as the operations of contraction and revision in belief change. In this paper, we deal with contraction and revision for the DL-Lite family through a model-theoretic approach. Standard description logic semantics yields an infinite number of models for DL-Lite knowledge bases, thus it is difficult to develop algorithms for contraction and revision that involve DL models. The key to our approach is the introduction of an alternative semantics called type semantics which can replace the standard semantics in characterising the standard inference tasks of DL-Lite. Type semantics has several advantages over the standard one. It is more succinct and importantly, with a finite signature, the semantics always yields a finite number of models. We then define model-based contraction and revision functions for DL-Lite knowledge bases under type semantics and provide representation theorems for them. Finally, the finiteness and succinctness of type semantics allow us to develop tractable algorithms for instantiating the functions.

ECAI Conference 2016 Conference Paper

Exploring Parallel Tractability of Ontology Materialization

• Zhangquan Zhou
• Guilin Qi
• Birte Glimm

Materialization is an important reasoning service for applications built on the Web Ontology Language (OWL). To make materialization efficient in practice, current research focuses on deciding tractability of an ontology language and designing parallel reasoning algorithms. However, some well-known large-scale ontologies, such as YAGO, have been shown to have good performance for parallel reasoning, but they are expressed in ontology languages that are not parallelly tractable, i. e. , the reasoning is inherently sequential in the worst case. This motivates us to study the problem of parallel tractability of ontology materialization from a theoretical perspective. That is, we aim to identify the ontologies for which materialization is parallelly tractable, i. e. , in NC complexity. In this work, we focus on datalog rewritable ontology languages. We identify several classes of datalog rewritable ontologies (called parallelly tractable classes) such that materialization over them is parallelly tractable. We further investigate the parallel tractability of materialization of a datalog rewritable OWL fragment DHL (Description Horn Logic) and an extension of DHL that allows complex role inclusion axioms. Based on the above results, we analyze real-world datasets and show that many ontologies expressed in DHL or its extension belong to the parallelly tractable classes.

AAAI Conference 2015 Conference Paper

Approximating Model-Based ABox Revision in DL-Lite: Theory and Practice

• Guilin Qi
• Zhe Wang
• Kewen Wang
• Xuefeng Fu
• Zhiqiang Zhuang

Model-based approaches provide a semantically well justified way to revise ontologies. However, in general, model-based revision operators are limited due to lack of efficient algorithms and inexpressibility of the revision results. In this paper, we make both theoretical and practical contribution to efficient computation of model-based revisions in DL-Lite. Specifically, we show that maximal approximations of two well-known model-based revisions for DL-LiteR can be computed using a syntactic algorithm. However, such a coincidence of model-based and syntactic approaches does not hold when role functionality axioms are allowed. As a result, we identify conditions that guarantee such a coincidence for DL-LiteFR. Our result shows that both model-based and syntactic revisions can co-exist seamlessly and the advantages of both approaches can be taken in one revision operator. Based on our theoretical results, we develop a graph-based algorithm for the revision operators and thus graph database techniques can be used to compute ontology revisions. Preliminary evaluation results show that the graph-based algorithm can efficiently handle revision of practical ontologies with large data.

AAAI Conference 2015 Conference Paper

Instance-Driven Ontology Evolution in DL-Lite

• Zhe Wang
• Kewen Wang
• Zhiqiang Zhuang
• Guilin Qi

The development and maintenance of large and complex ontologies are often time-consuming and error-prone. Thus, automated ontology learning and evolution have attracted intensive research interest. In data-centric applications where ontologies are designed from the data or automatically learnt from it, when new data instances are added that contradict the ontology, it is often desirable to incrementally revise the ontology according to the added data. In description logics, this problem can be intuitively formulated as the operation of TBox contraction, i. e. , rational elimination of certain axioms from the logical consequences of a TBox, and it is w. r. t. an ABox. In this paper we introduce a model-theoretic approach to such a contraction problem by using an alternative semantic characterisation of DL-Lite TBoxes. We show that entailment checking (without necessarily first computing the contraction result) is in coNP, which does not shift the corresponding complexity in propositional logic, and the problem is tractable when the size of the new data is bounded.

AAAI Conference 2014 Conference Paper

Contraction and Revision over DL-Lite TBoxes

• Zhiqiang Zhuang
• Zhe Wang
• Kewen Wang
• Guilin Qi

Two essential tasks in managing Description Logic (DL) ontologies are eliminating problematic axioms and incorporating newly formed axioms. Such elimination and incorporation are formalised as the operations of contraction and revision in belief change. In this paper, we deal with contraction and revision for the DL-Lite family through a model-theoretic approach. Standard DL semantics yields infinite numbers of models for DL-Lite TBoxes, thus it is not practical to develop algorithms for contraction and revision that involve DL models. The key to our approach is the introduction of an alternative semantics called type semantics which is more succinct than DL semantics. More importantly, with a finite signature, type semantics always yields finite humber of models. We then define model-based contraction and revision for DL-Lite TBoxes under type semantics and provide representation theorems for them. Finally, the succinctness of type semantics allows us to develop tractable algorithms for both operations.

AAAI Conference 2012 Conference Paper

A New Operator for ABox Revision in DL-Lite

• Sibei Gao
• Guilin Qi
• Haofen Wang

In this paper, we propose a new operator for revising ABoxes in DL-Lite ontologies. We present a graphbased algorithm for ABox revision in DL-Lite, which implements the revision operator and we show it runs in polynomial time

AAAI Conference 2012 Conference Paper

Conflict-Based Belief Revision Operators in Possibilistic Logic

• Guilin Qi
• Kewen Wang

In this paper, we investigate belief revision in possibilistic logic, which is a weighted logic proposed to deal with incomplete and uncertain information. Existing revision operators in possibilistic logic are restricted in the sense that the input information can only be a formula instead of a possibilistic knowledge base which is a set of weighted formulas. To break this restriction, we consider weighted prime implicants of a possibilistic knowledge base and use them to define novel revision operators in possibilistic logic. Intuitively, a weighted prime implicant of a possibilistic knowledge base is a logically weakest possibilistic term (i. e. , a set of weighted literals) that can entail the knowledge base. We first show that the existing definition of a weighted prime implicant is problematic and need a modification. To define a revision operator using weighted prime implicants, we face two problems. The first problem is that we need to define the notion of a conflict set between two weighted prime implicants of two possibilistic knowledge bases to achieve minimal change. The second problem is that we need to define the disjunction of possibilistic terms. We solve these problems and define two conflict-based revision operators in possibilistic logic. We then adapt the well-known postulates for revision proposed by Katsuno and Mendelzon and show that our revision operators satisfy four of the basic adapted postulates and satisfy two others in some special cases.

AAAI Conference 2012 Conference Paper

Large Scale Temporal RDFS Reasoning Using MapReduce

• Chang Liu
• Guilin Qi
• Yong Yu

In this work, we build a large scale reasoning engine under temporal RDFS semantics using MapReduce. We identify the major challenges of applying MapReduce framework to reason over temporal information, and present our solutions to tackle them.

ECAI Conference 2012 Conference Paper

Reasoning with Fuzzy-EL+ Ontologies Using MapReduce

• Zhangquan Zhou
• Guilin Qi
• Chang Liu 0021
• Pascal Hitzler
• Raghava Mutharaju

Fuzzy extension of Description Logics (DLs) allows the formal representation and handling of fuzzy knowledge. In this paper, we consider fuzzy-EL+, which is a fuzzy extension of EL+. We first present revised completion rules for fuzzy-EL+that can be handled by MapReduce programs. We then propose an algorithm for scale reasoning with fuzzy-EL+ontologies based on MapReduce.

AAAI Conference 2011 Conference Paper

Towards Practical ABox Abduction in Large OWL DL Ontologies

• Jianfeng Du
• Guilin Qi
• Yi-Dong Shen
• Jeff Pan

ABox abduction is an important aspect for abductive reasoning in Description Logics (DLs). It finds all minimal sets of ABox axioms that should be added to a background ontology to enforce entailment of a specified set of ABox axioms. As far as we know, by now there is only one ABox abduction method in expressive DLs computing abductive solutions with certain minimality. However, the method targets an ABox abduction problem that may have infinitely many abductive solutions and may not output an abductive solution in finite time. Hence, in this paper we propose a new ABox abduction problem which has only finitely many abductive solutions and also propose a novel method to solve it. The method reduces the original problem to an abduction problem in logic programming and solves it with Prolog engines. Experimental results show that the method is able to compute abductive solutions in benchmark OWL DL ontologies with large ABoxes.

KR Conference 2010 Conference Paper

Computing Inconsistency Measurements under Multi-Valued Semantics by Partial Max-SAT solvers

• Guohui Xiao
• Yue Ma
• Guilin Qi
• Zuoquan Lin

consistent subsets of formulas (Knight 2002) or minimal inMeasuring the inconsistency degree of a knowledge base can help us to deal with inconsistencies. Several inconsistency measures have been given under different multi-valued semantics, including 4-valued semantics, 3-valued semantics, LPm and Quasi Classical semantics. In this paper, we first carefully analyze the relationship between these inconsistency measures by showing that the inconsistency degrees under 4-valued semantics, 3-value semantics, LPm are the same, but different from the one based on Quasi Classical semantics. We then consider the computation of these inconsistency measures and show that computing inconsistency measurement under multi-valued semantics is usually intractable. To tackle this problem, we propose two novel algorithms that respectively encode the problems of computing inconsistency degrees under 4-valued semantics (3-valued semantics, LPm) and under Quasi Classical semantics into the partial MaxSAT problems. We implement these algorithms and do experiments on some benchmark data sets. The preliminary but encouraging experimental results show that our approach is efficient to handle large knowledge bases.

UAI Conference 2010 Conference Paper

Merging Knowledge Bases in Possibilistic Logic by Lexicographic Aggregation

• Guilin Qi
• Jianfeng Du
• Weiru Liu
• David A. Bell

Belief merging is an important but difficult problem in Artificial Intelligence, especially when sources of information are pervaded with uncertainty. Many merging operators have been proposed to deal with this problem in possibilistic logic, a weighted logic which is powerful for handling inconsistency and dealing with uncertainty. They often result in a possibilistic knowledge base which is a set of weighted formulas. Although possibilistic logic is inconsistency tolerant, it suffers from the well-known “drowning effect”. Therefore, we may still want to obtain a consistent possibilistic knowledge base as the result of merging. In such a case, we argue that it is not always necessary to keep weighted information after merging. In this paper, we define a merging operator that maps a set of possibilistic knowledge bases and a formula representing the integrity constraints to a classical knowledge base by using lexicographic ordering. We show that it satisfies nine postulates that generalize basic postulates for propositional merging given in [11]. These postulates capture the principle of minimal change in some sense. We then provide an algorithm for generating the resulting knowledge base of our merging operator. Finally, we discuss the compatibility of our merging operator with propositional merging and establish the advantage of our merging operator over existing semantic merging operators in the propositional case.

KR Conference 2010 Conference Paper

Preferential Semantics for Plausible Subsumption in Possibility Theory

• Guilin Qi
• Zhizheng Zhang

Handling exceptions in a knowledge-based system has been considered as an important issue in many domains of applications, such as medical domain. In this paper, we propose several preferential semantics for plausible subsumption to deal with exceptions in description logic-based knowledge bases. Our preferential semantics are defined in the framework of possibility theory, which is an uncertainty theory devoted to the handling of incomplete information. We consider the properties of these semantics and their relationships. Entailment of these plausible subsumption relative to a knowledge base is also considered. We show the close relationship between two of our semantics and the mutually dual preferential semantics given by Britz, Heidema and Meyer. Finally, we show that our semantics for plausible subsumption can be reduced to standard semantics of an expressive description logic. Thus, the problem of plausible subsumption checking under our semantics can be reduced to the problem of subsumption checking under the classical semantics.

IJCAI Conference 2009 Conference Paper

• Guilin Qi
• Jianfeng Du

The problem of revising an ontology consistently is closely related to the problem of belief revision which has been widely discussed in the literature. Some syntax-based belief revision operators have been adapted to revise ontologies in Description Logics (DLs). However, these operators remove the whole axioms to resolve logical contradictions and thus are not fine-grained. In this paper, we propose three model-based revision operators to revise terminologies in DLs. We show that one of them is more rational than others by comparing their logical properties. Therefore, we focus on this revision operator. We also consider the problem of computing the result of revision by our operator with the help of the notion of concept forgetting. Finally, we analyze the computational complexity of our revision operator.

AAAI Conference 2008 Conference Paper

A Semantic Approach for Iterated Revision in Possibilistic Logic

• Guilin Qi

In this paper, we propose a new approach for iterated revision in possibilistic logic by applying a one-step revision operator. We first argue that the set of KM postulates for revision is too strong to define a practical one-step revision operator and some of them should be weakened. We then present a semantic approach for iterated revision in possibilistic logic using a one-step revision operator. The computation of the semantic approach is given. We show that our revision approach satisfies almost all the DP postulates for iterated revision and some other important logical properties.

ECAI Conference 2008 Conference Paper

Prime Implicate-based Belief Revision Operators

• Meghyn Bienvenu
• Andreas Herzig
• Guilin Qi

AAAI Conference 2007 Conference Paper

A Model-based Approach for Merging Prioritized Knowledge Bases in Possibilistic Logic

• Guilin Qi

This paper presents a new approach for merging prioritized knowledge bases in possibilistic logic. Our approach is semantically defined by a model-based merging operator in propositional logic and the merged result of our approach is a normal possibility distribution. We also give an algorithm to obtain the syntactical counterpart of the semantic approach. The logical properties of our approach are considered. Finally, we analyze the computational complexity of our merging approach.

JELIA Conference 2006 Conference Paper

Knowledge Base Revision in Description Logics

• Guilin Qi
• Weiru Liu
• David A. Bell

Abstract Ontology evolution is an important problem in the Semantic Web research. Recently, Alchourrón, Gärdenfors and Markinson’s (AGM) theory on belief change has been applied to deal with this problem. However, most of current work only focuses on the feasibility of the application of AGM postulates on contraction to description logics (DLs), a family of ontology languages. So the explicit construction of a revision operator is ignored. In this paper, we first generalize the AGM postulates on revision to DLs. We then define two revision operators in DLs. One is the weakening-based revision operator which is defined by weakening of statements in a DL knowledge base and the other is its refinement. We show that both operators capture some notions of minimal change and satisfy the generalized AGM postulates for revision.

UAI Conference 2005 Conference Paper

A Revision-Based Approach to Resolving Conflicting Information

• Guilin Qi
• Weiru Liu
• David A. Bell

In this paper, we propose a revision-based approach for conflict resolution by generalizing the Disjunctive Maxi-Adjustment (DMA) approach (Benferhat et al. 2004). Revision operators can be classified into two different families: the model-based ones and the formula-based ones. So the revision-based approach has two different versions according to which family of revision operators is chosen. Two particular revision operators are considered, one is the Dalal's revision operator, which is a model-based revision operator, and the other is the cardinality-maximal based revision operator, which is a formulabased revision operator. When the Dalal's revision operator is chosen, the revision-based approach is independent of the syntactic form in each stratum and it captures some notion of minimal change. When the cardinalitymaximal based revision operator is chosen, the revision-based approach is equivalent to the DMA approach. We also show that both approaches are computationally easier than the DMA approach.

IJCAI Conference 2005 Conference Paper

Measuring conflict and agreement between two prioritized belief bases

• Guilin Qi
• Weiru Liu
• David A

In this paper, we investigate the relationship between two prioritized knowledge bases by measuring both the conflict and the agreement between them. First of all, a quantity of conflict and two quantities of agreement are defined. The former is shown to be a generalization of the Dalal distance. The latter are, respectively, a quantity of strong agreement which measures the amount of information on which two belief bases “totally” agree, and a quantity of weak agreement which measures the amount of information that is believed by one source but is unknown to the other. All three quantity measures are based on the weighted prime implicant, which represents beliefs in a prioritized belief base. We then define a degree of conflict and two degrees of agreement based on our quantity of conflict and the quantities of agreement. We also consider the impact of these measures on belief merging and information source ordering.

NMR Workshop 2004 Conference Paper

Combining individually inconsistent prioritized knowledge bases

• Guilin Qi
• Weiru Liu
• David H. Glass

It is well accepted that inconsistency may exist in a database system or an intelligent information system (Benferhat et al. 1993a; 1993b; 1997b; 1998; Benferhat & Kaci 2003; Elvang-Gøransson & Hunter 1995; Gabbay & Hunter 1991; Lin 1994; Priest et al. 1989; Priest 2001). Inconsistency can either appear in the given knowledge bases or as a result of combination or revision. In this paper, we will propose two different methods to combine individually inconsistent possibilistic knowledge bases. The first method, called an argument-based method, is a generalization of the merging method introduced in (Benferhat & Kaci 2003). When the knowledge bases to be merged are self-consistent, this method coincides with the original one. The second method, called a multiple-operator based method, combines the consistent and the conflict information using different operators. This method is more reasonable than the argument-based method because it differentiates the consistent and conflict information.