Recently, Omni-modal large language models (OLLMs) have sparked a new wave of research, achieving impressive results in tasks such as audio-video understanding and real-time environment perception. However, hallucination issues still persist. Similar to the bimodal setting, the priors from the text modality tend to dominate, leading OLLMs to rely more heavily on textual cues while neglecting visual and audio information. In addition, fully multimodal scenarios introduce new challenges. Most existing models align visual or auditory modalities with text independently during training, while ignoring the intrinsic correlations between video and its corresponding audio. This oversight results in hallucinations when reasoning requires interpreting hidden audio cues embedded in video content. To address these challenges, we propose OmniDPO, a preference-alignment framework designed to mitigate hallucinations in OLLMs. Specifically, OmniDPO incorporates two strategies: (1) constructing text-preference sample pairs to enhance the model’s understanding of audio-video interactions; and (2) constructing multimodal-preference sample pairs to strengthen the model’s attention to visual and auditory information. By tackling both challenges, OmniDPO effectively improves multimodal grounding and reduces hallucination. Experiments conducted on two OLLMs demonstrate that OmniDPO not only effectively mitigates multimodal hallucinations but also significantly enhances the models' reasoning capabilities across modalities.

Leyi Pan, Shuchang Tao, Yunpeng Zhai, Zheyu Fu, Liancheng Fang, Minghua He, Lingzhe Zhang, Zhaoyang Liu, Bolin Ding, Aiwei Liu, Lijie Wen. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

While person re-identification (Re-ID) has achieved remarkable progress in controlled laboratory settings, its widespread deployment in real-world urban surveillance remains impeded by the significant domain gaps caused by environmental dynamics and sparse camera topologies. In such operational scenarios, relying solely on visual appearance leads to severe visual ambiguity, while the continuous expansion of camera networks induces catastrophic forgetting. To bridge this “lab-to-real” gap, we present a robust, street-surveillance-oriented Re-ID solution validated across a diverse array of benchmarks. Our framework incorporates three key innovations: (1) We construct a deployable Teacher–Student framework to ensure stable feature transfer from labeled source domains to noisy target environments, enabling robust adaptation without manual supervision; (2) Addressing system sustainability, we design a diversity-preserving Dynamic Data Replay mechanism based on Farthest Point Sampling (FPS) to prevent catastrophic forgetting as the network continuously expands; (3) A Spatiotemporal Feature Fusion (STFF) module is developed to resolve visual ambiguity in sparse networks by imposing explicit physical constraints to filter out spatiotemporally infeasible candidates. Extensive evaluations on a proprietary benchmark derived from the live Hainan surveillance system and a series of mainstream datasets demonstrate that our method significantly outperforms existing approaches, offering a superior solution for practical deployment in smart city infrastructure.

Adversarial attacks pose a significant threat to machine learning models by inducing incorrect predictions through imperceptible perturbations to input data. While these attacks are well studied in unstructured domains such as images, their behaviour on tabular data remains underexplored due to mixed feature types and complex inter-feature dependencies. This study introduces a comprehensive benchmark that evaluates adversarial attacks on tabular datasets with respect to both effectiveness and imperceptibility. We assess five white-box attack algorithms (FGSM, BIM, PGD, DeepFool, and C&W) across four representative models (LR, MLP, TabTransformer and FT-Transformer) using eleven datasets spanning finance, energy, and healthcare domains. The benchmark employs four quantitative imperceptibility metrics (proximity, sparsity, deviation, and sensitivity) to characterise perturbation realism. The analysis quantifies the trade-off between these two aspects and reveals consistent differences between attack types, with ℓ ∞ -based attacks achieving higher success but lower subtlety, and ℓ 2 -based attacks offering more realistic perturbations. The benchmark findings offer actionable insights for designing more imperceptible adversarial attacks, advancing the understanding of adversarial vulnerability in tabular machine learning.

Leyi Pan, Aiwei Liu, Shiyu Huang, Yijian Lu, Xuming Hu, Lijie Wen, Irwin King, Philip S. Yu. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

The literature review is an indispensable step in the research process. It provides the benefit of comprehending the research problem and understanding the current research situation while conducting a comparative analysis of prior works. However, literature summary is challenging and time consuming. The previous LLM-based studies on literature review mainly focused on the complete process, including literature retrieval, screening, and summarization. However, for the summarization step, simple CoT method often lacks the ability to provide extensive comparative summary. In this work, we firstly focus on the independent literature summarization step and introduce ChatCite, an LLM agent with human workflow guidance for comparative literature summary. This agent, by mimicking the human workflow, first extracts key elements from relevant literature and then generates summaries using a Reflective Incremental Mechanism. In order to better evaluate the quality of the generated summaries, we devised a LLM-based automatic evaluation metric, G-Score, in refer to the human evaluation criteria. The ChatCite agent outperformed other models in various dimensions in the experiments. The literature summaries generated by ChatCite can also be directly used for drafting literature reviews.

Image Translation (IT) holds immense potential across diverse domains, enabling the translation of textual content within images into various languages. However, existing datasets often suffer from limitations in scale, diversity, and quality, hindering the development and evaluation of IT models. To address this issue, we introduce MIT-10M, a large-scale parallel corpus of multilingual image translation with over 10M image-text pairs derived from real-world data, which has undergone extensive data cleaning and multilingual translation validation. It contains 0.8M images in three sizes, 28 categories, tasks with three levels of difficulty and 14 languages image-text pairs, which is a considerable improvement on existing datasets. We conduct extensive experiments to evaluate and train models on MIT-10M. The experimental results clearly indicate that our dataset has higher adaptability when it comes to evaluating the performance of the models in tackling challenging and complex image translation tasks in the real world. Moreover, the performance of the model fine-tuned with MIT-10M has tripled compared to the baseline model, further confirming its superiority.

Despite the recent breakthroughs achieved by Large Vision Language Models (LVLMs) in understanding and responding to complex visual-textual contexts, their inherent hallucination tendencies limit their practical application in real-world scenarios that demand high levels of precision. Existing methods typically either fine-tune the LVLMs using additional data, which incurs extra costs in manual annotation and computational resources or perform comparisons at the decoding stage, which may eliminate useful language priors for reasoning while introducing inference time overhead. Therefore, we propose ICT, a lightweight, training-free method that calculates an intervention direction to shift the model’s focus towards different levels of visual information, enhancing its attention to high-level and fine-grained visual details. During the forward pass stage, the intervention is applied to the attention heads that encode the overall image information and the fine-grained object details, effectively mitigating the phenomenon of overly language priors, and thereby alleviating hallucinations. Extensive experiments demonstrate that ICT achieves strong performance with a small amount of data and generalizes well across different datasets and models. Our codes are publicly available at:https://github.com/THU-BPM/ICT/.

Key-value stores face challenges from skewed dynamic workloads. We presents CableCache, an in-network request deduplication system. It utilizes programmable switches to maintain an object request information directory and employs the INRD protocol to deduplicate requests for hot objects with the same key, which not only overcomes the limitation of target object value size but also shortens the access path to some extent. Additionally, it can address complex scenarios such as the presence of read-modify-write requests and multi-packet target objects. Experiments show that the system effectively alleviates load imbalance in key-value stores, improves system-wide throughput, and reduces tail latency of requests under both synthetic and real workloads.

Direct Preference Optimization (DPO) has been widely adopted for preference alignment of Large Language Models (LLMs) due to its simplicity and effectiveness. However, DPO is derived as a bandit problem in which the whole response is treated as a single arm, ignoring the importance differences between tokens, which may affect optimization efficiency and make it difficult to achieve optimal results. In this work, we propose that the optimal data for DPO has equal expected rewards for each token in winning and losing responses, as there is no difference in token importance. However, since the optimal dataset is unavailable in practice, we propose using the original dataset for importance sampling to achieve unbiased optimization. Accordingly, we propose a token-level importance sampling DPO objective named TIS-DPO that assigns importance weights to each token based on its reward. Inspired by previous works, we estimate the token importance weights using the difference in prediction probabilities from a pair of contrastive LLMs. We explore three methods to construct these contrastive LLMs: (1) guiding the original LLM with contrastive prompts, (2) training two separate LLMs using winning and losing responses, and (3) performing forward and reverse DPO training with winning and losing responses. Experiments show that TIS-DPO significantly outperforms various baseline methods on harmlessness and helpfulness alignment and summarization tasks. We also visualize the estimated weights, demonstrating their ability to identify key token positions.

Text watermarking for Large Language Models (LLMs) has made significant progress in detecting LLM outputs and preventing misuse. Current watermarking techniques offer high detectability, minimal impact on text quality, and robustness to text editing. However, current researches lack investigation into the imperceptibility of watermarking techniques in LLM services. This is crucial as LLM providers may not want to disclose the presence of watermarks in real-world scenarios, as it could reduce user willingness to use the service and make watermarks more vulnerable to attacks. This work is the first to investigate the imperceptibility of watermarked LLMs. We design an identification algorithm called Water-Probe that detects watermarks through well-designed prompts to the LLM. Our key motivation is that current watermarked LLMs expose consistent biases under the same watermark key, resulting in similar differences across prompts under different watermark keys. Experiments show that almost all mainstream watermarking algorithms are easily identified with our well-designed prompts, while Water-Probe demonstrates a minimal false positive rate for non-watermarked LLMs. Finally, we propose that the key to enhancing the imperceptibility of watermarked LLMs is to increase the randomness of watermark key selection. Based on this, we introduce the Water-Bag strategy, which significantly improves watermark imperceptibility by merging multiple watermark keys.

In biomedical fields, one named entity may consist of a series of non-adjacent tokens and overlap with other entities. Previous methods recognize discontinuous entities by connecting entity fragments or internal tokens, which face challenges of error propagation and decoding ambiguity due to the wide variety of span or word combinations. To address these issues, we deeply explore discontinuous entity structures and propose an effective Gap-aware grid tagging model for Discontinuous Named Entity Recognition, named GapDNER. Our GapDNER innovatively applies representation learning on the context gaps between entity fragments to resolve decoding ambiguity and enhance discontinuous NER performance. Specifically, we treat the context gap as an additional type of span and convert span classification into a token-pair grid tagging task. Subsequently, we design two interactive components to comprehensively model token-pair grid features from both intra- and inter-span perspectives. The intra-span regularity extraction module employs the biaffine mechanism along with linear attention to capture the internal regularity of each span, while the inter-span relation enhancement module utilizes criss-cross attention to obtain semantic relations among different spans. At the inference stage of entity decoding, we assign a directed edge to each entity fragment and context gap, then use the BFS algorithm to search for all valid paths from the head to tail of grids with entity tags. Experimental results on three datasets demonstrate that our GapDNER achieves new state-of-the-art performance on discontinuous NER and exhibits remarkable advantages in recognizing complex entity structures.

Traditional named entity recognition (NER) aims to identify text mentions into pre-defined entity types. Continual Named Entity Recognition (CNER) is introduced since entity categories are continuously increasing in various real-world scenarios. However, existing continual learning (CL) methods for NER face challenges of catastrophic forgetting and semantic shift of non-entity type. In this paper, we propose GenCNER, a simple but effective Generative framework for CNER to mitigate the above drawbacks. Specifically, we skillfully convert the CNER task into sustained entity triplet sequence generation problem and utilize a powerful pre-trained seq2seq model to solve it. Additionally, we design a type-specific confidence-based pseudo labeling strategy along with knowledge distillation (KD) to preserve learned knowledge and alleviate the impact of label noise at the triplet level. Experimental results on two benchmark datasets show that our framework outperforms previous state-of-the-art methods in multiple CNER settings, and achieves the smallest gap compared with non-CL results.

Recent studies exploited Large Language Models (LLMs) to autonomously generate heuristics for solving Combinatorial Optimization Problems (COPs), by prompting LLMs to first provide search directions and then derive heuristics accordingly. However, the absence of task-specific knowledge in prompts often leads LLMs to provide unspecific search directions, obstructing the derivation of well-performing heuristics. Moreover, evaluating the derived heuristics remains resource-intensive, especially for those semantically equivalent ones, often requiring omissible resource expenditure. To enable LLMs to provide specific search directions, we propose the Hercules algorithm, which leverages our designed Core Abstraction Prompting (CAP) method to abstract the core components from elite heuristics and incorporate them as prior knowledge in prompts. We theoretically prove the effectiveness of CAP in reducing unspecificity and provide empirical results in this work. To reduce computing resources required for evaluating the derived heuristics, we propose few-shot Performance Prediction Prompting (PPP), a first-of-its-kind method for the Heuristic Generation (HG) task. PPP leverages LLMs to predict the fitness values of newly derived heuristics by analyzing their semantic similarity to previously evaluated ones. We further develop two tailored mechanisms for PPP to enhance predictive accuracy and determine unreliable predictions, respectively. The use of PPP makes Hercules more resource-efficient and we name this variant Hercules-P. Extensive experiments across four HG tasks, five COPs, and eight LLMs demonstrate that Hercules outperforms the state-of-the-art LLM-based HG algorithms, while Hercules-P excels at minimizing required computing resources. In addition, we illustrate the effectiveness of CAP, PPP, and the other proposed mechanisms by conducting relevant ablation studies.

Process mining techniques can extract process models from event logs produced by information systems. However, in flexible environments, simply using existing methods often leads to complex process models that are hard to understand, due to the less structured and greater complexity of processes in real life. Trace clustering is a pre-processing technique that enhances the effectiveness of model mining by partitioning similar behaviors in logs. In this paper, we present a Multi-view Trace Clustering method, named MTC, that improves the homogeneity of trace subclusters. Our method consists of three parts: (1) We use trace profiles to depict the traces from different views, and each profile would be transformed into a graph based on the k-nearest neighbor algorithm; (2) A fusion graph is designed to capture the information among these graphs based on an attention coefficient matrix, and then the graph convolutional networks are used to encode all graphs for obtaining the common representation; (3) We also enhance the characterization of the common representation with an inner decoder. Finally, we adopt k-means to cluster the traces in the log based on the common representation. Extensive experiments using multiple datasets illustrate that MTC significantly surpasses state-of-the-art trace clustering methods

Junzhe Chen, Xuming Hu, Shuodi Liu, Shiyu Huang, Wei-Wei Tu, Zhaofeng He, Lijie Wen. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Aiwei Liu, Haoping Bai, Zhiyun Lu, Xiang Kong, Xiaoming Wang, Jiulong Shan, Meng Cao, Lijie Wen. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Large language models (LLMs) have recently driven striking performance improvements across a range of natural language processing tasks. The factual knowledge acquired during pretraining and instruction tuning can be useful in various downstream tasks, such as question answering, and language generation. Unlike conventional Knowledge Bases (KBs) that explicitly store factual knowledge, LLMs implicitly store facts in their parameters. Content generated by the LLMs can often exhibit inaccuracies or deviations from the truth, due to facts that can be incorrectly induced or become obsolete over time. To this end, we aim to explore the extent and scope of factual knowledge within LLMs by designing the benchmark Pinocchio. Pinocchio contains 20K diverse factual questions that span different sources, timelines, domains, regions, and languages. Furthermore, we investigate whether LLMs can compose multiple facts, update factual knowledge temporally, reason over multiple pieces of facts, identify subtle factual differences, and resist adversarial examples. Extensive experiments on different sizes and types of LLMs show that existing LLMs still lack factual knowledge and suffer from various spurious correlations. We believe this is a critical bottleneck for realizing trustworthy artificial intelligence. The dataset Pinocchio and our codes are publicly available at: https://github.com/THU-BPM/Pinocchio.

Recently, text watermarking algorithms for large language models (LLMs) have been proposed to mitigate the potential harms of text generated by LLMs, including fake news and copyright issues. However, current watermark detection algorithms require the secret key used in the watermark generation process, making them susceptible to security breaches and counterfeiting during public detection. To address this limitation, we propose an unforgeable publicly verifiable watermark algorithm named UPV that uses two different neural networks for watermark generation and detection, instead of using the same key at both stages. Meanwhile, the token embedding parameters are shared between the generation and detection networks, which makes the detection network achieve a high accuracy very efficiently. Experiments demonstrate that our algorithm attains high detection accuracy and computational efficiency through neural networks. Subsequent analysis confirms the high complexity involved in forging the watermark from the detection network. Our code is available at https://github.com/THU-BPM/unforgeable_watermark

Watermark algorithms for large language models (LLMs) have achieved extremely high accuracy in detecting text generated by LLMs. Such algorithms typically involve adding extra watermark logits to the LLM's logits at each generation step. However, prior algorithms face a trade-off between attack robustness and security robustness. This is because the watermark logits for a token are determined by a certain number of preceding tokens; a small number leads to low security robustness, while a large number results in insufficient attack robustness. In this work, we propose a semantic invariant watermarking method for LLMs that provides both attack robustness and security robustness. The watermark logits in our work are determined by the semantics of all preceding tokens. Specifically, we utilize another embedding LLM to generate semantic embeddings for all preceding tokens, and then these semantic embeddings are transformed into the watermark logits through our trained watermark model. Subsequent analyses and experiments demonstrated the attack robustness of our method in semantically invariant settings: synonym substitution and text paraphrasing settings. Finally, we also show that our watermark possesses adequate security robustness.

We introduce a novel method tailored for unconstrained multiview optical satellite photogrammetry in time-varying illumination and reflection conditions. Our approach uses continuous radiance fields to represent surface radiance and albedo based on radiometry principles, integrating both static and transient components for satellite photogrammetry. In addition, an innovative self-supervised mechanism is introduced to optimize the learning process which leverages dark regions’ accentuation, transient and static composition, and shadow regularization. Evaluations on multidate WorldView-3 images affirm that our model consistently surpasses the state-of-the-art techniques.

Open relation extraction is the task of extracting open-domain relation facts from natural language sentences. Existing works either utilize distant-supervised annotations to train a supervised classifier over pre-defined relations, or adopt unsupervised methods with additional dependency on external assumptions. However, these works can only obtain information signals from limited existing knowledge bases or datasets. In this work, we propose a self-supervised framework namedWeb-SelfORE, which exploits self-supervised signals by requiring a large pretrained language model to extensively read real-world relevant documents from the web, and obtain contextualized relational features by mixing contextualized representations of entities from different documents. We perform adaptive clustering on contextualized relational features and bootstrap the self-supervised signals by improving contextualized features in relation classification. We additionally compare the effectiveness of self-supervisions brought by different document sources, and introduce relevance and redundancy evaluation metrics to obtain higher-quality self-supervisions. Experimental results on four public datasets show the effectiveness and robustness ofWeb-SelfOREon open-domain relation extraction task when comparing with competitive baselines.

Process discovery aims to extract descriptive process models from event logs. To date, various process discovery algorithms have been proposed for different application settings. However, most of them meet challenges in handling event logs produced from hierarchical multi-instance business processes, in which multiple sub-process instances are invoked by the execution of a parent process. To address the problem, a novel approach is presented to support the discovery of hierarchical multi-instance process models. Specifically, taking event logs with multi-instance information as input, the detailed implementation of our method can be generally divided into four steps: nesting relation detection, hierarchical event log construction, sub-process case identification, and hierarchical multi-instance model discovery. We have implemented our approach properly as plugins in the openly accessible ProM toolkit, and compared its performance against the state-of-the-art process discovery approaches over six publicly available event logs. Based on the experimental result, it is demonstrated that the proposed approach can effectively discover hierarchical multi-instance process models with better quality.

Aspect-based sentiment analysis (ABSA) is a fine-grained sentiment classification task.Many recent works have used dependency trees to extract the relation between aspects and contexts and have achieved significant improvements.However, further improvement is limited due to the potential mismatch between the dependency tree as a syntactic structure and the sentiment classification as a semantic task.To alleviate this gap, we replace the syntactic dependency tree with the semantic structure named Abstract Meaning Representation (AMR) and propose a model called AMR-based Path Aggregation Relational Network (APARN) to take full advantage of semantic structures.In particular, we design the path aggregator and the relation-enhanced selfattention mechanism that complement each other.The path aggregator extracts semantic features from AMRs under the guidance of sentence information, while the relationenhanced self-attention mechanism in turn improves sentence features with refined semantic information.Experimental results on four public datasets demonstrate 1.13% average F1 improvement of APARN in ABSA when compared with state-of-the-art baselines.1

How to identify semantic relations among entities in a document when only a few labeled documents are available?Few-shot documentlevel relation extraction (FSDLRE) is crucial for addressing the pervasive data scarcity problem in real-world scenarios.Metric-based meta-learning is an effective framework widely adopted for FSDLRE, which constructs class prototypes for classification.However, existing works often struggle to obtain class prototypes with accurate relational semantics: 1) To build prototype for a target relation type, they aggregate the representations of all entity pairs holding that relation, while these entity pairs may also hold other relations, thus disturbing the prototype.2) They use a set of generic NOTA (none-of-the-above) prototypes across all tasks, neglecting that the NOTA semantics differs in tasks with different target relation types.In this paper, we propose a relation-aware prototype learning method for FSDLRE to strengthen the relational semantics of prototype representations.By judiciously leveraging the relation descriptions and realistic NOTA instances as guidance, our method effectively refines the relation prototypes and generates task-specific NOTA prototypes.Extensive experiments demonstrate that our method outperforms state-of-the-art approaches by average 2.61% F 1 across various settings of two FSDLRE benchmarks. 1

Named Entity Recognition (NER) is a well and widely studied task in natural language processing. Recently, the nested NER has attracted more attention since its practicality and difficulty. Existing works for nested NER ignore the recognition order and boundary position relation of nested entities. To address these issues, we propose a novel seq2seq model named GPRL, which formulates the nested NER task as an entity triplet sequence generation process. GPRL adopts the reinforcement learning method to generate entity triplets de-coupling the entity order in gold labels and expects to learn a reasonable recognition order of entities via trial and error. Based on statistics of boundary distance for nested entities, GPRL designs a Gaussian prior to represent the boundary distance distribution between nested entities and adjust the out-put probability distribution of nested boundary tokens. Experiments on three nested NER datasets demonstrate that GPRL outperforms previous nested NER models.

How can we better extract entities and relations from text? Using multimodal extraction with images and text obtains more signals for entities and relations, and aligns them through graphs or hierarchical fusion, aiding in extraction. Despite attempts at various fusions, previous works have overlooked many unlabeled image-caption pairs, such as NewsCLIPing. This paper proposes innovative pre-training objectives for entity-object and relation-image alignment, extracting objects from images and aligning them with entity and relation prompts for soft pseudo-labels. These labels are used as self-supervised signals for pre-training, enhancing the ability to extract entities and relations. Experiments on three datasets show an average 3.41% F1 improvement over prior SOTA. Additionally, our method is orthogonal to previous multimodal fusions, and using it on prior SOTA fusions further improves 5.47% F1.

Low-resource relation extraction (LRE) aims to extract potential relations from limited labeled corpus to handle the problem of scarcity of human annotations. Previous works mainly consist of two categories of methods: (1) Self-training methods, which improve themselves through the models' predictions, thus suffering from confirmation bias when the predictions are wrong. (2) Self-ensembling methods, which learn task-agnostic representations, therefore, generally do not work well for specific tasks. In our work, we propose a novel LRE architecture named SelfLRE, which leverages two complementary modules, one module uses self-training to obtain pseudo-labels for unlabeled data, and the other module uses self-ensembling learning to obtain the task-agnostic representations, and leverages the existing pseudo-labels to refine the better task-specific representations on unlabeled data. The two models are jointly trained through multi-task learning to iteratively improve the effect of LRE task. Experiments on three public datasets show that SelfLRE achieves 1.81% performance gain over the SOTA baseline. Source code is available at: https://github.com/THU-BPM/SelfLRE.

As social media platforms are evolving from text-based forums into multi-modal environments, the nature of misinformation in social media is also transforming accordingly. Misinformation spreaders have recently targeted contextual connections between the modalities e.g., text and image. However, existing datasets for rumor detection mainly focus on a single modality i.e., text. To bridge this gap, we construct MR2, a multimodal multilingual retrieval-augmented dataset for rumor detection. The dataset covers rumors with images and texts, and provides evidence from both modalities that are retrieved from the Internet. Further, we develop established baselines and conduct a detailed analysis of the systems evaluated on the dataset. Extensive experiments show that MR2 will provide a challenging testbed for developing rumor detection systems designed to retrieve and reason over social media posts. Source code and data are available at: https://github.com/THU-BPM/MR2.

Real-world fact verification task aims to verify the factuality of a claim by retrieving evidence from the source document. The quality of the retrieved evidence plays an important role in claim verification. Ideally, the retrieved evidence should be faithful (reflecting the model's decision-making process in claim verification) and plausible (convincing to humans), and can improve the accuracy of verification task. Although existing approaches leverage the similarity measure of semantic or surface form between claims and documents to retrieve evidence, they all rely on certain heuristics that prevent them from satisfying all three requirements. In light of this, we propose a fact verification model named ReRead to retrieve evidence and verify claim that: (1) Train the evidence retriever to obtain interpretable evidence (i.e., faithfulness and plausibility criteria); (2) Train the claim verifier to revisit the evidence retrieved by the optimized evidence retriever to improve the accuracy. The proposed system is able to achieve significant improvements upon best-reported models under different settings.

Natural Language Inference (NLI) is a growingly essential task in natural language understanding, which requires inferring the relationship between the sentence pairs (premiseandhypothesis). Recently, low-resource natural language inference has gained increasing attention, due to significant savings in manual annotation costs and a better fit with real-world scenarios. Existing works fail to characterize discriminative representations between different classes with limited training data, which may cause faults in label prediction. Here we propose a multi-level supervised contrastive learning framework named MultiSCL for low-resource natural language inference. MultiSCL leverages a sentence-level and pair-level contrastive learning objective to discriminate between different classes of sentence pairs by bringing those in one class together and pushing away those in different classes. MultiSCL adopts a data augmentation module that generates different views for input samples to better learn the latent representation. The pair-level representation is obtained from a cross attention module. We conduct extensive experiments on two public NLI datasets in low-resource settings, and the accuracy of MultiSCL exceeds other models by 1.8%, 3.1% and 4.1% on SNLI, MNLI and Sick with 5 instances per label respectively. Moreover, our method outperforms the previous state-of-the-art method on cross-domain tasks of text classification.

With the growth of the depth of neural networks and the scale of data, the difficulty of network training also increases. When the GPU memory is insufficient, it is challenging to train deeper models. Recent research uses tensor swapping and recomputation techniques in a combined manner to optimize memory usage. However, complex dependencies and enormous scales of the DNN graph limit the improvement of single GPU memory optimization. Improper swap and recomputation decisions even bring negative effects on training performance. In this article, we propose a novel hybrid tensor re-generation strategy, called STR, which combines swap and recomputation techniques to find the optimal execution plan for the DNN training when the memory is limited. We formalize our memory optimization problem with constraints that describe the dependency of the operator calculation and the bandwidth usage of the swap. Ahost checkpointmechanism is designed to make full use of the swapped tensors, which reduces the cost of the recomputation. We also present arecursive source tracingalgorithm to improve the optimization efficiency by constraint relaxation with a performance bound. To optimize large models, we further introduce an approximation method based on a weighted graph coarsening. We implement a prototype of STR as a plugin on TensorFlow and evaluated based on 5 popular DNN models. The experimental result shows that the approximate solution of STR improves the training throughput of ResNet series of models by up to 28.1% compared to the state-of-the-art hybrid optimization strategy.

Business processes are critical for information systems to control workflows and deliver services. Although existing process discovery techniques can generate flat process models from business event logs, few of them have investigated the notion of hierarchy (i.e., subprocesses) yet. To fill the gap, this article first defines the concept of hierarchical Petri nets (HPNs), which can support the formal modeling and correctness verification of processes with subprocesses. Followed by that, we propose an approach which can effectively discover HPNs from event logs with lifecycle information. Moreover, to quantify the quality of discovered HPNs, details on how to transform an HPN to a classical Petri net are given such that existing metrics can be applied. All proposed approaches have been fully implemented in ProM, and experiments over both synthetic and real-life event logs demonstrate that our approach can effectively discover hierarchical process models. Specifically, compared to exiting approaches on processes discovery, our approach can generally perform better in terms of model quality.

A scene graph is a semantic representation that expresses the objects, attributes, and relationships between objects in a scene. Scene graphs play an important role in many cross modality tasks, as they are able to capture the interactions between images and texts. In this paper, we focus on scene graph modification (SGM), where the system is required to learn how to update an existing scene graph based on a natural language query. Unlike previous approaches that rebuilt the entire scene graph, we frame SGM as a graph expansion task by introducing the incremental structure expanding (ISE). ISE constructs the target graph by incrementally expanding the source graph without changing the unmodified structure. Based on ISE, we further propose a model that iterates between nodes prediction and edges prediction, inferring more accurate and harmonious expansion decisions progressively. In addition, we construct a challenging dataset that contains more complicated queries and larger scene graphs than existing datasets. Experiments on four benchmarks demonstrate the effectiveness of our approach, which surpasses the previous state-of-the-art model by large margins.

We propose the first character-level white-box adversarial attack method against transformer models.The intuition of our method comes from the observation that words are split into subtokens before being fed into the transformer models and the substitution between two close subtokens has a similar effect to the character modification.Our method mainly contains three steps.First, a gradient-based method is adopted to find the most vulnerable words in the sentence.Then we split the selected words into subtokens to replace the origin tokenization result from the transformer tokenizer.Finally, we utilize an adversarial loss to guide the substitution of attachable subtokens in which the Gumbel-softmax trick is introduced to ensure gradient propagation.Meanwhile, we introduce the visual and length constraint in the optimization process to achieve minimum character modifications.Extensive experiments on both sentence-level and token-level tasks demonstrate that our method could outperform the previous attack methods in terms of success rate and edit distance.Furthermore, human evaluation verifies our adversarial examples could preserve their origin labels.

Natural language inference (NLI) is an increasingly important task for natural language understanding, which requires one to infer the relationship between the sentence pair (premise and hypothesis). Many recent works have used contrastive learning by incorporating the relationship of the sentence pair from NLI datasets to learn sentence representation. However, these methods only focus on comparisons with sentence-level representations. In this paper, we propose a Pair-level Supervised Contrastive Learning approach (PairSCL). We adopt a cross attention module to learn the joint representations of the sentence pairs. A contrastive learning objective is designed to distinguish the varied classes of sentence pairs by pulling those in one class together and pushing apart the pairs in other classes. We evaluate PairSCL on two public datasets of NLI where the accuracy of PairSCL outperforms other methods by 2.1% on average. Furthermore, our method outperforms the previous state-of-the-art method on seven transfer tasks of text classification.1

With the growing of the depth of neural networks and the scale of data, the difficulty of network training also increases. When the GPU memory is insufficient, it is challenging to train deeper models. Recent research uses tensor swapping and recomputation techniques in a combined manner to optimize the memory usage. However, complex dependencies of the DNN graph limit the improvement of the single GPU memory optimization. Improper swap decisions even brings negative effects because the source of the recomputation may have been swapped out. In this paper, we propose a novel swap dominated tensor re-generation strategy, called STR, which combines swap and recomputation techniques to find the optimal execution plan for the DNN training when the memory is limited. We formalize our memory optimization problem with constraints which describe the dependency of the operator calculation and the bandwidth usage of swap. A host checkpoint mechanism is designed to make full use of the swapped tensors, which reduces the cost of the recomputation. We also present an approximation method based on a recursive source tracing procedure to improve the optimization efficiency. We implement a prototype of STR as a plugin on TensorFlow. The experimental result shows that STR improves up to 21.3% throughput compared with the state-of-the-art hybrid optimization strategy.

The generalizability to new databases is of vital importance to Text-to-SQL systems which aim to parse human utterances into SQL statements. Existing works achieve this goal by leveraging the exact matching method to identify the lexical matching between the question words and the schema items. However, these methods fail in other challenging scenarios, such as the synonym substitution in which the surface form differs between the corresponding question words and schema items. In this paper, we propose a framework named ISESL-SQL to iteratively build a semantic enhanced schema-linking graph between question tokens and database schemas. First, we extract a schema linking graph from PLMs through a probing procedure in an unsupervised manner. Then the schema linking graph is further optimized during the training process through a deep graph learning method. Meanwhile, we also design an auxiliary task called graph regularization to improve the schema information mentioned in the schema-linking graph. Extensive experiments on three benchmarks demonstrate that ISESL-SQL could consistently outperform the baselines and further investigations show its generalizability and robustness.

Xuming Hu, Zhijiang Guo, GuanYu Wu, Aiwei Liu, Lijie Wen, Philip Yu. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Shuliang Liu, Xuming Hu, Chenwei Zhang, Shu’ang Li, Lijie Wen, Philip Yu. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Prediction over event sequences is critical for many real-world applications in Information Retrieval and Natural Language Processing. Future Event Generation (FEG) is a challenging task in event sequence prediction because it requires not only fluent text generation but also commonsense reasoning to maintain the logical coherence of the entire event story. In this paper, we propose a novel explainable FEG framework, Coep. It highlights and integrates two types of event knowledge, sequential knowledge of direct event-event relations and inferential knowledge that reflects the intermediate character psychology between events, such as intents, causes, reactions, which intrinsically pushes the story forward. To alleviate the knowledge forgetting issue, we design two modules, IM and GM, for each type of knowledge, which are combined via prompt tuning. First, IM focuses on understanding inferential knowledge to generate commonsense explanations and provide a soft prompt vector for GM. We also design a contrastive discriminator for better generalization ability. Second, GM generates future events by modeling direct sequential knowledge with the guidance of IM. Automatic and human evaluation demonstrate that our approach can generate more coherent, specific, and logical future events.

Flexibility and evolution have been a hot topic in the context of business process management. However, contemporary process mining techniques assume processes to be in a steady state, which will result in a mixed-model mined from the event log. Business process drift detection is a family of methods to detect changes by analyzing the event log, but existing methods have some disadvantages in dealing with concept drifts. First, most of these methods detect changes depending on an exploration of a potentially large feature space and are time consuming. Second, with the size of delay period becoming small, the accuracies of some methods will drop rapidly. In this article, we address these problems by proposing a novel drift detection technique called LCDD that significantly differs from existing methods. The core idea is to find out new features and disappeared features from traces after the log fragment meets local completeness. To begin with, we use the relations of direct succession as the lightweight feature, which are compared between two windows (complete window and detection window). Then, we find new direct successions and stable disappeared direct successions by just moving detection window. Finally, the forgetting mechanism is used to abandon some direct successions after finding a change point. An extensive empirical evaluation shows that LCDD is faster and more accurate.

Measuring business process similarity plays an important role in the analysis, management and optimization of business in big companies. In the early days, experts paid major attention to calculating business similarity according to corresponding process models. However, models only express ideal behavior of business processes without any undesired or unexpected business routines. In order to fully model business behavior, some researchers use system logs in similarity measuring. But previous system-logs-based similarity measurements have limitations on: (1) satisfaction of algorithm properties, (2) distribution of similarity values, and (3) complexity of algorithm. In this article, we take the advantages of token logs in process behavioral similarity measuring. Firstly, the Token Log Profile, modeled with a relation matrix, is defined as an abstraction of the initial token logs. Then, similarity between business processes is calculated based on their Token Log Profiles according to the proposed algorithm. Besides, we extend the properties that similarity algorithms should satisfy for evaluating the proposed algorithm. The experimental and analytical results show that our algorithm achieves very promising accuracy and efficiency while satisfying all the proposed properties compared with state-of-the-art algorithms.

Many activities in business process management, such as process retrieval, process mining, and process integration, need to determine the similarity between business processes. Along with many other relational behavior semantics, Transition Adjacency Relation (abbr. TAR) has been proposed as a kind of behavioral gene of process models and a useful perspective for process similarity measurement. In this article we explain why it is still relevant and necessary to improve TAR or pTAR (i.e., projected TAR) computation efficiency and put forward a novel approach for TAR computation based on Petri net unfolding. This approach not only improves the efficiency of TAR computation, but also enables the long-expected combined usage of TAR and Behavior Profiles (abbr. BP) in process model similarity estimation.

Distributed in-memory computing frameworks usually have lots of parameters (e.g., the buffer size of shuffle) to form a configuration for each execution. A well-tuned configuration can bring large improvements of performance. However, to improve resource utilization, jobs are often share the same cluster, which causes dynamic cluster load conditions. According to our observation, the variation of cluster load reduces effectiveness of configuration tuning. Besides, as a common problem of cluster computing jobs, overestimation of resources also occurs during configuration tuning. It is challenging to efficiently find the optimal configuration in a shared cluster with the consideration of memory-sparing. In this article, we introduce MespaConfig, a job-level configuration optimizer for distributed in-memory computing jobs. Advancements of MespaConfig over previous work are features including memory-sparing and load-sensitive. We evaluate MespaConfig by 6 typical Spark programs under different load conditions. The evaluation results show that MespaConfig improves the performance of six typical programs by up to 12× compared with default configurations. MespaConfig also achieves at most 41 percent reduction of configuration memory usage and reduces the optimization time overhead by 10.8× compared with the state-of-the-art approach.

Word embedding can represent a word as a point vector or a Gaussian distribution in high-dimensional spaces. Gaussian distribution is innately more expressive than point vector owing to the ability to additionally capture semantic uncertainties of words, and thus can express asymmetric relations among words more naturally (e.g., animal entails cat but not the reverse. However, previous Gaussian embedders neglect inner-word conceptual knowledge and lack tailored Gaussian contextualizer, leading to inferior performance on both intrinsic (context-agnostic) and extrinsic (context-sensitive) tasks. In this paper, we first propose a novel Gaussian embedder which explicitly accounts for inner-word conceptual units (sememes) to represent word semantics more precisely; during learning, we propose Gaussian Distribution Attention over Gaussian representations to adaptively aggregate multiple sememe distributions into a word distribution, which guarantees the Gaussian linear combination property. Additionally, we propose a Gaussian contextualizer to utilize outer-word contexts in a sentence, producing contextualized Gaussian representations for context-sensitive tasks. Extensive experiments on intrinsic and extrinsic tasks demonstrate the effectiveness of the proposed approach, achieving state-of-the-art performance with near 5.00% relative improvement.

Chen Qian, Fuli Feng, Lijie Wen, Chunping Ma, Pengjun Xie. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Low-resource Relation Extraction (LRE) aims to extract relation facts from limited labeled corpora when human annotation is scarce.Existing works either utilize self-training scheme to generate pseudo labels that will cause the gradual drift problem, or leverage metalearning scheme which does not solicit feedback explicitly.To alleviate selection bias due to the lack of feedback loops in existing LRE learning paradigms, we developed a Gradient Imitation Reinforcement Learning method to encourage pseudo label data to imitate the gradient descent direction on labeled data and bootstrap its optimization capability through trial and error.We also propose a framework called GradLRE, which handles two major scenarios in low-resource relation extraction.Besides the scenario where unlabeled data is sufficient, GradLRE handles the situation where no unlabeled data is available, by exploiting a contextualized augmentation method to generate data.Experimental results on two public datasets demonstrate the effectiveness of GradLRE on low resource relation extraction when comparing with baselines.Source code is available 1 .

Many recently proposed code clone detectors exploit neural networks to capture latent semantics of source code, thus achieving impressive results for detecting semantic clones. These neural clone detectors rely on the availability of large amounts of labeled training data. We identify a key oversight in the current evaluation methodology for neural clone detection: cross-functionality generalization (i.e., detecting semantic clones of which the functionalities are unseen in training). Specifically, we focus on this question: do neural clone detectors truly learn the ability to detect semantic clones, or they just learn how to model specific functionalities in training data while cannot generalize to realistic unseen functionalitiesƒ This paper investigates how the generalizability can be evaluated and improved.Our contributions are 3-folds: (1) We propose an evaluation methodology that can systematically measure the cross-functionality generalizability of neural clone detection. Based on this evaluation methodology, an empirical study is conducted and the results indicate that current neural clone detectors cannot generalize well as expected. (2) We conduct empirical analysis to understand key factors that can impact the generalizability. We investigate 3 factors: training data diversity, vocabulary, and locality. Results show that the performance loss on unseen functionalities can be reduced through addressing the out-of-vocabulary problem and increasing training data diversity. (3) We propose a human-in-the-loop mechanism that help adapt neural clone detectors to new code repositories containing lots of unseen functionalities. It improves annotation efficiency with the combination of transfer learning and active learning. Experimental results show that it reduces the amount of annotations by about 88%. Our code and data are publicly available1.

Event logs are the basis of process mining techniques and tools that extract process behavior information for better understanding and optimization of business processes. While it has been widely realized that the degree of completeness of event logs may largely determine the effectiveness of these techniques, how to estimate the completeness of event logs has not yet been fully addressed. This is mainly because ground-truth process models are usually unknown. To attack this problem, we pay a closer look to several concepts and implicit assumptions in the log completeness estimation problem and characterize it as a special case of the species estimation problem in the field of statistics. Although species estimation is still an open problem, a number of statistic models and techniques with approximate solutions have been available. To investigate the relevance of these methods for event log completeness estimation, we have designed and conducted a wide scope of empirical study and quantitative experiments on both real-world and synthesized event logs to compare the performance of these methods. In addition, the completeness estimation of several important and widely used real-world events logs are reported for the first time together with some best practice experience learned through this research.

The abundant semi-structured data on the Web, such as HTML-based tables and lists, provide commercial search engines a rich information source for question answering (QA).Different from plain text passages in Web documents, Web tables and lists have inherent structures, which carry semantic correlations among various elements in tables and lists.Many existing studies treat tables and lists as flat documents with pieces of text and do not make good use of semantic information hidden in structures.In this paper, we propose a novel graph representation of Web tables and lists based on a systematic categorization of the components in semi-structured data as well as their relations.We also develop pre-training and reasoning techniques on the graph model for the QA task.Extensive experiments on several real datasets collected from a commercial engine verify the effectiveness of our approach.Our method improves F1 score by 3.90 points over the state-of-the-art baselines.

Open relation extraction is the task of extracting open-domain relation facts from natural language sentences.Existing works either utilize heuristics or distant-supervised annotations to train a supervised classifier over pre-defined relations, or adopt unsupervised methods with additional assumptions that have less discriminative power.In this work, we propose a self-supervised framework named SelfORE, which exploits weak, self-supervised signals by leveraging large pretrained language model for adaptive clustering on contextualized relational features, and bootstraps the self-supervised signals by improving contextualized features in relation classification.Experimental results on three datasets show the effectiveness and robustness of SelfORE on open-domain Relation Extraction when comparing with competitive baselines.Source code is available 1 .

The real-world networks often compose of different types of nodes and edges with rich semantics, widely known as heterogeneous information network (HIN). Heterogeneous network embedding aims to embed nodes into low-dimensional vectors which capture rich intrinsic information of heterogeneous networks. However, existing models either depend on manually designing meta-paths, ignore mutual effects between different semantics, or omit some aspects of information from global networks. To address these limitations, we propose a novel Graph-Aggregated Heterogeneous Network Embedding (GAHNE), which is designed to extract the semantics of HINs as comprehensively as possible to improve the results of downstream tasks based on graph convolutional neural networks. In GAHNE model, we develop several mechanisms that can aggregate semantic representations from different single-type sub-networks as well as fuse the global information into final embeddings. Extensive experiments on three real-world HIN datasets show that our proposed model consistently outperforms the existing state-of-the-art methods.

Text classification in low-resource languages (eg Thai) is of great practical value for some information retrieval applications (eg sentiment-analysis-based restaurant recommendation). Due to lacking large-scale corpus for learning comprehensive text representation, bilingual text classification which borrows the linguistics knowledge from a rich-resource language becomes a promising solution. Despite the success of bilingual methods, they largely ignore another source of semantic information---the writing system. Noting that most low-resource languages are phonographic languages, we argue that a logographic language (eg Chinese) can provide helpful information for improving some phonographic languages' text classification, since a logographic character (ie logogram) could represent a sememe or a whole concept, not only a phoneme or a sound. In this paper, by using a phonographic labeled corpus and its machine-translated logographic corpus both, we devise a framework to explore the central theme of utilizing logograms as a "semantic detection assistant''. Specifically, from a logographic labeled corpus, we first devise a statistical-significance-based module to pick out informative text pieces. To represent them and further reduce the effects of translation errors, our approach is equipped with Gaussian embedding whose covariances serve as reliable signals of translation errors. For a test document, all seeds' Gaussian representations are used to convolute the document and produce a logographic embedding, before being fused with its phonographic embedding for final prediction. Extensive experiments validate the effectiveness of our approach and further investigations show its generalizability and robustness.

Abstract As the promotion of technologies and applications of Big Data, the research of business process management (BPM) has gradually deepened to consider the impacts and challenges of big business data on existing BPM technologies. Recently, parallel business process mining (e.g. discovering business models from business visual data, integrating runtime business data with interactive business process monitoring visualisation systems and summarising and visualising historical business data for further analysis, etc.) and multi‐perspective business data analytics (e.g. pattern detecting, decision‐making and process behaviour predicting, etc.) have been intensively studied considering the steep increase in business data size and type. However, comprehensive and in‐depth testing is needed to ensure their quality. Testing based solely on existing business processes and their system logs is far from sufficient. Large‐scale randomly generated models and corresponding complete logs should be used in testing. To test parallel algorithms for discovering process models, different log completeness and generation algorithms were proposed. However, they suffer from either state space explosion or non‐full‐covering task dependencies problem. Besides, most existing generation algorithms rely on random executing strategy, which leads to low and unstable efficiency. In this paper, we propose a novel log completeness type, that is, #TAR completeness, as well as its generation algorithm. The experimental results based on a series of randomly generated process models show that the #TAR complete logs outperform the state‐of‐the‐art ones with lower capacity, fuller dependencies covering and higher generating efficiency.

Online discussion forums can be used for reflecting on the overall user experience of a system. If a user forum is well‐structured, it can be a valuable source of requirements‐related information, which can potentially be accommodated in the requirements engineering process to enhance the current and future software. However, presently, there are limited approaches for extracting such requirements‐related information from the relevant community forums. To fill this gap, this study proposes an automated approach, which automatically identifies requirements information using natural language processing and machine learning. For this purpose, the authors analysed 3319 user comments collected from the seven discussion topics in the Reddit forum. Then, using a content analysis approach, they studied how frequently end‐users submit such information across each discussion topic. Also, they developed an automated approach that identifies key stakeholders, who frequently contribute his rationales in the forum discussion. Further, they employed different machine learning algorithms to classify user comments into rationale elements of different types. The authors' results show that online forums, such as Reddit, can be a rich source of requirements elicitation. Also, machine learning is a promising tool to detect user's rationale and identify different kinds of requirements modelling elements.

Abstract Due to the pervasive use of online forums and social media, users' feedback are more accessible today and can be used within a requirements engineering context. However, such information is often fragmented, with multiple perspectives from multiple parties involved during on‐going interactions. In this paper, the authors propose a Crowd‐based Requirements Engineering approach by Argumentation (CrowdRE‐Arg). The framework is based on the analysis of the textual conversations found in user forums, identification of features, issues and the arguments that are in favour or opposing a given requirements statement. The analysis is to generate an argumentation model of the involved user statements, retrieve the conflicting‐viewpoints, reason about the winning‐arguments and present that to systems analysts to make informed‐requirements decisions. For this purpose, the authors adopted a bipolar argumentation framework and a coalition‐based meta‐argumentation framework as well as user voting techniques. The CrowdRE‐Arg approach and its algorithms are illustrated through two sample conversations threads taken from the Reddit forum. Additionally, the authors devised algorithms that can identify conflict‐free features or issues based on their supporting and attacking arguments. The authors tested these machine learning algorithms on a set of 3,051 user comments, preprocessed using the content analysis technique. The results show that the proposed algorithms correctly and efficiently identify conflict‐free features and issues along with their winning arguments.

Sharing process models on the web has emerged as a common practice. Users can collect and share their experimental process models with others. However, some users always feel confused about the shared process models for lack of necessary guidelines or instructions. Therefore, several process translators have been proposed to explain the semantics of process models in natural language (NL). We find that previous studies suffer from information loss and generate semantically erroneous descriptions that diverge from original model behaviors. In this paper, we propose a novel process translator named BePT (Behavior-based Process Translator) based on the encoder-decoder paradigm, encoding a process model into a middle representation and decoding the representation into NL descriptions. Our theoretical analysis demonstrates that BePT satisfies behavior correctness, behavior completeness and description minimality. The qualitative and quantitative experiments show that BePT outperforms the state-of-the-art baselines.

The variational neural models have achieved significant progress in dialogue generation. They are of encoder-decoder architecture, with stochastic latent variables learned at the utterance level. However, latent variables are usually approximated by factorized-form distributions, the value space of which is too large relative to latent features to be encoded, leading to the sparsity problem. As a result, little useful information is carried in latent representations, and generated responses tend to be non-committal and meaningless. To address it, we initially propose the Latent-Constrained Variational Neural Dialogue Model (LC-VNDM). It follows variational neural dialogue framework, with an utterance encoder, a context encoder and a response decoder hierarchically organized. Particularly, LC-VNDM uses a hierarchically-structured variational distribution form, which considers inter-dependencies between latent variables. Thus it defines a constrained latent value space, and prevents latent global features from being diluted. Therefore, latent representations sampled from it would carry richer global information to facilitate the decoding, generating meaningful responses. We conduct extensive experiments on three datasets using automatic evaluation and human evaluation. Experiments prove that LC-VNDM significantly outperforms the state-of-the-arts and can generate information-richer responses by learning a better-quality latent space.

In the past, users were asked to express their needs and intentions by writing a structured requirements document in natural language. Due to the pervasive use of online forums and social media, user feedback is more accessible today. However, the information obtained is often fragmented, involving multipleperspectives from multiple parties on an on-going basis. In this paper, we propose a Crowd-based Requirements Engineering approach by Argumentation (CrowdRE-Arg), which analyses the conversations from user forum, identifies the arguments in favor or opposing of a given requirements related discussion topic. By generating the argumentation model of the involved user statements, we are able to recover the conflicting viewpoints, to reason about the winning arguments for informed requirements decisions. The proposed approach is illustrated with a data set of sample conversations about the design of a new Google-Map feature from Reddit. Also, we apply natural language processing techniques and machine learning algorithms to support the automated execution of the CrowdRE-Arg approach.

Event sequence prediction has wide applications on economics, electronic health and social media monitoring. Accurate prediction of event sequences can help provide better service to customers and prevent risks. Recent works try to address the problem aiming at learning the impact of past events on the future events using deep learning methods. Such works often take the past event sequences as input and model the self-change transformations of the events, and few of them concerned the effect of event attributes. We propose an RNN-based predictive model to encode multiple attributes as attached information of the event for predicting next event and its attributes given past sequences. To learn how important each attribute is for the event, we design a component modulator to customize weights for representations of the event and its attributes. The more important the information is, the relevant weight will be higher. Finally, the prediction of next event and its attributes are conducted simultaneously with a different modulator for each predictive task. The performance of the proposed model was evaluated on 5 real-life datasets, containing two different types of event logs. The results show that our model outperforms the baselines and the state-of-the-art, not only on the prediction of next event and its attributes but also the generation of event sequence suffix.

Deep Generative Models (DGMs) are able to extract high-level representations from massive unlabeled data and are explainable from a probabilistic perspective. Such characteristics favor sequence modeling tasks. However, it still remains a huge challenge to model sequences with DGMs. Unlike real-valued data that can be directly fed into models, sequence data consist of discrete elements and require being transformed into certain representations first. This leads to the following two challenges. First, high-level features are sensitive to small variations of inputs as well as the way of representing data. Second, the models are more likely to lose long-term information during multiple transformations. In this paper, we propose a Hierarchical Deep Generative Model With Dual Memory to address the two challenges. Furthermore, we provide a method to efficiently perform inference and learning on the model. The proposed model extends basic DGMs with an improved hierarchically organized multi-layer architecture. Besides, our model incorporates memories along dual directions, respectively denoted as broad memory and deep memory. The model is trained end-to-end by optimizing a variational lower bound on data log-likelihood using the improved stochastic variational method. We perform experiments on several tasks with various datasets and obtain excellent results. The results of language modeling show our method significantly outperforms state-of-the-art results in terms of generative performance. Extended experiments including document modeling and sentiment analysis, prove the high-effectiveness of dual memory mechanism and latent representations. Text random generation provides a straightforward perception for advantages of our model.

Conceptual elevation of data in business process modeling was first formulated in 2003. The research community responded to this new idea enthusiastically. In the past decade, there have been numerous research activities concerning the interactions between business processes/activities and data in many aspects of business process management. Many of the advancements have or will have impacted on design/modeling, analysis, implementation of business processes in practice. However, there is still a lack of techniques for developing a suite of "interrelated" processes realizing a business service. Interrelated processes are a cluster of processes that share data and other resources, are collectively constrained by regulations and policies, influence KPIs as a group. Many current practical applications of business workflow systems are in urgent need for tools and techniques for process clusters. In this paper, we formulate a broad notion of an "Enterprise Process Framework" (EPF) to address this need and outline several interesting research challenges arising from EPFs.

With the broad use of business process management technology, there are more and more business process models. Since the ability of different modelers is different, the quality of these models varies. A question arises here is that, can we refactor these models to improve the quality as practised in software engineering? Business process modeling can be regarded as declarative programming, and business process models can be used to drive the process aware information systems, which are generally developed with model driven architecture, so business process models are crucial for the efficiency of process aware information systems. In this paper, we propose a novel approach on how to systematically refactor business process models with parallel structures for sequence structures for the first time. More specifically, we analyze the real causal relations between business tasks based on data operation dependency analysis, and refactor business process models with process mining technology. After comprehensive model refactoring, parallel execution of business tasks can be maximized, so the efficiency of business processing can be improved, that is, the quality of business process models can be improved. Analysis and experiments show that our approach is effective and efficient.

Since the behavior is the essential characteristic of business process models, and ordering relations between execution of tasks can be used to describe the behavior of process models, we need to compute the ordering relations between tasks in process models. This computation can be used for compliance checking and querying process models based on behavior. There are three basic types of ordering relations between two events in a concurrent system, i.e., causal, conflict, and concurrency. In this paper, we refine the causal and concurrency relations with uncertainty according to whether one task is always executed with the other task in the same instance. To compute the refined ordering relations with uncertainty efficiently, we propose some rules for adjacent tasks and some transitive laws for nonadjacent tasks together with their proofs. Based on these rules and laws, we propose an algorithm to compute the refined ordering relations for acyclic process models based on unfolding technology. The algorithm has a biquadrate time to the size of complete prefix unfolding of the original model.

While many process mining algorithms have been proposed recently, there does not exist a widely accepted benchmark to evaluate and compare these process mining algorithms. As a result, it can be difficult to choose a suitable process mining algorithm for a given enterprise or application domain. Some recent benchmark systems have been developed and proposed to address this issue. However, evaluating available process mining algorithms against a large set of business models (e.g., in a large enterprise) can be computationally expensive, tedious, and time-consuming. This paper investigates a scalable solution that can evaluate, compare, and rank these process mining algorithms efficiently, and hence proposes a novel framework that can efficiently select the process mining algorithms that are most suitable for a given model set. In particular, using our framework, only a portion of process models need empirical evaluation and others can be recommended directly via a regression model. As a further optimization, this paper also proposes a metric and technique to select high-quality reference models to derive an effective regression model. Experiments using artificial and real data sets show that our approach is practical and outperforms the traditional approach.

Process mining links traditional model-driven Business Process Management and data mining by means of deriving knowledge from event logs to improve operational business processes. As an impact factor of the quality of process mining results, the degree of completeness of the given event log should be necessarily measured. In this paper an approach is proposed in the context of mining control-flow dependencies to evaluate the local completeness of an event log without knowing any information about the original process model. Experiment results show that the proposed approach works robustly and gives better estimation than approaches available.

While many process mining algorithms have been proposed recently, there does not exist a widely-accepted benchmark to evaluate and compare these process mining algorithms. As a result, it can be difficult to choose a suitable process mining algorithm for a given enterprise or application domain. Some recent benchmark systems have been developed and proposed to address this issue. However, evaluating available process mining algorithms against a large set of business models (e.g., in a large enterprise) can be computationally expensive, tedious and time-consuming. This paper proposes a novel framework that can efficiently select the process mining algorithms that are most suitable for a given model set. In particular, it attemptsto investigate how we can avoid evaluating numerous process mining algorithms on all given process models.

Workflow management is concerned with automated support for business processes. Workflow management systems are driven by process models specifying the tasks that need to be executed, the order in which they can be executed, which resources are authorised to perform which tasks, and data that is required for, and produced by, these tasks. As workflow instances may run over a sustained period of time, it is important that workflow specifications be checked before they are deployed. Workflow verification is usually concerned with control-flow dependencies only; however, transition conditions based on data may further restrict possible choices between tasks. In this paper we extend workflow nets where transitions have concrete conditions associated with them, called WTC-nets. We then demonstrate that we can determine which execution paths of a WTC-net that are possible according to the control-flow dependencies, are actually possible when considering the conditions based on data. Thus, we are able to more accurately determine at design time whether a workflow net with transition conditions is sound.

In workflow management systems (WFMSs), time management plays an essential role in controlling the lifecycle of business processes. Especially, run-time analysis of time constraints is necessary to help process manger proactively detect possible deadline violations and appropriately handle these violations. Traditional time constraint analyses either present deterministic results which are too restrictive in highly uncertain workflow processes, or only consider static analysis at workflow build-time. For such an issue, this paper proposes a dynamic approach for analyzing time constraints during process execution. To be specific, based on a Petri-net-extended stochastic model, this approach first analyzes activity instances’ continuous probabilities of satisfying time constraints when a process instance is initiated. Afterwards, during the execution of this process instance, the approach dynamically updates these probabilities whenever an activity instance is completed. Moreover, an example process instance in real-world WFMSs shows the practicality of our approach.

Many activities in business process management, such as process search, process clustering, and process mining, need to determine the similarity between two process models. Although several approaches have recently been proposed to measure the behavioral similarity between business processes, all of them require that tasks in processes are properly labeled. According to these approaches, similarity between two given processes can be dramatically different under different task labeling schemes. In this paper, we consider process similarity measure from another view point, i.e., focusing on the control flow structures and ignoring the task labels. Thus, we propose a label-free similarity measure between process models based on transition adjacent relations (TARs) in the context of workflow nets (WF-nets), as well as an efficient algorithm. The experimental results involving comparison of different similarity measures on artificial processes and evaluation of the efficient algorithm on real-life processes are discussed.

Inefficient development of workflow-based applications reduces the benefits from the adoption of workflow technology. In this paper, we propose a new workflow product structure, in which application data are integrated into workflow process models. The new structure can facilitate workflow-based application development, as well as promote application flexibility against process change.

A large number of cohesion measures based on method-attribute references have been proposed. However, virtually no attention has been paid to the abstract representation that objectively depicts the relationships among the members of a class. Specially, the flow dependence relationship among attributes, the indirect and potential dependence relationships among class members, and the direction of method-attribute references are ignored. To address this problem, we first identifies four types of basic dependence relationships and uses a class member dependence graph to represent all dependences among the members of a class. Then, a dependence relationships based measure for measuring the class cohesiveness is proposed. Finally, we compare our class cohesion measure with typical cohesion measures.