The pursuit of model explainability has prompted the selective rationalization (aka, rationale extraction) which can identify important features (i.e., rationales) from the original input to support prediction results. Existing methods typically involve a cascaded approach with a selector responsible for extracting rationales from the input, followed by a predictor that makes predictions based on the selected rationales. However, these approaches often neglect the information contained in the non-rationales, underutilizing the input. Therefore, in our prior work, we introduce the Disentanglement-Augmented Rationale Extraction (DARE) method, which disentangles the input into rationale and non-rationale components, and enhances rationale representations by minimizing the mutual information between them. While DARE demonstrates strong performance in rationalization, it may still rely on shortcuts in the training distribution, leading to unfaithful rationales. To this end, in this paper, we propose Faith-DARE, an extension of DARE that aims to extract more reliable rationales by mitigating shortcut dependencies. Specifically, we treat the non-rationale features identified by DARE as environments that are decorrelated from the predictions. By shuffling and recombining these environments with rationales, we generate counterfactual samples and identify invariant rationales that remain predictive across shifted distributions. Extensive experiments on graph and textual datasets validate the effectiveness of Faith-DARE.

The remarkable success in neural networks provokes the selective rationalization. It explains the prediction results by identifying a small subset of the inputs sufficient to support them. Since existing methods still suffer from adopting the shortcuts in data to compose rationales and limited large-scale annotated rationales by human, in this paper, we propose a Shortcuts-fused Selective Rationalization (SSR) method, which boosts the rationalization by discovering and exploiting potential shortcuts. Specifically, SSR first designs a shortcuts discovery approach to detect several potential shortcuts. Then, by introducing the identified shortcuts, we propose two strategies to mitigate the problem of utilizing shortcuts to compose rationales. Finally, we develop two data augmentations methods to close the gap in the number of annotated rationales. Extensive experimental results on real-world datasets clearly validate the effectiveness of our proposed method.

Graph Neural Networks (GNNs) have demonstrated remarkable performance in graph classification tasks. However, ensuring the explainability of their predictions remains a challenge. To address this, graph rationalization methods have been introduced to generate concise subsets of the original graph, known as rationales, which serve to explain the predictions made by GNNs. Existing rationalizations often rely on shortcuts in data for prediction and rationale composition. In response, de-shortcut rationalization methods have been proposed, which commonly leverage counterfactual augmentation to enhance data diversity for mitigating the shortcut problem. Nevertheless, these methods have predominantly focused on centralized datasets and have not been extensively explored in the Federated Learning (FL) scenarios. To this end, in this paper, we propose a Federated Graph Rationalization (FedGR) with anti-shortcut augmentations to achieve self-explaining GNNs, which involves two data augmenters. These augmenters are employed to produce client-specific shortcut conflicted samples at each client, which contributes to mitigating the shortcut problem under the FL scenarios. Experiments on real-world benchmarks and synthetic datasets validate the effectiveness of FedGR under the FL scenarios.

With the development of legal intelligence, Criminal Court View Generation has attracted much attention as a crucial task of legal intelligence, which aims to generate concise and coherent texts that summarize case facts and provide explanations for verdicts. Existing researches explore the key information in case facts to yield the court views. Most of them employ a coarse-grained approach that partitions the facts into broad segments (e.g., verdict-related sentences) to make predictions. However, this approach fails to capture the complex details present in the case facts, such as various criminal elements and legal events. To this end, in this paper, we propose an Event Grounded Generation (EGG) method for criminal court view generation with cooperative (Large) Language Models, which introduces the fine-grained event information into the generation. Specifically, we first design a LLMs-based extraction method that can extract events in case facts without massive annotated events. Then, we incorporate the extracted events into court view generation by merging case facts and events. Besides, considering the computational burden posed by the use of LLMs in the extraction phase of EGG, we propose a LLMs-free EGG method that can eliminate the requirement for event extraction using LLMs in the inference phase. Extensive experimental results on a real-world dataset clearly validate the effectiveness of our proposed method.

Selective rationalizations improve the explainability of neural networks by selecting a subsequence of the input (i.e., rationales) to explain the prediction results.Although existing methods have achieved promising results, they still suffer from adopting the spurious correlations in data (aka., shortcuts) to compose rationales and make predictions.Inspired by the causal theory, in this paper, we develop an interventional rationalization (Inter-RAT) to discover the causal rationales.Specifically, we first analyse the causalities among the input, rationales and results with a causal graph.Then, we discover spurious correlations between the input and rationales, and between rationales and results, respectively, by identifying the confounder in the causalities.Next, based on the backdoor adjustment, we propose a causal intervention method to remove the spurious correlations between input and rationales.Further, we discuss reasons why spurious correlations between the selected rationales and results exist by analysing the limitations of the sparsity constraint in the rationalization, and employ the causal intervention method to remove these correlations.Extensive experimental results on three realworld datasets clearly validate the effectiveness of our proposed method.The source code of Inter-RAT is available at https://github. com/yuelinan/Codes-of-Inter-RAT.

Rationale extraction can be considered as a straightforward method of improving the model explainability, where rationales are a subsequence of the original inputs, and can be extracted to support the prediction results. Existing methods are mainly cascaded with the selector which extracts the rationale tokens, and the predictor which makes the prediction based on selected tokens. Since previous works fail to fully exploit the original input, where the information of non-selected tokens is ignored, in this paper, we propose a Disentanglement-Augmented Rationale Extraction (DARE) method, which encapsulates more information from the input to extract rationales. Specifically, it first disentangles the input into the rationale representations and the non-rationale ones, and then learns more comprehensive rationale representations for extracting by minimizing the mutual information (MI) between the two disentangled representations. Besides, to improve the performance of MI minimization, we develop a new MI estimator by exploring existing MI estimation methods. Extensive experimental results on three real-world datasets and simulation studies clearly validate the effectiveness of our proposed method. Code is released at https://github.com/yuelinan/DARE.

Since fake news poses a serious threat to society and individuals, numerous studies have been brought by considering text, propagation and user profiles. Due to the data collection problem, these methods based on propagation and user profiles are less applicable in the early stages. A good alternative method is to detect news based on text as soon as they are released, and a lot of text-based methods were proposed, which usually utilized words, sentences or paragraphs as basic units. But, word is a too fine-grained unit to express coherent information well, sentence or paragraph is too coarse to show specific information. Which granularity is better and how to utilize it to enhance text representation for fake news detection are two key problems. In this paper, we introduce Elementary Discourse Unit (EDU) whose granularity is between word and sentence, and propose a multi-EDU-structure awareness model to improve text representation for fake news detection, namely EDU4FD. For the multi-EDU-structure awareness, we build the sequence-based EDU representations and the graph-based EDU representations. The former is gotten by modeling the coherence between consecutive EDUs with TextCNN that reflect the semantic coherence. For the latter, we first extract rhetorical relations to build the EDU dependency graph , which can show the global narrative logic and help deliver the main idea truthfully. Then a Relation Graph Attention Network (RGAT) is set to get the graph-based EDU representation. Finally, the two EDU representations are incorporated as the enhanced text representation for fake news detection, using a gated recursive unit combined with a global attention mechanism. Experiments on four cross-source fake news datasets show that our model outperforms the state-of-the-art text-based methods. Our results suggest that considering EDU and its structural features could enhance the text representation for fake news detection.

In recommender systems, the cold-start problem is a critical issue. To alleviate this problem, an emerging direction adopts meta-learning frameworks and achieves success. Most existing works aim to learn globally shared prior knowledge across all users so that it can be quickly adapted to a new user with sparse interactions. However, globally shared prior knowledge may be inadequate to discern users’ complicated behaviors and causes poor generalization. Therefore, we argue that prior knowledge should be locally shared by users with similar preferences who can be recognized by social relations. To this end, in this paper, we propose a Preference-Adaptive Meta-Learning approach (PAML) to improve existing meta-learning frameworks with better generalization capacity. Specifically, to address two challenges imposed by social relations, we first identify reliable implicit friends to strengthen a user’s social relations based on our defined palindrome paths. Then, a coarse-fine preference modeling method is proposed to leverage social relations and capture the preference. Afterwards, a novel preference-specific adapter is designed to adapt the globally shared prior knowledge to the preference-specific knowledge so that users who have similar tastes share similar knowledge. We conduct extensive experiments on two publicly available datasets. Experimental results validate the power of social relations and the effectiveness of PAML.

Criminal Court View Generation is an essential task in legal intelligence, which aims to automatically generate sentences interpreting judgment results. The court view could be seen as the summary of crime circumstances in a case, including ADjudging Circumstance (ADC) and SEntencing Circumstance (SEC). However, different circumstances vary widely, and adopting them to generate court views directly may limit the generation performance. Therefore, it is necessary to identify the ADC and SEC related sentences in case facts and enhance them into the court view generation, respectively. To this end, in this paper, we propose a novel Circumstances enhanced Criminal Court View Generation (C3VG) method, consisting of the extraction and generation stage. Specifically, in the extraction stage, we design a Circumstances Selector to select ADC and SEC related sentences. After that, we apply them to two generators to generate the circumstances enhanced court views, respectively. After merging the two types of court views, we could obtain the final court views. We evaluate C3VG by conducting extensive experiments on a real-world dataset and experimental results clearly validate the effectiveness of our proposed model.

Many networks in real applications are constantly evolving as the creation and elimination of nodes and edges. Dynamic link prediction aims to infer whether there will be an edge between a pair of nodes, given the recent evolution history of the network. In this paper, we devise a flexible framework for link prediction on dynamic networks regularly archived as different snapshots. On the basis of node vectors learned on individual snapshots, a gated recurrent unit (GRU) network is utilized to model the node vector evolution series and predict the node representation in the future. Then, the edge representation is not only constructed from the interaction between representations of the target node pair, but also enriched with local neighborhood representations---historical embeddings of their common neighbors. Finally, a binary classifier is trained to perform link prediction. The framework can be instantiated with many off-the-shelf outstanding node embedding and binary classification methods. Extensive experiments on three different datasets demonstrate the effectiveness and flexibility of our proposed framework. Ablation studies show that the node vector evolution and local neighborhood representation both have positive but different effects on dynamic link prediction on diverse networks.

Nowadays the indoor location context becomes an important element in a number of real applications. Use of WiFi signals to fulfil the location detection of WiFi-enabled devices is a promising approach. In this paper an AP (Access Point)- centred indoor positioning system is proposed to address some common concerns in the conventional MH (Mobile Handheld)-based positioning system, such as excessive involvement of MH, in particular for scenarios of positioning multiple MHs simultaneously. Meanwhile the popularly-used fingerprint technique is combined into the AP-centred architecture to achieve higher positioning accuracy. The proposed system is fully designed, implemented and tested in a real-world deployment. In the environment covered by the APs running the proposed system, the location of the WiFi-enabled MHs appearing in this environment can be computed by a positioning server without disturbing MHs. The accuracy of positioning result obtained from the AP- centred positioning system is evaluated in comparison with a traditional MH-based system in the real experiments. The proposed AP-centred system shows not only the feasibility of AP-centred positioning but also better performance on positioning accuracy and energy consumption of MH.

In this paper we present a middleware named SmartPrint to provide cloud print service in office, where many heterogeneous networks exist. The goal of the system is to shield the communication heterogeneity of the devices in the office and make authorized users freely connect to all the printers with no modification on their terminals. SmartPrint can manages all the printers in an office building, and it provides friendly service for the users who know nothing about the printers. SmartPrint can also automatically choose printers for the office staffs. We propose and implement two printer allocation methods, one aims to improve the experience of the user with short print job, and the other is a multiple attributes decision algorithm which considers all factors including spatial information that impact the user experiences. Through experiments we validate the methods, and prove that SmartPrint achieves high user satisfaction from collected real data.

Description Logic possesses strong knowledge representation and reasoning capabilities and offers logical foundation for Semantic Web ontology languages such as OWL and OWL-S. However, the present implementations of OWL and OWL-S are deficient in sem

Trojan's attack behavior has become increasingly common and diversifiable. How to judge Trojan-like features of the softwares which the users download has become the problem that the users concern about. In this paper, we first capture the software's behavior and related parameters from our virtual software test bed, then a modeling method using Stochastic Petri Nets is proposed, which supports quantitative analysis for the application software's behaviors. Based on the model, the similarity degree between application software and Trojan software is analyzed quantitatively. This analysis show that the model can be used to design an effective anti-Trojan system. The paper concludes with an example to illustrate the effectiveness of the model and analysis method.

The description logic, which possesses strong knowledge representation and reasoning capabilities, is the logic basis of the semantic Web ontology languages such as OWL and OWL-S, but OWL and OWL-S are deficient in the semantic modeling of the dynamic services composition and also do not consider the user preferences in the dynamic services composition. The AI planning, which provides an effective method for solving the planning problem and task decomposition in AI, possesses better modeling capability of the action state transformation, but the AI planning is limited in the knowledge representation and reasoning capabilities. Based on the merits of the description logic, OWL-S and the AI planning, this paper extends the OWL-S model, proposes a service composition mechanism and testifies its feasibility in description logic. The results show that this composition mechanism can not only be feasible but also be helpful for the semantic modeling of the services composite process in the semantic Web