Search

Emotion-Cause Pair Extraction (ECPE) aims to extract all emotion clauses and their corresponding cause clauses from a document. Existing approaches tackle this task through multi-task learning (MTL) framework in which the two subtasks provide indicative clues for ECPE. However, the previous MTL framework considers only one round of multi-task reasoning and ignores the reverse feedbacks from ECPE to the subtasks. Besides, its multi-task reasoning only relies on semantics-level interactions, which cannot capture the explicit dependencies, and both the encoder sharing and multi-task hidden states concatenations can hardly capture the causalities. To solve these issues, we first put forward a new MTL framework based on Co-evolving Reasoning. It (1) models the bidirectional feedbacks between ECPE and its subtasks; (2) allows the three tasks to evolve together and prompt each other recurrently; (3) integrates prediction-level interactions to capture explicit dependencies. Then we propose a novel multi-task relational graph (MRG) to sufficiently exploit the causal relations. Finally, we propose a Co-evolving Graph Reasoning Network (CGR-Net) that implements our MTL framework and conducts Co-evolving Reasoning on MRG. Experimental results show that our model achieves new state-of-the-art performance, and further analysis confirms the advantages of our method.

Multi-task learning (MTL) is a widely adopted machine learning paradigm in recommender systems. However, existing MTL models often suffer from performance degeneration with negative transfer and seesaw phenomena. Some works attempt to alleviate the negative transfer and seesaw issues by separating task-specific and shared experts to mitigate the harmful interference between task-specific and shared knowledge. Despite the success of these efforts, task-specific and shared knowledge have still not been thoroughly decoupled. There may still exist unnecessary mixture between the shared and task-specific knowledge, which may harm MLT models' performances. To tackle this problem, in this paper, we propose multi-task knowledge disentanglement network (MTKDN) to further reduce harmful interference between the shared and task-specific knowledge. Specifically, we propose a novel contrastive disentanglement mechanism to explicitly decouple the shared and task-specific knowledge in corresponding hidden spaces. In this way, the unnecessary mixture between shared and task-specific knowledge can be reduced. As for optimization objectives, we propose individual optimization objectives for shared and task-specific experts, by which we can encourage these two kinds of experts to focus more on extracting the shared and task-specific knowledge, respectively. Additionally, we propose a margin regularization to ensure that the fusion of shared and task-specific knowledge can outperform exploiting either of them alone. We conduct extensive experiments on open-source large-scale recommendation datasets. The experimental results demonstrate that MTKDN significantly outperforms state-of-the-art MTL models. In addition, the ablation experiments further verify the necessity of our proposed contrastive disentanglement mechanism and the novel loss settings.

Dual-task dialog language understanding aims to tackle two correlative dialog language understanding tasks simultaneously via leveraging their inherent correlations. In this paper, we put forward a new framework, whose core is relational temporal graph reasoning. We propose a speaker-aware temporal graph (SATG) and a dual-task relational temporal graph (DRTG) to facilitate relational temporal modeling in dialog understanding and dual-task reasoning. Besides, different from previous works that only achieve implicit semantics-level interactions, we propose to model the explicit dependencies via integrating prediction-level interactions. To implement our framework, we first propose a novel model Dual-tAsk temporal Relational rEcurrent Reasoning network (DARER), which first generates the context-, speaker- and temporal-sensitive utterance representations through relational temporal modeling of SATG, then conducts recurrent dual-task relational temporal graph reasoning on DRTG, in which process the estimated label distributions act as key clues in prediction-level interactions. And the relational temporal modeling in DARER is achieved by relational graph convolutional networks (RGCNs). Then we further propose Relational Temporal Transformer (ReTeFormer), which achieves fine-grained relational temporal modeling via Relation- and Structure-aware Disentangled Multi-head Attention. Accordingly, we propose DARER with ReTeFormer (DARER 2), which adopts two variants of ReTeFormer to achieve the relational temporal modeling of SATG and DTRG, respectively. The extensive experiments on different scenarios verify that our models outperform state-of-the-art models by a large margin. Remarkably, on the dialog sentiment classification task in the Mastodon dataset, DARER and DARER 2 gain relative improvements of about 28% and 34% over the previous best model in terms of F1.

Recent graph-based models for multi-intent SLU have obtained promising results through modeling the guidance from the prediction of intents to the decoding of slot filling. However, existing methods (1) only model the unidirectional guidance from intent to slot, while there are bidirectional inter-correlations between intent and slot; (2) adopt homogeneous graphs to model the interactions between the slot semantics nodes and intent label nodes, which limit the performance. In this paper, we propose a novel model termed Co-guiding Net, which implements a two-stage framework achieving the mutual guidances between the two tasks. In the first stage, the initial estimated labels of both tasks are produced, and then they are leveraged in the second stage to model the mutual guidances. Specifically, we propose two heterogeneous graph attention networks working on the proposed two heterogeneous semantics-label graphs, which effectively represent the relations among the semantics nodes and label nodes. Besides, we further propose Co-guiding-SCL Net, which exploits the single-task and dual-task semantics contrastive relations. For the first stage, we propose single-task supervised contrastive learning, and for the second stage, we propose co-guiding supervised contrastive learning, which considers the two tasks' mutual guidances in the contrastive learning procedure. Experiment results on multi-intent SLU show that our model outperforms existing models by a large margin, obtaining a relative improvement of 21.3% over the previous best model on MixATIS dataset in overall accuracy. We also evaluate our model on the zero-shot cross-lingual scenario and the results show that our model can relatively improve the state-of-the-art model by 33.5% on average in terms of overall accuracy for the total 9 languages.

Microplastics (MPs), as an emerging pollutant, may cause deleterious changes to the nitrogen cycle in terrestrial ecosystems. However, single impact of MPs and synergistic effects of MPs with hydrochar on ammonia (NH3) volatilization and soil microbiome in paddy fields has been largely unexplored. In this study, polyethylene (PE), polyacrylonitrile (PAN) and straw-derived hydrochar (HBC) were selected for observations in an entire rice cycle growth period. Results showed that under the condition of 0.5% (w/w) MPs concentration, presence of MPs alone and co-existence of MPs and HBC (MPs + HBC) unexpectedly mitigated cumulative NH3 volatilization from paddy soil compared with the control with no MPs or HBC addition. MPs + HBC increased NH3 volatilization by 37.8–46.2% compared with MPs alone, indicating that co-existence of MPs and HBC weaken the mitigation effect of MPs on NH3 volatilization. Additionally, results of nitrogen cycle related microorganisms closely related to NH3 volatilization demonstrated that MPs + HBC altered the bacterial community structure and species diversity. These findings provide an important opportunity to advance our understanding of the impacts of MPs in agricultural environment and soils, and provide a sound theoretical basis for rationalizing the application of HBC in soil with MPs.

Microplastics (MPs) are verified to affect the fate of ammonia (NH3) in agricultural soils. However, the impacts and mechanisms of MPs coupled with biochar (BC), a widely used agricultural conditioner, on NH3 losses are mostly untapped. The aim of this study was to investigate the mechanisms of common MPs (i.e., polyethylene, polyester, and polyacrylonitrile) and straw-derived BC on NH3 volatilization in rice-wheat rotation soils. Results showed that BC alone and MPs with BC (MPs + BC) reduced 5.5 % and 11.2–26.6 % cumulative NH3 volatilization than the control (CK), respectively, in the rice season. The increased nitrate concentration and soil cation exchange capacity were dominant contributors to the reduced soil NH3 volatilization in the rice season. BC and MPs + BC persistently reduced 44.5 % and 60.0–62.6 % NH3 losses than CK in the wheat season as influenced by pH and nitrate concentration. Moreover, BC and MPs + BC increased humic acid-like substances in soil dissolved organic matter by an average of 159.1 and 179.6 % than CK, respectively, in rice and wheat seasons. The increased adsorption of soil ammonium roots and the promotion of crop root growth were the main mechanism of NH3 reduction. Our findings partially revealed the mechanisms of the coexistence of MPs and BC on NH3 mitigation in rice-wheat rotational ecosystems.

High-silica phosphogypsum (PG) is a kind of industrial by-product with great utilization potential. However, it is difficult to reuse PG directly due to the related gangue minerals (e.g., SiO2), and thus efficient purification is required to allow its further applications. Herein, a typical high-silica phosphogypsum waste was purified by a new "reverse-direct flotation" method. The organic matters and fine slimes were removed by reverse flotation, and then, the silica impurity was removed by direct flotation. Via the closed-circuit flotation process, the whiteness of the PG concentrate is improved from 33.23 to 63.42, and the purity of gypsum in the PG concentrate increases from 83.90% to 96.70%, with a gypsum recovery of 85%. Additionally, the content of SiO2 is significantly reduced from 11.11% to 0.07%. In-depth investigations suggest that the difference in the floatability of gypsum and quartz is prominently intensified by flotation reagents at pH = 2-2.5, and thus leads to good desilication performance. Further characteristics of the PG concentrate prove that impurities have been well removed, and the PG concentrate meets the requirement of related standards for gypsum building materials. The flotation method reported here paves the way for the purification of high-silica phosphogypsum, which can be extended to the purification and value-added reutilization of other industrial solid wastes.

Recent years have witnessed the emerging success of leveraging syntax graphs for the target sentiment classification task. However, we discover that existing syntax-based models suffer from two issues: noisy information aggregation and loss of distant correlations. In this paper, we propose a novel model termed Neural Subgraph Explorer, which (1) reduces the noisy information via pruning target-irrelevant nodes on the syntax graph; (2) introduces beneficial first-order connections between the target and its related words into the obtained graph. Specifically, we design a multi-hop actions score estimator to evaluate the value of each word regarding the specific target. The discrete action sequence is sampled through Gumble-Softmax and then used for both of the syntax graph and the self-attention graph. To introduce the first-order connections between the target and its relevant words, the two pruned graphs are merged. Finally, graph convolution is conducted on the obtained unified graph to update the hidden states. And this process is stacked with multiple layers. To our knowledge, this is the first attempt of target-oriented syntax graph pruning in this task. Experimental results demonstrate the superiority of our model, which achieves new state-of-the-art performance.

To fully understand the potential ecological and human health risks from nanoplastics and microplastics (NMPs) in the environment, it is critical to make accurate measurements. Similar to past research on the toxicology of engineered nanomaterials, a broad range of measurement artifacts and biases are possible when testing their potential toxicity. For example, antimicrobials and surfactants may be present in commercially available NMP dispersions, and these compounds may account for toxicity observed instead of being caused by exposure to the NMP particles. Therefore, control measurements are needed to assess potential artifacts, and revisions to the protocol may be needed to eliminate or reduce the artifacts. In this paper, we comprehensively review and suggest a next generation of control experiments to identify measurement artifacts and biases that can occur while performing NMP toxicity experiments. This review covers the broad range of potential NMP toxicological experiments, such as in vitro studies with a single cell type or complex 3-D tissue constructs, in vivo mammalian studies, and ecotoxicity experiments testing pelagic, sediment, and soil organisms. Incorporation of these control experiments can reduce the likelihood of false positive and false negative results and more accurately elucidate the potential ecological and human health risks of NMPs.

Aspect-level sentiment classification (ASC) aims to predict the fine-grained sentiment polarity towards a given aspect mentioned in a review. Despite recent advances in ASC, enabling machines to preciously infer aspect sentiments is still challenging. This paper tackles two challenges in ASC: (1) due to lack of aspect knowledge, aspect representation derived in prior works is inadequate to represent aspect's exact meaning and property information; (2) prior works only capture either local syntactic information or global relational information, thus missing either one of them leads to insufficient syntactic information. To tackle these challenges, we propose a novel ASC model which not only end-to-end embeds and leverages aspect knowledge but also marries the two kinds of syntactic information and lets them compensate for each other. Our model includes four key components: (1) a knowledge-aware gated recurrent memory network recurrently integrates dynamically summarized aspect knowledge; (2) a dual syntax graph network combines both kinds of syntactic information to comprehensively capture sufficient syntactic information; (3) a knowledge integrating gate re-enhances the final representation with further needed aspect knowledge; (4) an aspect-to-context attention mechanism aggregates the aspect-related semantics from all hidden states into the final representation. Experimental results on several benchmark datasets demonstrate the effectiveness of our model, which overpass previous state-of-the-art models by large margins in terms of both Accuracy and Macro-F1. To facilitate further research in the community, we have released our source code at https://github.com/XingBowen714/KaGRMN-DSG-ABSA.

Aspect-based sentiment analysis (ABSA) aims to predict the sentiment expressed in a review with respect to a given aspect. The core of ABSA is to model the interaction between the context and given aspect to extract aspect-related information. In prior work, attention mechanisms and dependency graph networks are commonly adopted to capture the relations between the context and given aspect. And the weighted sum of context hidden states is used as the final representation fed to the classifier. However, the information related to the given aspect may be already discarded and adverse information may be retained in the context modeling processes of existing models. Such a problem cannot be solved by subsequent modules due to two reasons. First, their operations are conducted on the encoder-generated context hidden states, whose value cannot be changed after the encoder. Second, existing encoders only consider the context while not the given aspect. To address this problem, we argue the given aspect should be considered as a new clue out of context in the context modeling process. As for solutions, we design three streams of aspect-aware context encoders: an aspect-aware LSTM, an aspect-aware GCN, and three aspect-aware BERTs. They are dedicated to generating aspect-aware hidden states which are tailored for the ABSA task. In these aspect-aware context encoders, the semantics of the given aspect is used to regulate the information ow. Consequently, the aspect-related information can be retained and aspect-irrelevant information can be excluded in the generated hidden states. We conduct extensive experiments on several benchmark datasets with empirical analysis, demonstrating the efficacies and advantages of our proposed aspect-aware context encoders.

Background: The Lancet Commission on Global Surgery established the Three Delays framework, categorising delays in accessing timely surgical care into delays in seeking care (First Delay), reaching care (Second Delay), and receiving care (Third Delay). Globally, knowledge gaps regarding delays for fracture care, and the lack of large prospective studies informed the rationale for our international observational study. We investigated delays in hospital admission as a surrogate for accessing timely fracture care and explored factors associated with delayed hospital admission. Methods: In this prospective observational substudy of the ongoing International Orthopaedic Multicenter Study in Fracture Care (INORMUS), we enrolled patients with fracture across 49 hospitals in 18 low-income and middle-income countries, categorised into the regions of China, Africa, India, south and east Asia, and Latin America. Eligible patients were aged 18 years or older and had been admitted to a hospital within 3 months of sustaining an orthopaedic trauma. We collected demographic injury data and time to hospital admission. Our primary outcome was the number of patients with open and closed fractures who were delayed in their admission to a treating hospital. Delays for patients with open fractures were defined as being more than 2 h from the time of injury (in accordance with the Lancet Commission on Global Surgery) and for those with closed fractures as being a delay of more than 24 h. Secondary outcomes were reasons for delay for all patients with either open or closed fractures who were delayed for more than 24 h. We did logistic regression analyses to identify risk factors of delays of more than 2 h in patients with open fractures and delays of more than 24 h in patients with closed fractures. Logistic regressions were adjusted for region, age, employment, urban living, health insurance, interfacility referral, method of transportation, number of fractures, mechanism of injury, and