Soft prompt tuning is a widely studied parameter-efficient fine-tuning method.However, it has a clear drawback: many soft tokens must be inserted into the input sequences to guarantee downstream performance.As a result, soft prompt tuning is less considered than Low-rank adaptation (LoRA) in the large language modeling (LLM) era.In this work, we propose a novel prompt tuning method, Instruction-Aware Prompt Tuning (IAPT), that requires only four soft tokens.First, we install a parameter-efficient soft prompt generator at each Transformer layer to generate idiosyncratic soft prompts for each input instruction.The generated soft prompts can be seen as a semantic summary of the input instructions and can effectively guide the output generation.Second, the soft prompt generators are modules with a bottleneck architecture consisting of a self-attention pooling operation, two linear projections, and an activation function.Pilot experiments show that prompt generators at different Transformer layers require different activation functions.Thus, we propose to learn the idiosyncratic activation functions for prompt generators automatically with the help of rational functions.We have conducted experiments on various tasks, and the experimental results demonstrate that (a) our IAPT method can outperform the recent baselines with comparable tunable parameters.(b) Our IAPT method is more efficient than LoRA under the singlebackbone multi-tenant setting.

Learning from human feedback has become a pivot technique in aligning large language models (LLMs) with human preferences. However, acquiring vast and premium human feedback is bottlenecked by time, labor, and human capability, resulting in small sizes or limited topics of current datasets. This further hinders feedback learning as well as alignment research within the open-source community. To address this issue, we explore how to go beyond human feedback and collect high-quality AI feedback automatically for a scalable alternative. Specifically, we identify scale and diversity as the key factors for feedback data to take effect. Accordingly, we first broaden instructions and responses in both amount and breadth to encompass a wider range of user-assistant interactions. Then, we meticulously apply a series of techniques to mitigate annotation biases for more reliable AI feedback. We finally present UltraFeedback, a large-scale, high-quality, and diversified AI feedback dataset, which contains over 1 million GPT-4 feedback for 250k user-assistant conversations from various aspects. Built upon UltraFeedback, we align a LLaMA-based model by best-of-$n$ sampling and reinforcement learning, demonstrating its exceptional performance on chat benchmarks. Our work validates the effectiveness of scaled AI feedback data in constructing strong open-source chat language models, serving as a solid foundation for future feedback learning research.

MOTIVATION: Natural language is poised to become a key medium for human-machine interactions in the era of large language models. In the field of biochemistry, tasks such as property prediction and molecule mining are critically important yet technically challenging. Bridging molecular expressions in natural language and chemical language can significantly enhance the interpretability and ease of these tasks. Moreover, it can integrate chemical knowledge from various sources, leading to a deeper understanding of molecules. RESULTS: Recognizing these advantages, we introduce the concept of conversational molecular design, a novel task that utilizes natural language to describe and edit target molecules. To better accomplish this task, we develop ChatMol, a knowledgeable and versatile generative pretrained model. This model is enhanced by incorporating experimental property information, molecular spatial knowledge, and the associations between natural and chemical languages. Several typical solutions including large language models (e.g. ChatGPT) are evaluated, proving the challenge of conversational molecular design and the effectiveness of our knowledge enhancement approach. Case observations and analysis offer insights and directions for further exploration of natural-language interaction in molecular discovery. AVAILABILITY AND IMPLEMENTATION: Codes and data are provided in https://github.com/Ellenzzn/ChatMol/tree/main.

Xiaonan Li, Kai Lv, Hang Yan, Tianyang Lin, Wei Zhu, Yuan Ni, Guotong Xie, Xiaoling Wang, Xipeng Qiu. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

Yusheng Su, Chi-Min Chan, Jiali Cheng, Yujia Qin, Yankai Lin, Shengding Hu, Zonghan Yang, Ning Ding, Xingzhi Sun, Guotong Xie, Zhiyuan Liu, Maosong Sun. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

Filter pruning is widely adopted to compress and accelerate the Convolutional Neural Networks (CNNs), but most previous works ignore the relationship between filters and channels in different layers. Processing each layer independently fails to utilize the collaborative relationship across layers. In this paper, we intuitively propose a novel pruning method by explicitly leveraging the Filters Similarity in Consecutive Layers (FSCL). FSCL compresses models by pruning filters whose corresponding features are more worthless in the model. The extensive experiments demonstrate the effectiveness of FSCL, and it yields remarkable improvement over state-of-the-art on accuracy, FLOPs and parameter reduction on several benchmark models and datasets.

Contrastive learning (CL) has achieved great success in various fields with self-supervised learning. However, CL under the supervised setting is not fully explored, especially how to utilize the class labels in CL. We propose a novel aligned contrastive finetuning (ACF) approach in this work. Specifically, we consider the label embeddings as labeled instances and put them in an InfoNCE loss objective together with the instance representations, thus aligning the label embeddings and instance representation in the same semantic space. In addition, we design a correlation-based regularization term to alleviate the anisotropy problem. Extensive experiments are conducted on language understanding and image classification tasks, demonstrating our ACF method’s competitiveness. ACF is off-the-shelf and can be plugged into any pre-trained models without additional network architectures or computation overhead.

Rapid and accurate prediction of drug-target affinity can accelerate and improve the drug discovery process. Recent studies show that deep learning models may have the potential to provide fast and accurate drug-target affinity prediction. However, the existing deep learning models still have their own disadvantages that make it difficult to complete the task satisfactorily. Complex-based models rely heavily on the time-consuming docking process, and complex-free models lacks interpretability. In this study, we introduced a novel knowledge-distillation insights drug-target affinity prediction model with feature fusion inputs to make fast, accurate and explainable predictions. We benchmarked the model on public affinity prediction and virtual screening dataset. The results show that it outperformed previous state-of-the-art models and achieved comparable performance to previous complex-based models. Finally, we study the interpretability of this model through visualization and find it can provide meaningful explanations for pairwise interaction. We believe this model can further improve the drug-target affinity prediction for its higher accuracy and reliable interpretability.

Ningyu Zhang, Mosha Chen, Zhen Bi, Xiaozhuan Liang, Lei Li, Xin Shang, Kangping Yin, Chuanqi Tan, Jian Xu, Fei Huang, Luo Si, Yuan Ni, Guotong Xie, Zhifang Sui, Baobao Chang, Hui Zong, Zheng Yuan, Linfeng Li, Jun Yan, Hongying Zan, Kunli Zhang, Buzhou Tang, Qingcai Chen. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022.

Recognition of glomeruli lesions is the key for diagnosis and treatment planning in kidney pathology; however, the coexisting glomerular structures such as mesangial regions exacerbate the difficulties of this task. In this paper, we introduce a scheme to recognize fine-grained glomeruli lesions from whole slide images. First, a focal instance structural similarity loss is proposed to drive the model to locate all types of glomeruli precisely. Then an Uncertainty Aided Apportionment Network is designed to carry out the fine-grained visual classification without bounding-box annotations. This double branch-shaped structure extracts common features of the child class from the parent class and produces the uncertainty factor for reconstituting the training dataset. Results of slide-wise evaluation illustrate the effectiveness of the entire scheme, with an 8–22% improvement of the mean Average Precision compared with remarkable detection methods. The comprehensive results clearly demonstrate the effectiveness of the proposed method.

We proposed an approach for inpatients’ FWA (fraud, waste, and abuse) detection using multi-view data. In medical insurance, inpatients’ data consisted of daily records of clinical items they consumed and the medical conditions such as gender, age, diagnosis, length of stay, and so on. Even though hospitalized due to the same disease, the clinical daily records varied significantly. To detect the abnormal variations, our method considered to group inpatients in the view of clinical daily records, as well as in the view of their medical conditions, and detect the abnormal variations based on the mismatch of the two clustering results. To accomplish the goal, we accommodated a two-step BERT (Bidirectional Encoder Representations from Transformers) to extract features of the clinical path via the representation learning of the daily consumed items for each admission. The inpatients were then divided into different groups in view of the sequential clinical items’ consumption. We also clustered the inpatients according to their clinical conditions. And the suspicious FWA cases were defined as the ones that had a low probability of cooccurrence and were inconsistent on the properties of two fine-grouping results. We experimented with our approach in multiple insurance claim data with different kinds of surgeries to detect FWA. And 95% of the suspicious cases detected using this approach were confirmed by the medical insurance experts.

To derive expressive representations from high-dimensional and sparse multi-omics samples, there has been existing research attempting to incorporate Gene Regulatory Network (GRN) as a prior knowledge to enhance deep learning models. However, these methods generally just considered homogeneous GRNs with simple structures containing just a single gene/interaction type for single omics data analysis. In this paper, we propose a new framework MoHeG that infuses the abundant knowledge in heterogeneous GRN for representation learning of multi-omics data. Particularly, MoHeG first adopts an interaction-specific graph attention network to represent the graph structure of heterogeneous GRN. Then, a self-supervised learning strategy, which combines an auto-encoder task and a contrastive learning task, is utilized to pre-train the network with unlabeled gene data in both detail and distinctiveness level. Through a fine-tuning approach, we can derive expressive representations for various downstream tasks. Our experiments demonstrate the effectiveness of the new framework on both sufficient and insufficient datasets, compared to a series of state-of-the-art baselines. MoHeG has the potential to become an advanced encoder in analysis pipeline of multi-omics data.

In this work, we propose a novel framework, Gradient Aligned Mutual Learning BERT (GAML-BERT), for improving the early exiting of BERT.GAML-BERT's contributions are two-fold.We conduct a set of pilot experiments, which shows that mutual knowledge distillation between a shallow exit and a deep exit leads to better performances for both.From this observation, we use mutual learning to improve BERT's early exiting performances, that is, we ask each exit of a multi-exit BERT to distill knowledge from each other.Second, we propose GA, a novel training method that aligns the gradients from knowledge distillation to cross-entropy losses.Extensive experiments are conducted on the GLUE benchmark, which shows that our GAML-BERT can significantly outperform the state-of-the-art (SOTA) BERT early exiting methods.

In recent years, with the better availability of medical data such as Electronic Health Records (EHR), more and more data mining models have been developed to explore the data-driven insights for better human health. However, there are many challenges for analyzing EHR such as high-dimensionality, temporality, sparsity, etc., which make the data-driven models less reliable. Medical knowledge graph (MKG), which encodes comprehensive knowledge about the medical concepts and relationships extracted from medical literature, holds great promise to regularize the data-driven models as prior knowledge. Nonetheless, the MKGs are typically not complete, which limits its utility in helping with the data mining process. In this paper, we propose a mutual enhancement framework MendMKG for predictive modeling of clinical events with both EHR and MKG. In particular, MendMKG first conducts a self-supervised learning strategy to simultaneously pre-train a graph attention network for embedding nodes and complete the MKG. It iteratively performs (1) an embedding-based knowledge graph completion module to derive missing edges, (2) and a reconstruction module of unlabeled EHR data to select high-quality ones from these edges, which would be further appended to the MKG to update the embedding model. Through the iterations, the two modules mutually benefit each other. Then, MendMKG uses the pre-trained graph attention network and the updated MKG to generate the visit embeddings to represent patient’s historical visits, and predict the diagnosis in future visit, through a fine-tuning approach. Experimental results on real world EHR corpus are provided to demonstrate the superiority of the proposed framework, compared to a series of state-of-the-art baselines.11The source code and knowledge graph data have been anonymously uploaded to https://github.com/1317375434/MendMKG.

Self-supervised learning has gradually emerged as a powerful technique for graph representation learning. However, transferable, generalizable, and robust representation learning on graph data still remains a challenge for pre-training graph neural networks. In this paper, we propose a simple and effective self-supervised pre-training strategy, named Pairwise Half-graph Discrimination (PHD), that explicitly pre-trains a graph neural network at graph-level. PHD is designed as a simple binary classification task to discriminate whether two half-graphs come from the same source. Experiments demonstrate that the PHD is an effective pre-training strategy that offers comparable or superior performance on 13 graph classification tasks compared with state-of-the-art strategies, and achieves notable improvements when combined with node-level strategies. Moreover, the visualization of learned representation revealed that PHD strategy indeed empowers the model to learn graph-level knowledge like the molecular scaffold. These results have established PHD as a powerful and effective self-supervised learning strategy in graph-level representation learning.

In this paper, we study the problem of leveraging dialogue agents learned from reinforcement learning (RL) that can interact with patients for automatic disease screening. This application requires efficient and effective inquiry of appropriate symptoms to make accurate diagnosis recommendations. Existing studies have tried to use RL to perform both symptom inquiry and diagnosis simultaneously, which needs to deal with a large, heterogeneous action space that affects the learning efficiency and effectiveness. To address the challenge, we propose to leverage the models learned from the dialogue data to customize the settings of the reinforcement learning for more efficient action space exploration. In particular, a supervised diagnosis model is built and involved in the definition of state and reward. We also develop the clustering method to form a hierarchy in the action space. These customizations can make the learning task focus on checking the most relevant symptoms, which effectively boost the confidence of diagnosis. Besides, a novel hierarchical reinforcement learning framework with the pretraining strategy is used to reduce the dimension of action space and help the model to converge. For empirical evaluations, we conduct extensive experiments on both synthetic and real-world datasets. The results have demonstrated the superiority of our approach in diagnostic accuracy and interaction efficiency compared with other baseline methods.

How to produce expressive molecular representations is a fundamental challenge in artificial intelligence-driven drug discovery. Graph neural network (GNN) has emerged as a powerful technique for modeling molecular data. However, previous supervised approaches usually suffer from the scarcity of labeled data and poor generalization capability. Here, we propose a novel molecular pre-training graph-based deep learning framework, named MPG, that learns molecular representations from large-scale unlabeled molecules. In MPG, we proposed a powerful GNN for modelling molecular graph named MolGNet, and designed an effective self-supervised strategy for pre-training the model at both the node and graph-level. After pre-training on 11 million unlabeled molecules, we revealed that MolGNet can capture valuable chemical insights to produce interpretable representation. The pre-trained MolGNet can be fine-tuned with just one additional output layer to create state-of-the-art models for a wide range of drug discovery tasks, including molecular properties prediction, drug-drug interaction and drug-target interaction, on 14 benchmark datasets. The pre-trained MolGNet in MPG has the potential to become an advanced molecular encoder in the drug discovery pipeline.

The iterative design of radiotherapy treatment plans is time-consuming and labor-intensive. In order to provide a guidance to treatment planning, Asymmetric network (A-Net) is proposed to predict the optimal 3D dose distribution for lung cancer patients. A-Net was trained and tested in 392 lung cancer cases with the prescription doses of 50Gy and 60Gy. In A-Net, the encoder and decoder are asymmetric, able to preserve input information and to adapt the limitation of GPU memory. Squeeze and excitation (SE) units are used to improve the data-fitting ability. A loss function involving both the dose distribution and prescription dose as ground truth are designed. In the experiment, A-Net is separately trained and tested in the 50Gy and 60Gy dataset and most of the metrics A-Net achieve similar performance as HD-Unet and 3D-Unet, and some metrics slightly better. In the 50Gy-and-60Gy-combined dataset, most of the A-Net's metrics perform better than the other two. In conclusion, A-Net can accurately predict the IMRT dose distribution in the three datasets of 50Gy and 50Gy-and-60Gy-combined dataset.

Glaucoma is one of the three leading causes of blindness in the world and is predicted to affect around 80 million people by 2020. The optic cup (OC) to optic disc (OD) ratio (CDR) in fundus images plays a pivotal role in the screening and diagnosis of glaucoma. Existing methods usually crop the optic disc region first, and subsequently perform segmentation in this region. However, these approaches come up with high complexities due to the separate operations. To remedy this issue, we propose a Region Focus Network (RF-Net) that innovatively integrates detection and multi-class segmentation into a unified architecture for end-to-end joint optic disc and cup segmentation with global optimization. The key idea of our method is designing a novel multi-class mask branch which generates a high-quality segmentation in the detected region for both disc and cup. To bridge the connection between the backbone and multi-class mask branch, a Fusion Feature Pooling (FFP) structure is presented to extract features from each level of the pyramid network and fuse them into a final feature representation for segmentation. Extensive experimental results on the REFUGE-2018 challenge dataset and the Drishti-GS dataset show that the proposed method achieves the best performance, compared with competitive approaches reported in the literature and the official leaderboard. Our code will be released soon.

When taking care of complex patients with multiple morbidities, accurately predicting the occurrence of each cause-specific event is critical for designing optimal treatment plans. However, standard survival analysis cannot deal with the multiple (usually competing) adverse events and views those competing events as censored. This will result in biased estimation of the incidence rate. In this paper, we propose a deep learning based survival analysis algorithm called DeepComp to jointly predict the progress of the competing events, which can thus inform the doctors which event is more likely to hit the patient first. DeepComp constructs a multi-task recurrent neural network (RNN) and views the conditional probability of each competing event at each time point as the output of each RNN cell. Then the probability chain rule is utilized to combine them together. In this way, the survival probability and the risk for each competing event over the time space are obtained. The multitask structure not only prevents the model from unreasonable censoring but also aids the model in capturing the complex hidden association among the competing events. A novel penalty is added to the loss function to better discriminate the competing risks for each particular patient, which could benefit treatment decision-making. We conduct comprehensive experiments on two real-world clinical data sets and one synthetic data set. The proposed DeepComp method achieves significant performance improvement compared to the state-of-the-art baseline methods.

Phrase mining is a fundamental task for text analysis and has various downstream applications such as named entity recognition, topic modeling, and relation extraction. In this paper, we focus on mining high-quality phrases from domain-specific corpora with special consideration of infrequent ones. Previous methods might miss infrequent high-quality phrases in the candidate selection stage. And these methods rely on explicit features to mine phrases while rarely considering the implicit features. In addition, completeness is rarely explicitly considered in the evaluation of a high-quality phrase. In this paper, we propose a novel approach that exploits a sequence labeling model to capture infrequent phrases. And we employ implicit semantic features and contextual POS tag statistics to measure meaningfulness and completeness, respectively. Experiments over four real-world corpora demonstrate that our method achieves significant improvements over previous state-of-the-art methods across different domains and languages.

In this paper, we integrate span-related information into pre-trained encoder for entity relation extraction task.Instead of using generalpurpose sentence encoder (e.g., existing universal pre-trained models), we introduce a span encoder and a span pair encoder to the pre-training network, which makes it easier to import intra-span and inter-span information into the pre-trained model.To learn the encoders, we devise three customized pretraining objectives from different perspectives, which target on tokens, spans, and span pairs.In particular, a span encoder is trained to recover a random shuffling of tokens in a span, and a span pair encoder is trained to predict positive pairs that are from the same sentences and negative pairs that are from different sentences using contrastive loss.Experimental results show that the proposed pre-training method outperforms distantly supervised pretraining, and achieves promising performance on two entity relation extraction benchmark datasets (ACE05, SciERC).

Pre-trained language models (PLMs), such as BERT, have achieved outstanding performance on multiple natural language processing (NLP) tasks. However, such pre-trained models usually contain a huge number of parameters and are computationally expensive. The high resource demand hinders their application on resource-restricted devices like mobile phones. Knowledge distillation (KD) is an effective compression approach, aiming at encouraging a light-weight student network to imitate the teacher network, and accordingly latent knowledge is transferred from the teacher to student. However, the great majority of student networks in previous KD methods are manually designed, normally a subnetwork of the teacher network. Transformer is generally utilized as the student for compressing BERT but still contains masses of parameters. Motivated by this, we propose a novel approach named NAS-KD, which automatically generates an optimal student network using neural architecture search (NAS) to enhance the distillation for BERT. Experiment on 7 classification tasks in NLP domain demonstrates that NAS-KD can substantially reduce the size of BERT without much performance sacrifice.

Representation learning (RL) plays an important role in extracting proper representations from complex medical data for various analyzing tasks, such as patient grouping, clinical endpoint prediction and medication recommendation. Medical data can be divided into two typical categories, outpatient and inpatient, that have different data characteristics. However, few existing RL methods are specially designed for inpatients data, which have strong temporal relations and consistent diagnosis. In addition, for unordered medical activity set, existing medical RL methods utilize a simple pooling strategy, which would result in indistinguishable contributions among the activities for learning. In this work, we propose Inpatient2Vec, a novel model for learning three kinds of representations for inpatient, including medical activity, hospital day and diagnosis. A multilayer self-attention mechanism with two training tasks is designed to capture the inpatient data characteristics and process the unordered set. Using a real-world dataset, we demonstrate that the proposed approach outperforms the competitive baselines on semantic similarity measurement and clinical events prediction tasks.

With the better availability of healthcare data, such as Electronic Health Records (EHR), more and more data analytics methodologies are developed aiming at digging insights from them to improve the quality of care delivery. There are many challenges on analyzing EHR, such as high dimensionality and event sparsity. Moreover, different from other application domains, the EHR analysis algorithms need to be highly interpretable to make them clinically useful. This makes representation learning from EHRs of key importance. In this paper, we propose an algorithm called Predictive Task Guided Tensor Decomposition (TaGiTeD), to analyze EHRs. Specifically, TaGiTeD learns event interaction patterns that are highly predictive for certain tasks from EHRs with supervised tensor decomposition. Compared with unsupervised methods, TaGiTeD can learn effective EHR representations in a more focused way. This is crucial because most of the medical problems have very limited patient samples, which are not enough for unsupervised algorithms to learn meaningful representations form. We apply TaGiTeD on real world EHR data warehouse and demonstrate that TaGiTeD can learn representations that are both interpretable and predictive.

Anomaly detection in healthcare data like patient records is no trivial task. The anomalies in these datasets are often caused by mismatches between different types of feature, e.g., medications that do not match with the diagnoses. Existing anomaly detection methods do not perform well when detecting "mismatches" between multiple types of feature, especially when the feature space is high-dimensional and sparse. This paper introduces a novel anomaly detection paradigm: Probabilistic-Mismatch Anomaly Detection (PMAD), which detects mismatches between features by modeling a normal instance with a common latent probability distribution that governs the generation of all types of feature. Under this paradigm, the target of anomaly detection is to find instances with dissimilar latent distributions. We further propose Topical PMAD based on an extended Latent Dirichlet Allocation (LDA) model, which is able to capture the latent relationship between features in a high-dimensional space. Experiments on both synthetic data and real-world patient records show that Topical PMAD can effectively detect anomalies with mismatched features, and is highly robust against high-dimensional data as well as inaccurate model selection. The real-world anomalies detected on a patient record dataset show a promising application prospect.

We show that existing mature, relational optimizers can be exploited with a novel schema to give better performance for property graph storage and retrieval than popular noSQL graph stores. The schema combines relational storage for adjacency information with JSON storage for vertex and edge attributes. We demonstrate that this particular schema design has benefits compared to a purely relational or purely JSON solution. The query translation mechanism translates Gremlin queries with no side effects into SQL queries so that one can leverage relational query optimizers. We also conduct an empirical evaluation of our schema design and query translation mechanism with two existing popular property graph stores. We show that our system is 2-8 times better on query performance, and 10-30 times better in throughput on 4.3 billion edge graphs compared to existing stores.

A Care Pathway is a knowledge-centric process to guide clinicians to provide evidence-based care to patients with specific conditions. One existing problem for care pathways is that they often fail to reflect the best clinical practice as a result of not being adequately updated. A better understanding of the gaps between a care pathway and real practice requires aligning patient records with the pathway. Patient records are unlabeled in practice making it difficult to align them with a care pathway which is inherently complex due to its representation as a hierarchical and declarative process model (HDPM). This paper proposes to solve this problem by developing a Hierarchical Markov Random Field (HMRF) method so that a set of patient records can best fit a given care pathway. We validate the effectiveness of the method with experiments on both synthesized data and real clinical data.

Ontology matching, aiming to obtain semantic correspondences between two ontologies, has played a key role in data exchange, data integration and metadata management. Among numerous matching scenarios, especially the applications cross multiple domains, we observe an important problem, denoted as unbalanced ontology matching which requires to find the matches between an ontology describing a local domain knowledge and another ontology covering the information over multiple domains, is not well studied in the community.

With the proliferation of data APIs, it is not uncommon that users who have no clear ideas about data APIs will encounter difficulties to build Mashups to satisfy their requirements. In this paper, we present a semantic-based mashup navigation system, sMash that makes mashup building easy by constructing and visualizing a real-life data API network. We build a sample network by gathering more than 300 popular APIs and find that the relationships between them are so complex that our system will play an important role in navigating users and give them inspiration to build interesting mashups easily. The system is accessible at: http://www.dart.zju.edu.cn/mashup.

We present sMash, a system for facilitating users to mashup Web data. The aspects emphasized by the demo are: (1) how to help novice users master data APIs and relationships amongst them easily; (2) how to inspire various users to build more amazing Web data mashups. First, a real-life data API network is constructed and visualized to enable users to surf and mashup. Second, two kinds of recommendations are generated dynamically based on a comprehensive analysis of the network, user's traces and a repository of mashups to provide navigation.

Augmenting web services with explicit semantics forms the foundation of Service Oriented Architectures (SOAs) automation. As more and more Semantic Web Services (SWSs) are deployed, similar SWSs could have quite different quality-of-service (QoS) levels. The QoS-aware discovery becomes an important challenge. While some efforts try to solve it via Constraint Programming (CP), they suffer from the purely syntactic matchmaking method. Furthermore, the construction of constraints and the selection of services are completely dependent on the literal translation from QoS descriptions, which increase obstacles to actually apply CP. In this paper, we propose a semantic QoS-aware framework for SWSs discovery by combining the semantic matchmaking and CP. Initially, a QoS ontology is presented to define QoS data into service descriptions. Then the ontology reasoning is adopted to change previous syntactic matchmaking into a semantic way. Through confirming the compatibility of concepts, complex QoS conditions are solved as constraints and a selection algorithm is proposed to obtain the optimal offer. Finally, the prototype implementation of our framework is discussed and a SWSs discovery case is used to illustrate the comprehensive discovery process.