arXiv subjects

Yin Zhang

Publications and source records attributed to Yin Zhang.

At least 127 records · Page 7Linked to original sources

arXiv2021

Biomedical named entity recognition using BERT in the machine reading comprehension framework

Recognition of biomedical entities from literature is a challenging research focus, which is the foundation for extracting a large amount of biomedical knowledge existing in unstructured texts into structured formats. Using the sequence labeling framework to implement biomedical named entity recognition (BioNER) is currently a conventional method. This method, however, often cannot take full advantage of the semantic information in the dataset, and the performance is not always satisfactory. In this work, instead of treating the BioNER task as a sequence labeling problem, we formulate it as a machine reading comprehension (MRC) problem. This formulation can introduce more prior knowledge utilizing well-designed queries, and no longer need decoding processes such as conditional random fields (CRF). We conduct experiments on six BioNER datasets, and the experimental results demonstrate the effectiveness of our method. Our method achieves state-of-the-art (SOTA) performance on the BC4CHEMD, BC5CDR-Chem, BC5CDR-Disease, NCBI-Disease, BC2GM and JNLPBA datasets, achieving F1-scores of 92.92%, 94.19%, 87.83%, 90.04%, 85.48% and 78.93%, respectively.

cs.CL↗

arXiv2021

A Model of Two Tales: Dual Transfer Learning Framework for Improved Long-tail Item Recommendation

Highly skewed long-tail item distribution is very common in recommendation systems. It significantly hurts model performance on tail items. To improve tail-item recommendation, we conduct research to transfer knowledge from head items to tail items, leveraging the rich user feedback in head items and the semantic connections between head and tail items. Specifically, we propose a novel dual transfer learning framework that jointly learns the knowledge transfer from both model-level and item-level: 1. The model-level knowledge transfer builds a generic meta-mapping of model parameters from few-shot to many-shot model. It captures the implicit data augmentation on the model-level to improve the representation learning of tail items. 2. The item-level transfer connects head and tail items through item-level features, to ensure a smooth transfer of meta-mapping from head items to tail items. The two types of transfers are incorporated to ensure the learned knowledge from head items can be well applied for tail item representation learning in the long-tail distribution settings. Through extensive experiments on two benchmark datasets, results show that our proposed dual transfer learning framework significantly outperforms other state-of-the-art methods for tail item recommendation in hit ratio and NDCG. It is also very encouraging that our framework further improves head items and overall performance on top of the gains on tail items.

cs.IR↗

arXiv2020

Improving Commonsense Question Answering by Graph-based Iterative Retrieval over Multiple Knowledge Sources

In order to facilitate natural language understanding, the key is to engage commonsense or background knowledge. However, how to engage commonsense effectively in question answering systems is still under exploration in both research academia and industry. In this paper, we propose a novel question-answering method by integrating multiple knowledge sources, i.e. ConceptNet, Wikipedia, and the Cambridge Dictionary, to boost the performance. More concretely, we first introduce a novel graph-based iterative knowledge retrieval module, which iteratively retrieves concepts and entities related to the given question and its choices from multiple knowledge sources. Afterward, we use a pre-trained language model to encode the question, retrieved knowledge and choices, and propose an answer choice-aware attention mechanism to fuse all hidden representations of the previous modules. Finally, the linear classifier for specific tasks is used to predict the answer. Experimental results on the CommonsenseQA dataset show that our method significantly outperforms other competitive methods and achieves the new state-of-the-art. In addition, further ablation studies demonstrate the effectiveness of our graph-based iterative knowledge retrieval module and the answer choice-aware attention module in retrieving and synthesizing background knowledge from multiple knowledge sources.

cs.CL↗

arXiv2020

Federated Unsupervised Representation Learning

To leverage enormous unlabeled data on distributed edge devices, we formulate a new problem in federated learning called Federated Unsupervised Representation Learning (FURL) to learn a common representation model without supervision while preserving data privacy. FURL poses two new challenges: (1) data distribution shift (Non-IID distribution) among clients would make local models focus on different categories, leading to the inconsistency of representation spaces. (2) without the unified information among clients in FURL, the representations across clients would be misaligned. To address these challenges, we propose Federated Constrastive Averaging with dictionary and alignment (FedCA) algorithm. FedCA is composed of two key modules: (1) dictionary module to aggregate the representations of samples from each client and share with all clients for consistency of representation space and (2) alignment module to align the representation of each client on a base model trained on a public data. We adopt the contrastive loss for local model training. Through extensive experiments with three evaluation protocols in IID and Non-IID settings, we demonstrate that FedCA outperforms all baselines with significant margins.

cs.LG↗

arXiv2020

PARADE: A New Dataset for Paraphrase Identification Requiring Computer Science Domain Knowledge

We present a new benchmark dataset called PARADE for paraphrase identification that requires specialized domain knowledge. PARADE contains paraphrases that overlap very little at the lexical and syntactic level but are semantically equivalent based on computer science domain knowledge, as well as non-paraphrases that overlap greatly at the lexical and syntactic level but are not semantically equivalent based on this domain knowledge. Experiments show that both state-of-the-art neural models and non-expert human annotators have poor performance on PARADE. For example, BERT after fine-tuning achieves an F1 score of 0.709, which is much lower than its performance on other paraphrase identification datasets. PARADE can serve as a resource for researchers interested in testing models that incorporate domain knowledge. We make our data and code freely available.

cs.CL↗

arXiv2020

Infusing Disease Knowledge into BERT for Health Question Answering, Medical Inference and Disease Name Recognition

Knowledge of a disease includes information of various aspects of the disease, such as signs and symptoms, diagnosis and treatment. This disease knowledge is critical for many health-related and biomedical tasks, including consumer health question answering, medical language inference and disease name recognition. While pre-trained language models like BERT have shown success in capturing syntactic, semantic, and world knowledge from text, we find they can be further complemented by specific information like knowledge of symptoms, diagnoses, treatments, and other disease aspects. Hence, we integrate BERT with disease knowledge for improving these important tasks. Specifically, we propose a new disease knowledge infusion training procedure and evaluate it on a suite of BERT models including BERT, BioBERT, SciBERT, ClinicalBERT, BlueBERT, and ALBERT. Experiments over the three tasks show that these models can be enhanced in nearly all cases, demonstrating the viability of disease knowledge infusion. For example, accuracy of BioBERT on consumer health question answering is improved from 68.29% to 72.09%, while new SOTA results are observed in two datasets. We make our data and code freely available.

cs.CL↗

arXiv2020

Teacher-Student Framework Enhanced Multi-domain Dialogue Generation

Dialogue systems dealing with multi-domain tasks are highly required. How to record the state remains a key problem in a task-oriented dialogue system. Normally we use human-defined features as dialogue states and apply a state tracker to extract these features. However, the performance of such a system is limited by the error propagation of a state tracker. In this paper, we propose a dialogue generation model that needs no external state trackers and still benefits from human-labeled semantic data. By using a teacher-student framework, several teacher models are firstly trained in their individual domains, learn dialogue policies from labeled states. And then the learned knowledge and experience are merged and transferred to a universal student model, which takes raw utterance as its input. Experiments show that the dialogue system trained under our framework outperforms the one uses a belief tracker.

cs.CL↗

arXiv2020

MTSS: Learn from Multiple Domain Teachers and Become a Multi-domain Dialogue Expert

How to build a high-quality multi-domain dialogue system is a challenging work due to its complicated and entangled dialogue state space among each domain, which seriously limits the quality of dialogue policy, and further affects the generated response. In this paper, we propose a novel method to acquire a satisfying policy and subtly circumvent the knotty dialogue state representation problem in the multi-domain setting. Inspired by real school teaching scenarios, our method is composed of multiple domain-specific teachers and a universal student. Each individual teacher only focuses on one specific domain and learns its corresponding domain knowledge and dialogue policy based on a precisely extracted single domain dialogue state representation. Then, these domain-specific teachers impart their domain knowledge and policies to a universal student model and collectively make this student model a multi-domain dialogue expert. Experiment results show that our method reaches competitive results with SOTAs in both multi-domain and single domain setting.

cs.CL↗

arXiv2020

Chemical-protein Interaction Extraction via Gaussian Probability Distribution and External Biomedical Knowledge

Motivation: The biomedical literature contains a wealth of chemical-protein interactions (CPIs). Automatically extracting CPIs described in biomedical literature is essential for drug discovery, precision medicine, as well as basic biomedical research. Most existing methods focus only on the sentence sequence to identify these CPIs. However, the local structure of sentences and external biomedical knowledge also contain valuable information. Effective use of such information may improve the performance of CPI extraction. Results: In this paper, we propose a novel neural network-based approach to improve CPI extraction. Specifically, the approach first employs BERT to generate high-quality contextual representations of the title sequence, instance sequence, and knowledge sequence. Then, the Gaussian probability distribution is introduced to capture the local structure of the instance. Meanwhile, the attention mechanism is applied to fuse the title information and biomedical knowledge, respectively. Finally, the related representations are concatenated and fed into the softmax function to extract CPIs. We evaluate our proposed model on the CHEMPROT corpus. Our proposed model is superior in performance as compared with other state-of-the-art models. The experimental results show that the Gaussian probability distribution and external knowledge are complementary to each other. Integrating them can effectively improve the CPI extraction performance. Furthermore, the Gaussian probability distribution can effectively improve the extraction performance of sentences with overlapping relations in biomedical relation extraction tasks. Availability: Data and code are available at https://github.com/CongSun-dlut/CPI_extraction. Contact: yangzh@dlut.edu.cn, wangleibihami@gmail.com Supplementary information: Supplementary data are available at Bioinformatics online.

cs.CL↗

arXiv2020

Lifelong Learning with Searchable Extension Units

Lifelong learning remains an open problem. One of its main difficulties is catastrophic forgetting. Many dynamic expansion approaches have been proposed to address this problem, but they all use homogeneous models of predefined structure for all tasks. The common original model and expansion structures ignore the requirement of different model structures on different tasks, which leads to a less compact model for multiple tasks and causes the model size to increase rapidly as the number of tasks increases. Moreover, they can not perform best on all tasks. To solve those problems, in this paper, we propose a new lifelong learning framework named Searchable Extension Units (SEU) by introducing Neural Architecture Search into lifelong learning, which breaks down the need for a predefined original model and searches for specific extension units for different tasks, without compromising the performance of the model on different tasks. Our approach can obtain a much more compact model without catastrophic forgetting. The experimental results on the PMNIST, the split CIFAR10 dataset, the split CIFAR100 dataset, and the Mixture dataset empirically prove that our method can achieve higher accuracy with much smaller model, whose size is about 25-33 percentage of that of the state-of-the-art methods.

cs.LG↗

arXiv2020

Magnetic Gradient: A Natural Driver of Solar Eruptions

It is well-known that there is a gradient, there will drive a flow inevitably. For example, a density-gradient may drive a diffusion flow, an electrical potential-gradient may drive an electric current in plasmas, etc. Then, what will be driven when a magnetic-gradient occurs in solar atmospheric plasmas? Considering the ubiquitous distribution of magnetic-gradient in solar plasma loops, this work demonstrates that magnetic-gradient pumping (MGP) mechanism is valid even in the partial ionized solar photosphere, chromosphere as well as in the corona. It drives energetic particle flows which carry and convey kinetic energy from the underlying atmosphere to move upwards, accumulate around the looptop and increase there temperature and pressure, and finally lead to eruptions around the looptop by triggering ballooning instabilities. This mechanism may explain the formation of the observing hot cusp-structures above flaring loops in most preflare phases, therefore, the magnetic-gradient should be a natural driver of solar eruptions. Furthermore, we may also apply to understand many other astrophysical phenomena, such as the temperature distribution above sunspots, the formation of solar plasma jets, type-II spicule, and fast solar wind above coronal holes, as well as the fast plasma jets related to white dwarfs, neutron stars and black holes.

astro-ph.SR↗

arXiv2019

Consistency-Aware Recommendation for User-Generated ItemList Continuation

User-generated item lists are popular on many platforms. Examples include video-based playlists on YouTube, image-based lists (or"boards") on Pinterest, book-based lists on Goodreads, and answer-based lists on question-answer forums like Zhihu. As users create these lists, a common challenge is in identifying what items to curate next. Some lists are organized around particular genres or topics, while others are seemingly incoherent, reflecting individual preferences for what items belong together. Furthermore, this heterogeneity in item consistency may vary from platform to platform, and from sub-community to sub-community. Hence, this paper proposes a generalizable approach for user-generated item list continuation. Complementary to methods that exploit specific content patterns (e.g., as in song-based playlists that rely on audio features), the proposed approach models the consistency of item lists based on human curation patterns, and so can be deployed across a wide range of varying item types (e.g., videos, images, books). A key contribution is in intelligently combining two preference models via a novel consistency-aware gating network - a general user preference model that captures a user's overall interests, and a current preference priority model that captures a user's current (as of the most recent item) interests. In this way, the proposed consistency-aware recommender can dynamically adapt as user preferences evolve. Evaluation over four datasets(of songs, books, and answers) confirms these observations and demonstrates the effectiveness of the proposed model versus state-of-the-art alternatives. Further, all code and data are available at https://github.com/heyunh2015/ListContinuation_WSDM2020.

cs.IR↗

arXiv2019

Numerical Study of Evaporative Cooling in the Space Station

In this paper, we numerically studied the effects of mechanical vibration and magnetic fields on evaporative cooling process carried in space station by direct simulation Monte Carlo method. Simulated with the vibration data of international space station, we found that the cooling process would suffer great atomic losses until the accelerations reduced tenfold at least. In addition, if we enlarge the s-wave scattering length five times by Feshbach resonance, the PSD increased to 50 compared to 3 of no magnetic fields situation after 5 seconds evaporative cooling. We also simulated the two stages crossed beam evaporative cooling process (TSCBC) under both physical impacts and obtain $4\times10^5$ $^{85}$Rb atoms with a temperature of 8 pK. These results are of significance to the cold atom experiments carried out on space station in the future.

cond-mat.quant-gas↗

arXiv2019

Big-Data Clustering: K-Means or K-Indicators?

The K-means algorithm is arguably the most popular data clustering method, commonly applied to processed datasets in some "feature spaces", as is in spectral clustering. Highly sensitive to initializations, however, K-means encounters a scalability bottleneck with respect to the number of clusters K as this number grows in big data applications. In this work, we promote a closely related model called K-indicators model and construct an efficient, semi-convex-relaxation algorithm that requires no randomized initializations. We present extensive empirical results to show advantages of the new algorithm when K is large. In particular, using the new algorithm to start the K-means algorithm, without any replication, can significantly outperform the standard K-means with a large number of currently state-of-the-art random replications.

cs.LG↗

arXiv2018

Anatomy of electrical signals and dc-voltage lineshape in spin torque ferromagnetic resonance

The electrical detection of spin torque ferromagnetic resonance (st-FMR) is becoming a popular method for measuring the spin-Hall angle of heavy metals (HM). However, various sensible analysis on the same material with either the same or different experimental setups yielded different spin-Hall angles with large discrepancy, indicating some missing ingredients in our current understanding of st-FMR. Here we carry out a careful analysis of electrical signals of the st-FMR in a HM/ferromagnet (HM/FM) bilayer with an arbitrary magnetic anisotropy. The FM magnetization is driven by two radio-frequency (rf) forces: the rf Oersted field generated by an applied rf electric current and the so called rf spin-orbit torque from the spin current flowing perpendicularly from the HM to the FM due to the spin-Hall effect. By using the universal form of the dynamic susceptibility matrix of magnetic materials at the st-FMR, the electrical signals originated from the anisotropic magnetoresistance, anomalous Hall effect and inverse spin-Hall effect are analysed and dc-voltage lineshape near the st-FMR are obtained. Angle-dependence of dc-voltage is given for two setups. A way of experimentally extracting the spin-Hall angle of a HM is proposed.

cond-mat.mes-hall↗

arXiv2018

Fairness-Aware Recommendation of Information Curators

This paper highlights our ongoing efforts to create effective information curator recommendation models that can be personalized for individual users, while maintaining important fairness properties. Concretely, we introduce the problem of information curator recommendation, provide a high-level overview of a fairness-aware recommender, and introduce some preliminary experimental evidence over a real-world Twitter dataset. We conclude with some thoughts on future directions.

cs.IR↗

arXiv2018

Breaking the current density threshold in spin-orbit-torque magnetic random access memory

Spin-orbit-torque magnetic random access memory (SOT-MRAM) is a promising technology for the next generation of data storage devices. The main bottleneck of this technology is the high reversal current density threshold. This outstanding problem of SOT-MRAM is now solved by using a current density of constant magnitude and varying flow direction that reduces the reversal current density threshold by a factor of more than the Gilbert damping coefficient. The Euler-Lagrange equation for the fastest magnetization reversal path and the optimal current pulse are derived for an arbitrary magnetic cell. The theoretical limit of minimal reversal current density and current density for a GHz switching rate of the new reversal strategy for CoFeB/Ta SOT-MRAMs are respectively of the order of $10^5$ A/cm$^2$ and $10^6$ A/cm$^2$ far below $10^7$ A/cm$^2$ and $10^8$ A/cm$^2$ in the conventional strategy. Furthermore, no external magnetic field is needed for a deterministic reversal in the new strategy.

cond-mat.mes-hall↗

arXiv2016

Very Long-period Pulsations before the Onset of Solar Flares

Solar flares are the most powerful explosions occurring in the solar system, which may lead to disastrous space weather events and impact various aspects of our Earth. So far, it is still a big challenge in modern astrophysics to understand the origin of solar flares and predict their onset. Based on the analysis of soft X-ray emission observed by the Geostationary Operational Environmental Satellite (GOES), this work reported a new discovery of very long-periodic pulsations occurred in the preflare phase before the onset of solar flares (preflare-VLPs). These pulsations are typically with period of 8 - 30 min and last for about 1 - 2 hours. They are possibly generated from LRC oscillations of plasma loops where electric current dominates the physical process during magnetic energy accumulation in the source region. The preflare-VLP provides an essential information for understanding the triggering mechanism and origin of solar flares, and may help us to response to solar explosions and the corresponding disastrous space weather events as a convenient precursory indicator.

astro-ph.SR↗