arXiv subjects

Qi Zhang

Publications and source records attributed to Qi Zhang.

At least 973 records · Page 54Linked to original sources

arXiv2021

MeDiaQA: A Question Answering Dataset on Medical Dialogues

In this paper, we introduce MeDiaQA, a novel question answering(QA) dataset, which constructed on real online Medical Dialogues. It contains 22k multiple-choice questions annotated by human for over 11k dialogues with 120k utterances between patients and doctors, covering 150 specialties of diseases, which are collected from haodf.com and dxy.com. MeDiaQA is the first QA dataset where reasoning over medical dialogues, especially their quantitative contents. The dataset has the potential to test the computing, reasoning and understanding ability of models across multi-turn dialogues, which is challenging compared with the existing datasets. To address the challenges, we design MeDia-BERT, and it achieves 64.3% accuracy, while human performance of 93% accuracy, which indicates that there still remains a large room for improvement.

cs.CL↗

arXiv2021

Sleep Staging Based on Multi Scale Dual Attention Network

Sleep staging plays an important role on the diagnosis of sleep disorders. In general, experts classify sleep stages manually based on polysomnography (PSG), which is quite time-consuming. Meanwhile, the acquisition process of multiple signals is much complex, which can affect the subject's sleep. Therefore, the use of single-channel electroencephalogram (EEG) for automatic sleep staging has become a popular research topic. In the literature, a large number of sleep staging methods based on single-channel EEG have been proposed with promising results and achieve the preliminary automation of sleep staging. However, the performance for most of these methods in the N1 stage do not satisfy the needs of the diagnosis. In this paper, we propose a deep learning model multi scale dual attention network(MSDAN) based on raw EEG, which utilizes multi-scale convolution to extract features in different waveforms contained in the EEG signal, connects channel attention and spatial attention mechanisms in series to filter and highlight key information, and uses soft thresholding to remove redundant information. Experiments were conducted using two datasets with 5-fold cross-validation and hold-out validation method. The final average accuracy, overall accuracy, macro F1 score and Cohen's Kappa coefficient of the model reach 96.70%, 91.74%, 0.8231 and 0.8723 on the Sleep-EDF dataset, 96.14%, 90.35%, 0.7945 and 0.8284 on the Sleep-EDFx dataset. Significantly, our model performed superiorly in the N1 stage, with F1 scores of 54.41% and 52.79% on the two datasets respectively. The results show the superiority of our network over the existing methods, reaching a new state-of-the-art. In particular, the proposed method achieves excellent results in the N1 sleep stage compared to other methods.

cs.LG↗

arXiv2021

Ensemble nonequivalence and Bose-Einstein condensation in weighted networks

The asymptotic (non)equivalence of canonical and microcanonical ensembles, describing systems with soft and hard constraints respectively, is a central concept in statistical physics. Traditionally, the breakdown of ensemble equivalence (EE) has been associated with nonvanishing relative canonical fluctuations of the constraints in the thermodynamic limit. Recently, it has been reformulated in terms of a nonvanishing relative entropy density between microcanonical and canonical probabilities. The earliest observations of EE violation required phase transitions or long-range interactions. More recent research on binary networks found that an extensive number of local constraints can also break EE, even in absence of phase transitions. Here we study for the first time ensemble nonequivalence in weighted networks with local constraints. Unlike their binary counterparts, these networks can undergo a form of Bose-Einstein condensation (BEC) producing a core-periphery structure where a finite fraction of the link weights concentrates in the core. This phenomenon creates a unique setting where local constraints coexist with a phase transition. We find surviving relative fluctuations only in the condensed phase, as in more traditional BEC settings. However, we also find a non-vanishing relative entropy density for all temperatures, signalling a breakdown of EE due to the presence of an extensive number of constraints, irrespective of BEC. Therefore, in presence of extensively many local constraints, vanishing relative fluctuations no longer guarantee EE.

cond-mat.stat-mech↗

arXiv2021

Quotients of probabilistic Boolean networks

A probabilistic Boolean network (PBN) is a discrete-time system composed of a collection of Boolean networks between which the PBN switches in a stochastic manner. This paper focuses on the study of quotients of PBNs. Given a PBN and an equivalence relation on its state set, we consider a probabilistic transition system that is generated by the PBN; the resulting quotient transition system then automatically captures the quotient behavior of this PBN. We therefore describe a method for obtaining a probabilistic Boolean system that generates the transitions of the quotient transition system. Applications of this quotient description are discussed, and it is shown that for PBNs, controller synthesis can be performed easily by first controlling a quotient system and then lifting the control law back to the original network. A biological example is given to show the usefulness of the developed results.

math.OC↗

arXiv2021

BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstruction

We study the challenging task of neural network quantization without end-to-end retraining, called Post-training Quantization (PTQ). PTQ usually requires a small subset of training data but produces less powerful quantized models than Quantization-Aware Training (QAT). In this work, we propose a novel PTQ framework, dubbed BRECQ, which pushes the limits of bitwidth in PTQ down to INT2 for the first time. BRECQ leverages the basic building blocks in neural networks and reconstructs them one-by-one. In a comprehensive theoretical study of the second-order error, we show that BRECQ achieves a good balance between cross-layer dependency and generalization error. To further employ the power of quantization, the mixed precision technique is incorporated in our framework by approximating the inter-layer and intra-layer sensitivity. Extensive experiments on various handcrafted and searched neural architectures are conducted for both image classification and object detection tasks. And for the first time we prove that, without bells and whistles, PTQ can attain 4-bit ResNet and MobileNetV2 comparable with QAT and enjoy 240 times faster production of quantized models. Codes are available at https://github.com/yhhhli/BRECQ.

cs.LG↗

arXiv2021

AbdomenCT-1K: Is Abdominal Organ Segmentation A Solved Problem?

With the unprecedented developments in deep learning, automatic segmentation of main abdominal organs seems to be a solved problem as state-of-the-art (SOTA) methods have achieved comparable results with inter-rater variability on many benchmark datasets. However, most of the existing abdominal datasets only contain single-center, single-phase, single-vendor, or single-disease cases, and it is unclear whether the excellent performance can generalize on diverse datasets. This paper presents a large and diverse abdominal CT organ segmentation dataset, termed AbdomenCT-1K, with more than 1000 (1K) CT scans from 12 medical centers, including multi-phase, multi-vendor, and multi-disease cases. Furthermore, we conduct a large-scale study for liver, kidney, spleen, and pancreas segmentation and reveal the unsolved segmentation problems of the SOTA methods, such as the limited generalization ability on distinct medical centers, phases, and unseen diseases. To advance the unsolved problems, we further build four organ segmentation benchmarks for fully supervised, semi-supervised, weakly supervised, and continual learning, which are currently challenging and active research topics. Accordingly, we develop a simple and effective method for each benchmark, which can be used as out-of-the-box methods and strong baselines. We believe the AbdomenCT-1K dataset will promote future in-depth research towards clinical applicable abdominal organ segmentation methods. The datasets, codes, and trained models are publicly available at https://github.com/JunMa11/AbdomenCT-1K.

cs.CV↗

arXiv2021

Efficient Top-k Ego-Betweenness Search

Betweenness centrality, measured by the number of times a vertex occurs on all shortest paths of a graph, has been recognized as a key indicator for the importance of a vertex in the network. However, the betweenness of a vertex is often very hard to compute because it needs to explore all the shortest paths between the other vertices. Recently, a relaxed concept called ego-betweenness was introduced which focuses on computing the betweenness of a vertex in its ego network. In this work, we study a problem of finding the top-k vertices with the highest ego-betweennesses. We first develop two novel search algorithms equipped with a basic upper bound and a dynamic upper bound to efficiently solve this problem. Then, we propose local-update and lazy-update solutions to maintain the ego-betweennesses for all vertices and the top-k results when the graph is updated, respectively. In addition, we also present two efficient parallel algorithms to further improve the efficiency. The results of extensive experiments on five large real-life datasets demonstrate the efficiency, scalability, and effectiveness of our algorithms.

cs.DB↗

arXiv2021

Intrinsic nature of chiral charge order in the kagome superconductor RbV3Sb5

Superconductors with kagome lattices have been identified for over 40 years, with a superconducting transition temperature TC up to 7K. Recently, certain kagome superconductors have been found to exhibit an exotic charge order, which intertwines with superconductivity and persists to a temperature being one order of magnitude higher than TC. In this work, we use scanning tunneling microscopy (STM) to study the charge order in kagome superconductor RbV3Sb5. We observe both a 2x2 chiral charge order and nematic surface superlattices (predominantly 1x4). We find that the 2x2 charge order exhibits intrinsic chirality with magnetic field tunability. Defects can scatter electrons to introduce standing waves, which couple with the charge order to cause extrinsic effects. While the chiral charge order resembles that discovered in KV3Sb5, it further interacts with the nematic surface superlattices that are absent in KV3Sb5 but exist in CsV3Sb5.

cond-mat.str-el↗

arXiv2021

Strong ensemble nonequivalence in systems with local constraints

The asymptotic equivalence of canonical and microcanonical ensembles is a central concept in statistical physics, with important consequences for both theoretical research and practical applications. However, this property breaks down under certain circumstances. The most studied violation of ensemble equivalence requires phase transitions, in which case it has a `restricted' (i.e. confined to a certain region in parameter space) but `strong' (i.e. characterized by a difference between the entropies of the two ensembles that is of the same order as the entropies themselves) form. However, recent research on networks has shown that the presence of an extensive number of local constraints can lead to ensemble nonequivalence even in the absence of phase transitions. This occurs in a `weak' (i.e. leading to a subleading entropy difference) but remarkably `unrestricted' (i.e. valid in the entire parameter space) form. Here we look for more general manifestations of ensemble nonequivalence in arbitrary ensembles of matrices with given margins. These models have widespread applications in the study of spatially heterogeneous and/or temporally nonstationary systems, with consequences for the analysis of multivariate financial and neural time-series, multi-platform social activity, gene expression profiles and other Big Data. We confirm that ensemble nonequivalence appears in `unrestricted' form throughout the entire parameter space due to the extensivity of local constraints. Surprisingly, at the same time it can also exhibit the `strong' form. This novel, simultaneously `strong and unrestricted' form of nonequivalence is very robust and imposes a principled choice of the ensemble. We calculate the proper mathematical quantities to be used in real-world applications.

cond-mat.stat-mech↗

arXiv2021

Giant enhancement of anomalous Hall effect in Cr modulation-doped non-collinear antiferromagnetic Mn3Sn thin films

We report on Cr doping effect in Mn3Sn polycrystalline films with both uniform and modulation doping. It is found that Cr doping with low concentration does not cause notable changes to the structural and magnetic properties of Mn3Sn, but it significantly enhances the anomalous Hall conductivity, particularly for modulation-doped samples at low temperature. A Hall conductivity as high as 184.8 Ω-1 cm-1 is obtained for modulation-doped samples at 50 K, in a sharp contrast to vanishingly small values for undoped samples at the same temperature. We attribute the enhancement to the change of Fermi level induced by Cr doping

cond-mat.mtrl-sci↗

arXiv2021

Unsupervised Summarization for Chat Logs with Topic-Oriented Ranking and Context-Aware Auto-Encoders

Automatic chat summarization can help people quickly grasp important information from numerous chat messages. Unlike conventional documents, chat logs usually have fragmented and evolving topics. In addition, these logs contain a quantity of elliptical and interrogative sentences, which make the chat summarization highly context dependent. In this work, we propose a novel unsupervised framework called RankAE to perform chat summarization without employing manually labeled data. RankAE consists of a topic-oriented ranking strategy that selects topic utterances according to centrality and diversity simultaneously, as well as a denoising auto-encoder that is carefully designed to generate succinct but context-informative summaries based on the selected utterances. To evaluate the proposed method, we collect a large-scale dataset of chat logs from a customer service environment and build an annotated set only for model evaluation. Experimental results show that RankAE significantly outperforms other unsupervised methods and is able to generate high-quality summaries in terms of relevance and topic coverage.

cs.CL↗

arXiv2021

Topic-Oriented Spoken Dialogue Summarization for Customer Service with Saliency-Aware Topic Modeling

In a customer service system, dialogue summarization can boost service efficiency by automatically creating summaries for long spoken dialogues in which customers and agents try to address issues about specific topics. In this work, we focus on topic-oriented dialogue summarization, which generates highly abstractive summaries that preserve the main ideas from dialogues. In spoken dialogues, abundant dialogue noise and common semantics could obscure the underlying informative content, making the general topic modeling approaches difficult to apply. In addition, for customer service, role-specific information matters and is an indispensable part of a summary. To effectively perform topic modeling on dialogues and capture multi-role information, in this work we propose a novel topic-augmented two-stage dialogue summarizer (TDS) jointly with a saliency-aware neural topic model (SATM) for topic-oriented summarization of customer service dialogues. Comprehensive studies on a real-world Chinese customer service dataset demonstrated the superiority of our method against several strong baselines.

cs.CL↗

arXiv2021

Knowledge Infused Policy Gradients with Upper Confidence Bound for Relational Bandits

Contextual Bandits find important use cases in various real-life scenarios such as online advertising, recommendation systems, healthcare, etc. However, most of the algorithms use flat feature vectors to represent context whereas, in the real world, there is a varying number of objects and relations among them to model in the context. For example, in a music recommendation system, the user context contains what music they listen to, which artists create this music, the artist albums, etc. Adding richer relational context representations also introduces a much larger context space making exploration-exploitation harder. To improve the efficiency of exploration-exploitation knowledge about the context can be infused to guide the exploration-exploitation strategy. Relational context representations allow a natural way for humans to specify knowledge owing to their descriptive nature. We propose an adaptation of Knowledge Infused Policy Gradients to the Contextual Bandit setting and a novel Knowledge Infused Policy Gradients Upper Confidence Bound algorithm and perform an experimental analysis of a simulated music recommendation dataset and various real-life datasets where expert knowledge can drastically reduce the total regret and where it cannot.

cs.LG↗

arXiv2021

SENT: Sentence-level Distant Relation Extraction via Negative Training

Distant supervision for relation extraction provides uniform bag labels for each sentence inside the bag, while accurate sentence labels are important for downstream applications that need the exact relation type. Directly using bag labels for sentence-level training will introduce much noise, thus severely degrading performance. In this work, we propose the use of negative training (NT), in which a model is trained using complementary labels regarding that ``the instance does not belong to these complementary labels". Since the probability of selecting a true label as a complementary label is low, NT provides less noisy information. Furthermore, the model trained with NT is able to separate the noisy data from the training data. Based on NT, we propose a sentence-level framework, SENT, for distant relation extraction. SENT not only filters the noisy data to construct a cleaner dataset, but also performs a re-labeling process to transform the noisy data into useful training data, thus further benefiting the model's performance. Experimental results show the significant improvement of the proposed method over previous methods on sentence-level evaluation and de-noise effect.

cs.CL↗

arXiv2021

Connection between forest fire emission and COVID-19 incidents in West Coast regions of the United States

Forest fires impact on soil, water and biota resources has been widely researched. Although forest fires profoundly impact the atmosphere and air quality across the ecosystems, much less research has been developed to examine its impact on the current pandemic. In-situ air pollution data were utilized to examine the effects of the 2020 forest fire on atmosphere and coronavirus (COVID 19) casualties. The spatiotemporal concentrations of particulate matter (PM2.5 and PM10) and Nitrogen Dioxide (NO2) were collected from August 1 to October 30 for 2020 (fire year) and 2019 (reference year). Both spatial (Multiscale Geographically Weighted Regression) and non spatial (negative binomial regression) regression analysis was performed to assess the adverse effects of fire emission on human health. The in situ data led measurements showed that the maximum increases in PM2.5, PM10, and NO2 concentrations were clustered in the West Coastal fire-prone states during the August 1 to October 30 period. The average concentration of particulate matter (PM2.5 and PM10) and NO2 were increased in all the fire states affected badly by forest fires. The average PM2.5 concentration over the period was recorded as 7.9, 6.3, 5.5, and 5.2 for California, Colorado, Oregon, and Washington in 2019, which was increased up to 24.9, 13.4, 25, and 17 in 2020. Both spatial and non-spatial regression models exhibited a statistically significant association between fire emission and COVID 19 incidents. A total of 30 models were developed for analyzing the spatial non-stationary and local association between the predictor and response factors. All these spatial models have demonstrated a statistically significant association between fire emissions and COVID counts. More thorough research is needed to better understand the complex association between forest fire and human health.

stat.AP↗

arXiv2021

One2Set: Generating Diverse Keyphrases as a Set

Recently, the sequence-to-sequence models have made remarkable progress on the task of keyphrase generation (KG) by concatenating multiple keyphrases in a predefined order as a target sequence during training. However, the keyphrases are inherently an unordered set rather than an ordered sequence. Imposing a predefined order will introduce wrong bias during training, which can highly penalize shifts in the order between keyphrases. In this work, we propose a new training paradigm One2Set without predefining an order to concatenate the keyphrases. To fit this paradigm, we propose a novel model that utilizes a fixed set of learned control codes as conditions to generate a set of keyphrases in parallel. To solve the problem that there is no correspondence between each prediction and target during training, we propose a $K$-step target assignment mechanism via bipartite matching, which greatly increases the diversity and reduces the duplication ratio of generated keyphrases. The experimental results on multiple benchmarks demonstrate that our approach significantly outperforms the state-of-the-art methods.

cs.CL↗

arXiv2021

TextFlint: Unified Multilingual Robustness Evaluation Toolkit for Natural Language Processing

Various robustness evaluation methodologies from different perspectives have been proposed for different natural language processing (NLP) tasks. These methods have often focused on either universal or task-specific generalization capabilities. In this work, we propose a multilingual robustness evaluation platform for NLP tasks (TextFlint) that incorporates universal text transformation, task-specific transformation, adversarial attack, subpopulation, and their combinations to provide comprehensive robustness analysis. TextFlint enables practitioners to automatically evaluate their models from all aspects or to customize their evaluations as desired with just a few lines of code. To guarantee user acceptability, all the text transformations are linguistically based, and we provide a human evaluation for each one. TextFlint generates complete analytical reports as well as targeted augmented data to address the shortcomings of the model's robustness. To validate TextFlint's utility, we performed large-scale empirical evaluations (over 67,000 evaluations) on state-of-the-art deep learning models, classic supervised methods, and real-world systems. Almost all models showed significant performance degradation, including a decline of more than 50% of BERT's prediction accuracy on tasks such as aspect-level sentiment classification, named entity recognition, and natural language inference. Therefore, we call for the robustness to be included in the model evaluation, so as to promote the healthy development of NLP technology.

cs.CL↗

arXiv2021

Discovery of unconventional chiral charge order in kagome superconductor KV3Sb5

Intertwining quantum order and nontrivial topology is at the frontier of condensed matter physics. A charge density wave (CDW) like order with orbital currents has been proposed as a powerful resource for achieving the quantum anomalous Hall effect in topological materials and for the hidden phase in cuprate high-temperature superconductors. However, the experimental realization of such an order is challenging. Here we use high-resolution scanning tunnelling microscopy (STM) to discover an unconventional charge order in a kagome material KV3Sb5, with both a topological band structure and a superconducting ground state. Through both topography and spectroscopic imaging, we observe a robust 2x2 superlattice. Spectroscopically, an energy gap opens at the Fermi level, across which the 2x2 charge modulation exhibits an intensity reversal in real-space, signaling charge ordering. At impurity-pinning free region, the strength of intrinsic charge modulations further exhibits chiral anisotropy with unusual magnetic field response. Theoretical analysis of our experiments suggests a tantalizing unconventional chiral CDW in the frustrated kagome lattice, which can not only lead to large anomalous Hall effect with orbital magnetism, but also be a precursor of unconventional superconductivity.

cond-mat.supr-con↗