arXiv subjects

Qi Jia

Publications and source records attributed to Qi Jia.

82 records · Page 5Linked to original sources

arXiv2022

Post-Training Dialogue Summarization using Pseudo-Paraphrasing

Previous dialogue summarization techniques adapt large language models pretrained on the narrative text by injecting dialogue-specific features into the models. These features either require additional knowledge to recognize or make the resulting models harder to tune. To bridge the format gap between dialogues and narrative summaries in dialogue summarization tasks, we propose to post-train pretrained language models (PLMs) to rephrase from dialogue to narratives. After that, the model is fine-tuned for dialogue summarization as usual. Comprehensive experiments show that our approach significantly improves vanilla PLMs on dialogue summarization and outperforms other SOTA models by the summary quality and implementation costs.

cs.CL↗

arXiv2022

Learning Weighting Map for Bit-Depth Expansion within a Rational Range

Bit-depth expansion (BDE) is one of the emerging technologies to display high bit-depth (HBD) image from low bit-depth (LBD) source. Existing BDE methods have no unified solution for various BDE situations, and directly learn a mapping for each pixel from LBD image to the desired value in HBD image, which may change the given high-order bits and lead to a huge deviation from the ground truth. In this paper, we design a bit restoration network (BRNet) to learn a weight for each pixel, which indicates the ratio of the replenished value within a rational range, invoking an accurate solution without modifying the given high-order bit information. To make the network adaptive for any bit-depth degradation, we investigate the issue in an optimization perspective and train the network under progressive training strategy for better performance. Moreover, we employ Wasserstein distance as a visual quality indicator to evaluate the difference of color distribution between restored image and the ground truth. Experimental results show our method can restore colorful images with fewer artifacts and false contours, and outperforms state-of-the-art methods with higher PSNR/SSIM results and lower Wasserstein distance. The source code will be made available at https://github.com/yuqing-liu-dut/bit-depth-expansion

cs.CV↗

arXiv2022

Irrational self-similar sets

Let $K\subset\mathbb{R}$ be a self-similar set defined on $\mathbb{R}$. It is easy to prove that if the Lebesgue measure of $K$ is zero, then for Lebesgue almost every $t$, $$K+t=\{x+t:x\in K\}$$ only consists of irrational or transcendental numbers. In this note, we shall consider some classes of self-similar sets, and explicitly construct such $t$'s. Our main idea is from the $q$-expansions.

math.NT↗

arXiv2022

Cross-SRN: Structure-Preserving Super-Resolution Network with Cross Convolution

It is challenging to restore low-resolution (LR) images to super-resolution (SR) images with correct and clear details. Existing deep learning works almost neglect the inherent structural information of images, which acts as an important role for visual perception of SR results. In this paper, we design a hierarchical feature exploitation network to probe and preserve structural information in a multi-scale feature fusion manner. First, we propose a cross convolution upon traditional edge detectors to localize and represent edge features. Then, cross convolution blocks (CCBs) are designed with feature normalization and channel attention to consider the inherent correlations of features. Finally, we leverage multi-scale feature fusion group (MFFG) to embed the cross convolution blocks and develop the relations of structural features in different scales hierarchically, invoking a lightweight structure-preserving network named as Cross-SRN. Experimental results demonstrate the Cross-SRN achieves competitive or superior restoration performances against the state-of-the-art methods with accurate and clear structural details. Moreover, we set a criterion to select images with rich structural textures. The proposed Cross-SRN outperforms the state-of-the-art methods on the selected benchmark, which demonstrates that our network has a significant advantage in preserving edges.

eess.IV↗

arXiv2020

DDRel: A New Dataset for Interpersonal Relation Classification in Dyadic Dialogues

Interpersonal language style shifting in dialogues is an interesting and almost instinctive ability of human. Understanding interpersonal relationship from language content is also a crucial step toward further understanding dialogues. Previous work mainly focuses on relation extraction between named entities in texts. In this paper, we propose the task of relation classification of interlocutors based on their dialogues. We crawled movie scripts from IMSDb, and annotated the relation labels for each session according to 13 pre-defined relationships. The annotated dataset DDRel consists of 6300 dyadic dialogue sessions between 694 pair of speakers with 53,126 utterances in total. We also construct session-level and pair-level relation classification tasks with widely-accepted baselines. The experimental results show that this task is challenging for existing models and the dataset will be useful for future research.

cs.CL↗

arXiv2020

Matching Questions and Answers in Dialogues from Online Forums

Matching question-answer relations between two turns in conversations is not only the first step in analyzing dialogue structures, but also valuable for training dialogue systems. This paper presents a QA matching model considering both distance information and dialogue history by two simultaneous attention mechanisms called mutual attention. Given scores computed by the trained model between each non-question turn with its candidate questions, a greedy matching strategy is used for final predictions. Because existing dialogue datasets such as the Ubuntu dataset are not suitable for the QA matching task, we further create a dataset with 1,000 labeled dialogues and demonstrate that our proposed model outperforms the state-of-the-art and other strong baselines, particularly for matching long-distance QA pairs.

cs.CL↗

arXiv2018

Enhancing intrinsic detection efficiency of superconducting nanowire single-photon detectors via helium ion irradiation

Realizing an NbN superconducting nanowire single-photon detector (SNSPD) with a 100% intrinsic detection efficiency (IDE) at the near-infrared wavelengths is still challenging. Herein, we developed a post-processing method to increase the IDE of NbN SNSPDs to near unity using a 20 keV helium ion irradiation. The IDE enhancement was achieved owing to the ion-induced reduction of the superconducting energy gap and the electron density of states at the Fermi level, determined with the electrical and magnetic transport measurements. The change in optical absorptance of the irradiated SNSPD was negligible as confirmed by the measured optical reflectance and system detection efficiency (SDE). Benefited with the IDE enhancement, the SDE of an irradiated device was significantly increased from 49% to 92% at 2.2 K for a 1550 nm wavelength.

physics.app-ph↗

arXiv2017

Sequential Dual Deep Learning with Shape and Texture Features for Sketch Recognition

Recognizing freehand sketches with high arbitrariness is greatly challenging. Most existing methods either ignore the geometric characteristics or treat sketches as handwritten characters with fixed structural ordering. Consequently, they can hardly yield high recognition performance even though sophisticated learning techniques are employed. In this paper, we propose a sequential deep learning strategy that combines both shape and texture features. A coded shape descriptor is exploited to characterize the geometry of sketch strokes with high flexibility, while the outputs of constitutional neural networks (CNN) are taken as the abstract texture feature. We develop dual deep networks with memorable gated recurrent units (GRUs), and sequentially feed these two types of features into the dual networks, respectively. These dual networks enable the feature fusion by another gated recurrent unit (GRU), and thus accurately recognize sketches invariant to stroke ordering. The experiments on the TU-Berlin data set show that our method outperforms the average of human and state-of-the-art algorithms even when significant shape and appearance variations occur.

cs.CV↗

arXiv2017

Angstrom-resolution single-molecule fluorescence resonance energy transfer reveals mechanisms of DNA helicases

Single-molecule FRET is widely used to study helicases by detecting distance changes between a fluorescent donor and an acceptor anchored to overhangs of a forked DNA duplex. However, it has lacked single-base pair (1-bp) resolution required for revealing stepping dynamics in unwinding because FRET signals are usually blurred by thermal fluctuations of the overhangs. We designed a nanotensioner in which a short DNA is bent to exert a force on the overhangs, just as in optical/magnetic tweezers. The strategy improved the resolution of FRET to 0.5 bp, high enough to uncover the differences in DNA unwinding by yeast Pif1 and E. coli RecQ whose unwinding behaviors cannot be differentiated by currently practiced methods. We found that Pif1 exhibits 1-bp-stepping kinetics, while RecQ breaks 1 bp at a time but questers the nascent nucleotides and releases them randomly. The high-resolution data allowed us to propose a three-parameter model to quantitatively interpret the apparently different unwinding behaviors of the two helicases which belong to two superfamilies.

physics.bio-ph↗

arXiv2016

A Fast Ellipse Detector Using Projective Invariant Pruning

Detecting elliptical objects from an image is a central task in robot navigation and industrial diagnosis where the detection time is always a critical issue. Existing methods are hardly applicable to these real-time scenarios of limited hardware resource due to the huge number of fragment candidates (edges or arcs) for fitting ellipse equations. In this paper, we present a fast algorithm detecting ellipses with high accuracy. The algorithm leverage a newly developed projective invariant to significantly prune the undesired candidates and to pick out elliptical ones. The invariant is able to reflect the intrinsic geometry of a planar curve, giving the value of -1 on any three collinear points and +1 for any six points on an ellipse. Thus, we apply the pruning and picking by simply comparing these binary values. Moreover, the calculation of the invariant only involves the determinant of a 3*3 matrix. Extensive experiments on three challenging data sets with 650 images demonstrate that our detector runs 20%-50% faster than the state-of-the-art algorithms with the comparable or higher precision.

cs.CV↗