Search

Existing research on news summarization primarily focuses on single-language single-document (SLSD), single-language multi-document (SLMD) or cross-language single-document (CLSD). However, in real-world scenarios, news about a international event often involves multiple documents in different languages, i.e., mixed-language multi-document (MLMD). Therefore, summarizing MLMD news is of great significance. However, the lack of datasets for MLMD news summarization has constrained the development of research in this area. To fill this gap, we construct a mixed-language multi-document news summarization dataset (MLMD-news), which contains four different languages and 10,992 source document cluster and target summary pairs. Additionally, we propose a graph-based extract-generate model and benchmark various methods on the MLMD-news dataset and publicly release our dataset and code\footnote[1]{https://github.com/Southnf9/MLMD-news}, aiming to advance research in summarization within MLMD scenarios.

Since the advent of ChatGPT, Large Language Models (LLMs) have excelled in various tasks but remain as black-box systems. Consequently, the reasoning bottlenecks of LLMs are mainly influenced by their internal architecture. As a result, many researchers have begun exploring the potential internal mechanisms of LLMs, with most studies focusing on attention heads. Our survey aims to shed light on the internal reasoning processes of LLMs by concentrating on the underlying mechanisms of attention heads. We first distill the human thought process into a four-stage framework: Knowledge Recalling, In-Context Identification, Latent Reasoning, and Expression Preparation. Using this framework, we systematically review existing research to identify and categorize the functions of specific attention heads. Furthermore, we summarize the experimental methodologies used to discover these special heads, dividing them into two categories: Modeling-Free methods and Modeling-Required methods. Also, we outline relevant evaluation methods and benchmarks. Finally, we discuss the limitations of current research and propose several potential future directions., Comment: 29 pages, 11 figures, 4 tables, 5 equations

We apply numerical optimization and linear algebra algorithms for classical computers to the problem of automatically synthesizing algorithms for quantum computers. Using our framework, we apply several common techniques from these classical domains and numerically examine their suitability for and performance on this problem. Our methods are evaluated on single-qubit systems as well as on larger systems. While the first part of our proposed method outputs a single unitary matrix representing the composite effects of a quantum circuit or algorithm, we use existing tools - and assess the performance of these - to factor such a matrix into a product of elementary quantum gates. This enables our pipeline to be truly end-to-end: starting from desired input/output examples, our code ultimately results in a quantum circuit diagram. We release our code to the research community (upon acceptance)., Comment: 17 pages, 9 figures

This paper presents the first public data release (DR1) of the FRB Line-of-sight Ionization Measurement From Lightcone AAOmega Mapping (FLIMFLAM) Survey, a wide field spectroscopic survey targeted on the fields of 10 precisely localized Fast Radio Bursts (FRBs). DR1 encompasses spectroscopic data for 10,468 galaxy redshifts across 10 FRBs fields with z<0.4, covering approximately 26 deg^2 of the sky in total. FLIMFLAM is composed of several layers, encompassing the `Wide' (covering ~ degree or >10 Mpc scales), `Narrow', (several-arcminute or ~ Mpc) and integral field unit (`IFU'; ~ arcminute or ~ 100 kpc ) components. The bulk of the data comprise spectroscopy from the 2dF-AAOmega on the 3.9-meter Anglo-Australian Telescope, while most of the Narrow and IFU data was achieved using an ensemble of 8-10-meter class telescopes. We summarize the information on our selected FRB fields, the criteria for target selection, methodologies employed for data reduction, spectral analysis processes, and an overview of our data products. An evaluation of our data reveals an average spectroscopic completeness of 48.43%, with over 80% of the observed targets having secure redshifts. Additionally, we describe our approach on generating angular masks and calculating the target selection functions, setting the stage for the impending reconstruction of the matter density field., Comment: 23 pages, 8 figures, Submitted to ApJS

The dispersion measure of fast radio bursts (FRBs), arising from the interactions of the pulses with free electrons along the propagation path, constitutes a unique probe of the cosmic baryon distribution. Their constraining power is further enhanced in combination with observations of the foreground large-scale structure and intervening galaxies. In this work, we present the first constraints on the partition of the cosmic baryons between the intergalactic medium (IGM) and circumgalactic medium (CGM), inferred from the FLIMFLAM spectroscopic survey. In its first data release, the FLIMFLAM survey targeted galaxies in the foreground of 8 localized FRBs. Using Bayesian techniques, we reconstruct the underlying ~Mpc-scale matter density field that is traced by the IGM gas. Simultaneously, deeper spectroscopy of intervening foreground galaxies (at impact parameters $b_\perp \lesssim r_{200}$) and the FRB host galaxies constrains the contribution from the CGM. Applying Bayesian parameter inference to our data and assuming a fiducial set of priors, we infer the IGM cosmic baryon fraction to be $f_{\rm igm}=0.59^{+0.11}_{-0.10}$, and a CGM gas fraction of $f_{\rm gas} = 0.55^{+0.26}_{-0.29}$ for $10^{10}\,M_\odot \lesssim M_{\rm halo}\lesssim 10^{13}\,M_\odot$ halos. The mean FRB host dispersion measure (rest-frame) in our sample is $\langle \rm{DM_{host}}\rangle = 90^{+29}_{-19}\rm{pc~cm^{-3}}$, of which $\langle{\rm DM_{host}^{unk}}\rangle =69^{+28}_{-19}~\rm{pc~cm^{-3}}$ arises from the host galaxy ISM and/or the FRB progenitor environment. While our current $f_{\rm igm}$ and $f_{\rm gas}$ uncertainties are too broad to constrain most galactic feedback models, this result marks the first measurement of the IGM and CGM baryon fractions, as well as the first systematic separation of the FRB host dispersion measure into two components: arising from the halo and from the inner ISM/FRB engine., Comment: 21 pages, 9 figures, 5 tables; Submitted to ApJ

Multifocus image fusion is an effective way to overcome the limitation of optical lenses. Many existing methods obtain fused results by generating decision maps. However, such methods often assume that the focused areas of the two source images are complementary, making it impossible to achieve simultaneous fusion of multiple images. Additionally, the existing methods ignore the impact of hard pixels on fusion performance, limiting the visual quality improvement of fusion image. To address these issues, a combining generation and recombination model, termed as GRFusion, is proposed. In GRFusion, focus property detection of each source image can be implemented independently, enabling simultaneous fusion of multiple source images and avoiding information loss caused by alternating fusion. This makes GRFusion free from the number of inputs. To distinguish the hard pixels from the source images, we achieve the determination of hard pixels by considering the inconsistency among the detection results of focus areas in source images. Furthermore, a multi-directional gradient embedding method for generating full focus images is proposed. Subsequently, a hard-pixel-guided recombination mechanism for constructing fused result is devised, effectively integrating the complementary advantages of feature reconstruction-based method and focused pixel recombination-based method. Extensive experimental results demonstrate the effectiveness and the superiority of the proposed method.The source code will be released on https://github.com/xxx/xxx.

Nowadays, autonomous cars can drive smoothly in ordinary cases, and it is widely recognized that realistic sensor simulation will play a critical role in solving remaining corner cases by simulating them. To this end, we propose an autonomous driving simulator based upon neural radiance fields (NeRFs). Compared with existing works, ours has three notable features: (1) Instance-aware. Our simulator models the foreground instances and background environments separately with independent networks so that the static (e.g., size and appearance) and dynamic (e.g., trajectory) properties of instances can be controlled separately. (2) Modular. Our simulator allows flexible switching between different modern NeRF-related backbones, sampling strategies, input modalities, etc. We expect this modular design to boost academic progress and industrial deployment of NeRF-based autonomous driving simulation. (3) Realistic. Our simulator set new state-of-the-art photo-realism results given the best module selection. Our simulator will be open-sourced while most of our counterparts are not. Project page: https://open-air-sun.github.io/mars/., Comment: CICAI 2023, project page with code: https://open-air-sun.github.io/mars/

The repeating fast radio burst FRB20190520B is an anomaly of the FRB population thanks to its high dispersion measure (DM$=1205\,$pc/cc) despite its low redshift of $z_\mathrm{frb}=0.241$. This excess has been attributed to a large host contribution of $DM_{host}\approx 900\,$pc/cc, far larger than any other known FRB. In this paper, we describe spectroscopic observations of the FRB20190520B field obtained as part of the FLIMFLAM survey, which yielded 701 galaxy redshifts in the field. We find multiple foreground galaxy groups and clusters, for which we then estimated halo masses by comparing their richness with numerical simulations. We discover two separate $M_{halo} >10^{14}\,M_\odot$ galaxy clusters, at $z=0.1867$ and $z=0.2170$, respectively, that are directly intersected by the FRB sightline within their characteristic halo radius \rvir{}. Subtracting off their estimated DM contributions as well that of the diffuse intergalactic medium, we estimate a host contribution of $DM_{host}=430^{+140}_{-220}\,$pc/cc or $DM_{host}=280^{+140}_{-170}\,$pc/cc (observed frame) depending on whether we assume the halo gas extends to $r_{200}$ or $2\times r_{200}$. This significantly smaller $DM_{host}$ -- no longer the largest known value -- is now consistent with H$\alpha$ emission measures of the host galaxy without invoking unusually high gas temperatures. Combined with the observed FRB scattering timescale, we estimate the turbulent fluctuation and geometric amplification factor of the scattering layer to be $\tilde{F} G\approx4.5 - 11\,(\mathrm{pc^2\;km})^{-1/3}$, suggesting most of the gas is close to the FRB host. This result illustrates the importance of incorporating foreground data for FRB analyses, both for understanding the nature of FRBs and to realize their potential as a cosmological probe., Comment: Accepted by ApJL. Interactive figure included (link in text). Note numerical values have changed from v1

Accurate estimation of multiple quality variables is critical for building industrial soft sensor models, which have long been confronted with data efficiency and negative transfer issues. Methods sharing backbone parameters among tasks address the data efficiency issue; however, they still fail to mitigate the negative transfer problem. To address this issue, a balanced Mixture-of-Experts (BMoE) is proposed in this work, which consists of a multi-gate mixture of experts (MMoE) module and a task gradient balancing (TGB) module. The MoE module aims to portray task relationships, while the TGB module balances the gradients among tasks dynamically. Both of them cooperate to mitigate the negative transfer problem. Experiments on the typical sulfur recovery unit demonstrate that BMoE models task relationship and balances the training process effectively, and achieves better performance than baseline models significantly.

The FLIMFLAM survey is collecting spectroscopic data of field galaxies near fast radio burst (FRB) sightlines to constrain key parameters describing the distribution of matter in the Universe. In this work, we leverage the survey data to determine the source of the excess extragalactic dispersion measure (DM), compared to the Macquart relation estimate of four FRBs: FRB20190714A, FRB20200430A, FRB20200906A, and FRB20210117A. By modeling the gas distribution around the foreground galaxy halos and galaxy groups of the sightlines, we estimate $\rm DM_{halos}$, their contribution to the FRB dispersion measures. The FRB20190714A sightline shows a clear excess of foreground halos which contribute roughly 2/3$^{rd}$ of the observed excess DM, thus implying a baryon-dense sightline. FRB20200906A shows a smaller but non-negligible foreground halo contribution, and further analysis of the IGM is necessary to ascertain the true cosmic contribution to its DM. RB20200430A and FRB20210117A show negligible foreground contributions, implying a large host galaxy excess and/or progenitor environment excess., Comment: 14 pages, 6 figures, 3 tables. Data access available here: https://drive.google.com/drive/folders/1gVVwTAmiHJtFB_GD56hgmDylXHYbOTFS

An accurate and explainable automatic monitoring system is critical for the safety of high efficiency energy conversion plants that operate under extreme working condition. Nonetheless, currently available data-driven monitoring systems often fall short in meeting the requirements for either high-accuracy or interpretability, which hinders their application in practice. To overcome this limitation, a data-driven approach, AttentionMixer, is proposed under a generalized message passing framework, with the goal of establishing an accurate and interpretable radiation monitoring framework for energy conversion plants. To improve the model accuracy, the first technical contribution involves the development of spatial and temporal adaptive message passing blocks, which enable the capture of spatial and temporal correlations, respectively; the two blocks are cascaded through a mixing operator. To enhance the model interpretability, the second technical contribution involves the implementation of a sparse message passing regularizer, which eliminates spurious and noisy message passing routes. The effectiveness of the AttentionMixer approach is validated through extensive evaluations on a monitoring benchmark collected from the national radiation monitoring network for nuclear power plants, resulting in enhanced monitoring accuracy and interpretability in practice.

It has been shown that learning radiance fields with depth rendering and depth supervision can effectively promote the quality and convergence of view synthesis. However, this paradigm requires input RGB-D sequences to be synchronized, hindering its usage in the UAV city modeling scenario. As there exists asynchrony between RGB images and depth images due to high-speed flight, we propose a novel time-pose function, which is an implicit network that maps timestamps to $\rm SE(3)$ elements. To simplify the training process, we also design a joint optimization scheme to jointly learn the large-scale depth-regularized radiance fields and the time-pose function. Our algorithm consists of three steps: (1) time-pose function fitting, (2) radiance field bootstrapping, (3) joint pose error compensation and radiance field refinement. In addition, we propose a large synthetic dataset with diverse controlled mismatches and ground truth to evaluate this new problem setting systematically. Through extensive experiments, we demonstrate that our method outperforms baselines without regularization. We also show qualitatively improved results on a real-world asynchronous RGB-D sequence captured by drone. Codes, data, and models will be made publicly available.

As a basic research problem for building effective recommender systems, post-click conversion rate (CVR) estimation has long been plagued by sample selection bias and data sparsity issues. To address the data sparsity issue, prevalent methods based on entire space multi-task model leverage the sequential pattern of user actions, i.e. exposure $\rightarrow$ click $\rightarrow$ conversion to construct auxiliary learning tasks. However, they still fall short of guaranteeing the unbiasedness of CVR estimates. This paper theoretically demonstrates two defects of these entire space multi-task models: (1) inherent estimation bias (IEB) for CVR estimation, where the CVR estimate is inherently higher than the ground truth; (2) potential independence priority (PIP) for CTCVR estimation, where the causality from click to conversion might be overlooked. This paper further proposes a principled method named entire space counterfactual multi-task model (ESCM$^2$), which employs a counterfactual risk minimizer to handle both IEB and PIP issues at once. To demonstrate the effectiveness of the proposed method, this paper explores its parameter tuning in practice, derives its analytic properties, and showcases its effectiveness in industrial CVR estimation, where ESCM$^2$ can effectively alleviate the intrinsic IEB and PIP issues and outperform baseline models., Comment: This submission is an extension of arXiv:2204.05125

We present ConchShell, a multi-modal generative adversarial framework that takes pictures as input to the network and generates piano music samples that match the picture context. Inspired by I3D, we introduce a novel image feature representation method: time-convolutional neural network (TCNN), which is used to forge features for images in the temporal dimension. Although our image data consists of only six categories, our proposed framework will be innovative and commercially meaningful. The project will provide technical ideas for work such as 3D game voice overs, short-video soundtracks, and real-time generation of metaverse background music.We have also released a new dataset, the Beach-Ocean-Piano Dataset (BOPD) 1, which contains more than 3,000 images and more than 1,500 piano pieces. This dataset will support multimodal image-to-music research., Comment: 5 pages

Intellectual properties is increasingly important in the economic development. To solve the pain points by traditional methods in IP evaluation, we are developing a new technology with machine learning as the core. We have built an online platform and will expand our business in the Greater Bay Area with plans., Comment: 5 pages, 2 figures

PaddleSpeech is an open-source all-in-one speech toolkit. It aims at facilitating the development and research of speech processing technologies by providing an easy-to-use command-line interface and a simple code structure. This paper describes the design philosophy and core architecture of PaddleSpeech to support several essential speech-to-text and text-to-speech tasks. PaddleSpeech achieves competitive or state-of-the-art performance on various speech datasets and implements the most popular methods. It also provides recipes and pretrained models to quickly reproduce the experimental results in this paper. PaddleSpeech is publicly avaiable at https://github.com/PaddlePaddle/PaddleSpeech.