Search

Recognizing overlapping speech from multiple speakers in conversational scenarios is one of the most challenging problem for automatic speech recognition (ASR). Serialized output training (SOT) is a classic method to address multi-talker ASR, with the idea of concatenating transcriptions from multiple speakers according to the emission times of their speech for training. However, SOT-style transcriptions, derived from concatenating multiple related utterances in a conversation, depend significantly on modeling long contexts. Therefore, compared to traditional methods that primarily emphasize encoder performance in attention-based encoder-decoder (AED) architectures, a novel approach utilizing large language models (LLMs) that leverages the capabilities of pre-trained decoders may be better suited for such complex and challenging scenarios. In this paper, we propose an LLM-based SOT approach for multi-talker ASR, leveraging pre-trained speech encoder and LLM, fine-tuning them on multi-talker dataset using appropriate strategies. Experimental results demonstrate that our approach surpasses traditional AED-based methods on the simulated dataset LibriMix and achieves state-of-the-art performance on the evaluation set of the real-world dataset AMI, outperforming the AED model trained with 1000 times more supervised data in previous works., Comment: 8 pages, accepted by IEEE SLT 2024

Precise direct cosmic-ray (CR) measurements provide an important probe to study the energetic particle sources in our Galaxy, and the interstellar environment through which these particles propagate. Uncertainties on hadronic models, ion-nucleon cross sections in particular, are currently the limiting factor towards obtaining more accurate CR ion flux measurements with calorimetric space-based experiments. We present an energy-dependent measurement of the inelastic cross section of protons and helium-4 nuclei (alpha particles) on a Bi$_4$Ge$_3$O$_{12}$ target, using 88 months of data collected by the DAMPE space mission. The kinetic energy range per nucleon of the measurement points ranges from 18 GeV to 9 TeV for protons, and from 5 GeV/n to 3 TeV/n for helium-4 nuclei. Our results lead to a significant improvement of the CR flux normalisation. In the case of helium-4, these results correspond to the first cross section measurements on a heavy target material at energies above 10 GeV/n., Comment: 17 pages, submitted to PRD

Egocentric human pose estimation (HPE) using wearable sensors is essential for VR/AR applications. Most methods rely solely on either egocentric-view images or sparse Inertial Measurement Unit (IMU) signals, leading to inaccuracies due to self-occlusion in images or the sparseness and drift of inertial sensors. Most importantly, the lack of real-world datasets containing both modalities is a major obstacle to progress in this field. To overcome the barrier, we propose EMHI, a multimodal \textbf{E}gocentric human \textbf{M}otion dataset with \textbf{H}ead-Mounted Display (HMD) and body-worn \textbf{I}MUs, with all data collected under the real VR product suite. Specifically, EMHI provides synchronized stereo images from downward-sloping cameras on the headset and IMU data from body-worn sensors, along with pose annotations in SMPL format. This dataset consists of 885 sequences captured by 58 subjects performing 39 actions, totaling about 28.5 hours of recording. We evaluate the annotations by comparing them with optical marker-based SMPL fitting results. To substantiate the reliability of our dataset, we introduce MEPoser, a new baseline method for multimodal egocentric HPE, which employs a multimodal fusion encoder, temporal feature encoder, and MLP-based regression heads. The experiments on EMHI show that MEPoser outperforms existing single-modal methods and demonstrates the value of our dataset in solving the problem of egocentric HPE. We believe the release of EMHI and the method could advance the research of egocentric HPE and expedite the practical implementation of this technology in VR/AR products.

We present Flow Matching for Reaction Coordinates (FMRC), a novel deep learning algorithm designed to identify optimal reaction coordinates (RC) in biomolecular reversible dynamics. FMRC is based on the mathematical principles of lumpability and decomposability, which we reformulate into a conditional probability framework for efficient data-driven optimization using deep generative models. While FMRC does not explicitly learn the well-established transfer operator or its eigenfunctions, it can effectively encode the dynamics of leading eigenfunctions of the system transfer operator into its low-dimensional RC space. We further quantitatively compare its performance with several state-of-the-art algorithms by evaluating the quality of Markov State Models (MSM) constructed in their respective RC spaces, demonstrating the superiority of FMRC in three increasingly complex biomolecular systems. Finally, we discuss its potential applications in downstream applications such as enhanced sampling methods and MSM construction.

Using $(2712.4 \pm 14.3) \times 10^6~\psi$(3686) events collected with the BESIII detector operating at the BEPCII collider, we search for the hadronic transition $h_c \to \pi^+\pi^-J/\psi$ via $\psi(3686)\to \pi^0 h_c$. No significant signal is observed. We set the most stringent upper limits to date on the branching fractions $\mathcal{B}(\psi(3686)\to \pi^0 h_c)\times\mathcal{B}(h_c\to\pi^+\pi^-J/\psi)$ and $\mathcal{B}(h_c \to \pi^+\pi^-J/\psi)$ at the 90$\%$ confidence level, which are determined to be $6.7\times 10^{-7}$ and $9.4 \times10^{-4}$, respectively.

Quantum-mechanics-based transport simulation is of importance for the design of ultra-short channel field-effect transistors (FETs) with its capability of understanding the physical mechanism, while facing the primary challenge of the high computational intensity. Traditional machine learning is expected to accelerate the optimization of FET design, yet its application in this field is limited by the lack of both high-fidelity datasets and the integration of physical knowledge. Here, we introduced a physics-integrated neural network framework to predict the transport curves of sub-5-nm gate-all-around (GAA) FETs using an in-house developed high-fidelity database. The transport curves in the database are collected from literature and our first-principles calculations. Beyond silicon, we included indium arsenide, indium phosphide, and selenium nanowires with different structural phases as the FET channel materials. Then, we built a physical-knowledge-integrated hyper vector neural network (PHVNN), in which five new physical features were added into the inputs for prediction transport characteristics, achieving a sufficiently low mean absolute error of 0.39. In particular, ~98% of the current prediction residuals are within one order of magnitude. Using PHVNN, we efficiently screened out the symmetric p-type GAA FETs that possess the same figures of merit with the n-type ones, which are crucial for the fabrication of homogeneous CMOS circuits. Finally, our automatic differentiation analysis provides interpretable insights into the PHVNN, which highlights the important contributions of our new input parameters and improves the reliability of PHVNN. Our approach provides an effective method for rapidly screening appropriate GAA FETs with the prospect of accelerating the design process of next-generation electronic devices., Comment: 13 pages, 4 figures

In this Letter, a systematic study of the weak radiative hyperon decay $\Xi^{0}\to\Lambda\gamma$ at an electron-positron collider using entangled $\Xi^{0}\bar{\Xi}^{0}$ pair events is presented. The absolute branching fraction for this decay has been measured for the first time, and is $\left(1.347 \pm 0.066_{\mathrm stat.}\pm0.054_{\mathrm syst.}\right)\times 10^{-3}$. The decay asymmetry parameter, which characterizes the effect of parity violation in the decay, is determined to be $-0.741 \pm 0.062_{\mathrm stat.}\pm 0.019_{\mathrm syst.}$. The obtained results are consistent with the world average values within the uncertainties, offering valuable insights into the underlying mechanism governing the weak radiative hyperon decays. The charge conjugation parity ($CP$) symmetries of branching fraction and decay asymmetry parameter in the decay are also studied. No statistically significant violation of charge conjugation parity symmetry is observed., Comment: 10 pages, 3 figures

The rare electromagnetic $J/\psi \to \mu^+\mu^-\mu^+\mu^-$ decay is observed with a significance greatly exceeding the discovery threshold, using proton-proton collision data collected by the LHCb experiment during 2016-2018 at a center-of-mass energy of 13 TeV, corresponding to an integrated luminosity of $5.4\,\text{fb}^{-1}$. The rate of this decay is measured relative to that of the $J/\psi \to \mu^+\mu^-$ mode. Using the QED model for the four-muon decay in the efficiency estimation, its branching fraction is determined to be \begin{equation*} {\mathcal{B}}(J/\psi \to \mu^+\mu^-\mu^+\mu^-) = (1.13\pm0.10\pm0.05\pm0.01)\times 10^{-6}, \end{equation*} where the uncertainties are statistical, systematic and due to the uncertainty on the branching fraction of the $J/\psi \to \mu^+\mu^-$ decay., Comment: All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://lbfence.cern.ch/alcm/public/analysis/full-details/3453 (LHCb public pages)

Beginning with VisualGLM and CogVLM, we are continuously exploring VLMs in pursuit of enhanced vision-language fusion, efficient higher-resolution architecture, and broader modalities and applications. Here we propose the CogVLM2 family, a new generation of visual language models for image and video understanding including CogVLM2, CogVLM2-Video and GLM-4V. As an image understanding model, CogVLM2 inherits the visual expert architecture with improved training recipes in both pre-training and post-training stages, supporting input resolution up to $1344 \times 1344$ pixels. As a video understanding model, CogVLM2-Video integrates multi-frame input with timestamps and proposes automated temporal grounding data construction. Notably, CogVLM2 family has achieved state-of-the-art results on benchmarks like MMBench, MM-Vet, TextVQA, MVBench and VCGBench. All models are open-sourced in https://github.com/THUDM/CogVLM2 and https://github.com/THUDM/GLM-4, contributing to the advancement of the field.

Sound source localization aims to localize objects emitting the sound in visual scenes. Recent works obtaining impressive results typically rely on contrastive learning. However, the common practice of randomly sampling negatives in prior arts can lead to the false negative issue, where the sounds semantically similar to visual instance are sampled as negatives and incorrectly pushed away from the visual anchor/query. As a result, this misalignment of audio and visual features could yield inferior performance. To address this issue, we propose a novel audio-visual learning framework which is instantiated with two individual learning schemes: self-supervised predictive learning (SSPL) and semantic-aware contrastive learning (SACL). SSPL explores image-audio positive pairs alone to discover semantically coherent similarities between audio and visual features, while a predictive coding module for feature alignment is introduced to facilitate the positive-only learning. In this regard SSPL acts as a negative-free method to eliminate false negatives. By contrast, SACL is designed to compact visual features and remove false negatives, providing reliable visual anchor and audio negatives for contrast. Different from SSPL, SACL releases the potential of audio-visual contrastive learning, offering an effective alternative to achieve the same goal. Comprehensive experiments demonstrate the superiority of our approach over the state-of-the-arts. Furthermore, we highlight the versatility of the learned representation by extending the approach to audio-visual event classification and object detection tasks. Code and models are available at: https://github.com/zjsong/SACL., Comment: arXiv admin note: substantial text overlap with arXiv:2203.13412

In this paper, we introduce an innovative hierarchical joint source-channel coding (HJSCC) framework for image transmission, utilizing a hierarchical variational autoencoder (VAE). Our approach leverages a combination of bottom-up and top-down paths at the transmitter to autoregressively generate multiple hierarchical representations of the original image. These representations are then directly mapped to channel symbols for transmission by the JSCC encoder. We extend this framework to scenarios with a feedback link, modeling transmission over a noisy channel as a probabilistic sampling process and deriving a novel generative formulation for JSCC with feedback. Compared with existing approaches, our proposed HJSCC provides enhanced adaptability by dynamically adjusting transmission bandwidth, encoding these representations into varying amounts of channel symbols. Additionally, we introduce a rate attention module to guide the JSCC encoder in optimizing its encoding strategy based on prior information. Extensive experiments on images of varying resolutions demonstrate that our proposed model outperforms existing baselines in rate-distortion performance and maintains robustness against channel noise.

Data-enabled predictive control (DeePC) is a data-driven control algorithm that utilizes data matrices to form a non-parametric representation of the underlying system, predicting future behaviors and generating optimal control actions. DeePC typically requires solving an online optimization problem, the complexity of which is heavily influenced by the amount of data used, potentially leading to expensive online computation. In this paper, we leverage deep learning to propose a highly computationally efficient DeePC approach for general nonlinear processes, referred to as Deep DeePC. Specifically, a deep neural network is employed to learn the DeePC vector operator, which is an essential component of the non-parametric representation of DeePC. This neural network is trained offline using historical open-loop input and output data of the nonlinear process. With the trained neural network, the Deep DeePC framework is formed for online control implementation. At each sampling instant, this neural network directly outputs the DeePC operator, eliminating the need for online optimization as conventional DeePC. The optimal control action is obtained based on the DeePC operator updated by the trained neural network. To address constrained scenarios, a constraint handling scheme is further proposed and integrated with the Deep DeePC to handle hard constraints during online implementation. The efficacy and superiority of the proposed Deep DeePC approach are demonstrated using two benchmark process examples., Comment: 34 pages, 7 figures

Enhancing the safety of autonomous vehicles is crucial, especially given recent accidents involving automated systems. As passengers in these vehicles, humans' sensory perception and decision-making can be integrated with autonomous systems to improve safety. This study explores neural mechanisms in passenger-vehicle interactions, leading to the development of a Passenger Cognitive Model (PCM) and the Passenger EEG Decoding Strategy (PEDS). Central to PEDS is a novel Convolutional Recurrent Neural Network (CRNN) that captures spatial and temporal EEG data patterns. The CRNN, combined with stacking algorithms, achieves an accuracy of $85.0\% \pm 3.18\%$. Our findings highlight the predictive power of pre-event EEG data, enhancing the detection of hazardous scenarios and offering a network-driven framework for safer autonomous vehicles.

Measurements of the strong-phase difference between $D^0$ and $\bar{D}^0\to\pi^+\pi^-\pi^+\pi^-$ are performed in bins of phase space. The study exploits a sample of quantum-correlated $D\bar{D}$ mesons collected by the BESIII experiment in $e^+e^-$ collisions at a center-of-mass energy of 3.773~GeV, corresponding to an integrated luminosity of 2.93~fb$^{-1}$. Here, $D$ denotes a neutral charm meson in a superposition of flavor eigenstates. The reported results are valuable for measurements of the $C\!P$-violating phase $\gamma$ (also denoted $\phi_3$) in $B^\pm \to DK^\pm$, $D \to \pi^+\pi^-\pi^+\pi^-$ decays, and the binning schemes are designed to provide good statistical sensitivity to this parameter. The expected uncertainty on $\gamma$ arising from the precision of the strong-phase measurements, when applied to very large samples of $B$-meson decays, is around $1.5^\circ$ or $2^\circ$, depending on the binning scheme. The binned strong-phase parameters are combined to give a value of $F_+^{4\pi} = 0.746 \pm 0.010 \pm 0.004$ for the $C\!P$-even fraction of $D^0 \to \pi^+\pi^-\pi^+\pi^-$ decays, which is around 30\% more precise than the previous best measurement of this quantity.

The AlphaFold series has transformed protein structure prediction with remarkable accuracy, often matching experimental methods. AlphaFold2, AlphaFold-Multimer, and the latest AlphaFold3 represent significant strides in predicting single protein chains, protein complexes, and biomolecular structures. While AlphaFold2 and AlphaFold-Multimer are open-sourced, facilitating rapid and reliable predictions, AlphaFold3 remains partially accessible through a limited online server and has not been open-sourced, restricting further development. To address these challenges, the PaddleHelix team is developing HelixFold3, aiming to replicate AlphaFold3's capabilities. Using insights from previous models and extensive datasets, HelixFold3 achieves an accuracy comparable to AlphaFold3 in predicting the structures of conventional ligands, nucleic acids, and proteins. The initial release of HelixFold3 is available as open source on GitHub for academic research, promising to advance biomolecular research and accelerate discoveries. We also provide online service at PaddleHelix website at https://paddlehelix.baidu.com/app/all/helixfold3/forecast.

This review paper describes the energy-upgraded CEBAF accelerator. This superconducting linac has achieved 12 GeV beam energy by adding 11 new high-performance cryomodules containing eighty-eight superconducting cavities that have operated CW at an average accelerating gradient of 20 MV/m. After reviewing the attributes and performance of the previous 6 GeV CEBAF accelerator, we discuss the upgraded CEBAF accelerator system in detail with particular attention paid to the new beam acceleration systems. In addition to doubling the acceleration in each linac, the upgrade included improving the beam recirculation magnets, adding more helium cooling capacity to allow the newly installed modules to run cold, adding a new experimental hall, and improving numerous other accelerator components. We review several of the techniques deployed to operate and analyze the accelerator performance, and document system operating experience and performance. In the final portion of the document, we present much of the current planning regarding projects to improve accelerator performance and enhance operating margins, and our plans for ensuring CEBAF operates reliably into the future. For the benefit of potential users of CEBAF, the performance and quality measures for beam delivered to each of the experimental halls is summarized in the appendix., Comment: 66 pages, 73 figures, 21 tables

Unveiling atomic structures is significant for the relationship construction between microscopic configurations and macroscopic properties of materials. However, we still lack a rapid, accurate, and robust approach to automatically resolve complex patterns in atomic-resolution microscopy. Here, we present a Trident strategy-enhanced disentangled representation learning method (a generative model), which utilizes a few unlabeled experimental images with abundant low-cost simulated images to generate a large corpus of annotated simulation data that closely resembles experimental conditions, realizing simultaneous achievement of high quality and large volumes of the training dataset. A structural inference model is then trained via a residual neural network which can directly deduce the interlayer slip and rotation of diversified and complicated stacking patterns at van der Waals (vdWs) interfaces with picometer-scale accuracy across various materials (ReS2, ReSe2, and MoS2) with different layer numbers (bilayer and trilayers) and demonstrates robustness to defects, imaging quality, and surface contaminations. The framework can also identify pattern transition interfaces, quantify subtle motif variations, and discriminate moir\'e patterns that are undistinguishable in frequency domains. The high-throughput processing ability of our method helps discover a novel vdW epitaxy where various thermodynamically favorable slip stackings can coexist, demonstrating the machine learning contribution to the new knowledge emergence., Comment: 25 pages,5 figures

The limitations of task-specific and general image restoration methods for specific degradation have prompted the development of all-in-one image restoration techniques. However, the diversity of patterns among multiple degradation, along with the significant uncertainties in mapping between degraded images of different severities and their corresponding undistorted versions, pose significant challenges to the all-in-one restoration tasks. To address these challenges, we propose Perceive-IR, an all-in-one image restorer designed to achieve fine-grained quality control that enables restored images to more closely resemble their undistorted counterparts, regardless of the type or severity of degradation. Specifically, Perceive-IR contains two stages: (1) prompt learning stage and (2) restoration stage. In the prompt learning stage, we leverage prompt learning to acquire a fine-grained quality perceiver capable of distinguishing three-tier quality levels by constraining the prompt-image similarity in the CLIP perception space. Subsequently, this quality perceiver and difficulty-adaptive perceptual loss are integrated as a quality-aware learning strategy to realize fine-grained quality control in restoration stage. For the restoration stage, a semantic guidance module (SGM) and compact feature extraction (CFE) are proposed to further promote the restoration process by utilizing the robust semantic information from the pre-trained large scale vision models and distinguishing degradation-specific features. Extensive experiments demonstrate that our Perceive-IR outperforms state-of-the-art methods in all-in-one image restoration tasks and exhibit superior generalization ability when dealing with unseen tasks., Comment: 13 pages, 8 figures

Robust and realistic rendering for large-scale road scenes is essential in autonomous driving simulation. Recently, 3D Gaussian Splatting (3D-GS) has made groundbreaking progress in neural rendering, but the general fidelity of large-scale road scene renderings is often limited by the input imagery, which usually has a narrow field of view and focuses mainly on the street-level local area. Intuitively, the data from the drone's perspective can provide a complementary viewpoint for the data from the ground vehicle's perspective, enhancing the completeness of scene reconstruction and rendering. However, training naively with aerial and ground images, which exhibit large view disparity, poses a significant convergence challenge for 3D-GS, and does not demonstrate remarkable improvements in performance on road views. In order to enhance the novel view synthesis of road views and to effectively use the aerial information, we design an uncertainty-aware training method that allows aerial images to assist in the synthesis of areas where ground images have poor learning outcomes instead of weighting all pixels equally in 3D-GS training like prior work did. We are the first to introduce the cross-view uncertainty to 3D-GS by matching the car-view ensemble-based rendering uncertainty to aerial images, weighting the contribution of each pixel to the training process. Additionally, to systematically quantify evaluation metrics, we assemble a high-quality synthesized dataset comprising both aerial and ground images for road scenes., Comment: BMVC2024 Project Page: https://sainingzhang.github.io/project/uc-gs/ Code: https://github.com/SainingZhang/uc-gs/

The polar magnetic fields of the Sun play an important role in governing solar activity and powering fast solar wind. However, because our view of the Sun is limited in the ecliptic plane, the polar regions remain largely uncharted. Using the high spatial resolution and polarimetric precision vector magnetograms observed by Hinode from 2012 to 2021, we investigate the long-term variation of the magnetic fields in polar caps at different latitudes. The Hinode magnetic measurements show that the polarity reversal processes in the north and south polar caps are non-simultaneous. The variation of the averaged radial magnetic flux density reveals that, in each polar cap, the polarity reversal is completed successively from the 70 degree latitude to the pole, reflecting a poleward magnetic flux migration therein. These results clarify the polar magnetic polarity reversal process at different latitudes., Comment: 1 table, 4 figures, published, 2024, Res. Astron. Astrophys., 24, 075015

Single-cell RNA sequencing (scRNA-seq) allows transcriptional profiling, and cell-type annotation of individual cells. However, sample preparation in typical scRNA-seq experiments often homogenizes the samples, thus spatial locations of individual cells are often lost. Although spatial transcriptomic techniques, such as in situ hybridization (ISH) or Slide-seq, can be used to measure gene expression in specific locations in samples, it remains a challenge to measure or infer expression level for every gene at a single-cell resolution in every location in tissues. Existing computational methods show promise in reconstructing these missing data by integrating scRNA-seq data with spatial expression data such as those obtained from spatial transcriptomics. Here we describe Laplacian Linear Optimal Transport (LLOT), an interpretable method to integrate single-cell and spatial transcriptomics data to reconstruct missing information at a whole-genome and single-cell resolution. LLOT iteratively corrects platform effects and employs Laplacian Optimal Transport to decompose each spot in spatial transcriptomics data into a spatially-smooth probabilistic mixture of single cells. We benchmarked LLOT against several other methods on datasets of Drosophila embryo, mouse cerebellum and synthetic datasets generated by scDesign3 in the paper, and another three datasets in the supplementary. The results showed that LLOT consistently outperformed others in reconstructing spatial expressions.

The general capabilities of Large Language Models (LLM) highly rely on the composition and selection on extensive pretraining datasets, treated as commercial secrets by several institutions. To mitigate this issue, we open-source the details of a universally applicable data processing pipeline and validate its effectiveness and potential by introducing a competitive LLM baseline. Specifically, the data processing pipeline consists of broad collection to scale up and reweighting to improve quality. We then pretrain a 7B model BaichuanSEED with 3T tokens processed by our pipeline without any deliberate downstream task-related optimization, followed by an easy but effective supervised fine-tuning stage. BaichuanSEED demonstrates consistency and predictability throughout training and achieves comparable performance on comprehensive benchmarks with several commercial advanced large language models, such as Qwen1.5 and Llama3. We also conduct several heuristic experiments to discuss the potential for further optimization of downstream tasks, such as mathematics and coding., Comment: 19 pages, 6 figures

Computational pathology (CPath) has significantly advanced the clinical practice of pathology. Despite the progress made, Multiple Instance Learning (MIL), a promising paradigm within CPath, continues to face challenges, particularly related to incomplete information utilization. Existing frameworks, such as those based on Convolutional Neural Networks (CNNs), attention, and selective scan space state sequential model (SSM), lack sufficient flexibility and scalability in fusing diverse features, and cannot effectively fuse diverse features. Additionally, current approaches do not adequately exploit order-related and order-independent features, resulting in suboptimal utilization of sequence information. To address these limitations, we propose a novel MIL framework called Mamba2MIL. Our framework utilizes the state space duality model (SSD) to model long sequences of patches of whole slide images (WSIs), which, combined with weighted feature selection, supports the fusion processing of more branching features and can be extended according to specific application needs. Moreover, we introduce a sequence transformation method tailored to varying WSI sizes, which enhances sequence-independent features while preserving local sequence information, thereby improving sequence information utilization. Extensive experiments demonstrate that Mamba2MIL surpasses state-of-the-art MIL methods. We conducted extensive experiments across multiple datasets, achieving improvements in nearly all performance metrics. Specifically, on the NSCLC dataset, Mamba2MIL achieves a binary tumor classification AUC of 0.9533 and an accuracy of 0.8794. On the BRACS dataset, it achieves a multiclass classification AUC of 0.7986 and an accuracy of 0.4981. The code is available at https://github.com/YuqiZhang-Buaa/Mamba2MIL.

Providing high-quality item recall for text queries is crucial in large-scale e-commerce search systems. Current Embedding-based Retrieval Systems (ERS) embed queries and items into a shared low-dimensional space, but uni-modality ERS rely too heavily on textual features, making them unreliable in complex contexts. While multi-modality ERS incorporate various data sources, they often overlook individual preferences for different modalities, leading to suboptimal results. To address these issues, we propose MRSE, a Multi-modality Retrieval System that integrates text, item images, and user preferences through lightweight mixture-of-expert (LMoE) modules to better align features across and within modalities. MRSE also builds user profiles at a multi-modality level and introduces a novel hybrid loss function that enhances consistency and robustness using hard negative sampling. Experiments on a large-scale dataset from Shopee and online A/B testing show that MRSE achieves an 18.9% improvement in offline relevance and a 3.7% gain in online core metrics compared to Shopee's state-of-the-art uni-modality system.

We study the 1D quantum many-body dynamics with a screened Coulomb potential in the mean-field setting. Combining the quantum mean-field, semiclassical, and Debye length limits, we prove the global derivation of the 1D Vlasov-Poisson equation. We tackle the difficulties brought by the pure state data, whose Wigner transforms converge to Wigner measures. We find new weighted uniform estimates around which we build the proof. As a result, we obtain, globally, stronger limits, and hence the global existence of solutions to the 1D Vlasov-Poisson equation subject to such Wigner measure data, which satisfy conservation laws of mass, momentum, and energy, despite being measure solutions. This happens to solve the 1D case of an open problem regarding the conservation law of the Vlasov-Poisson equation raised in [18] by Diperna and Lions.

With increasing privacy concerns in artificial intelligence, regulations have mandated the right to be forgotten, granting individuals the right to withdraw their data from models. Machine unlearning has emerged as a potential solution to enable selective forgetting in models, particularly in recommender systems where historical data contains sensitive user information. Despite recent advances in recommendation unlearning, evaluating unlearning methods comprehensively remains challenging due to the absence of a unified evaluation framework and overlooked aspects of deeper influence, e.g., fairness. To address these gaps, we propose CURE4Rec, the first comprehensive benchmark for recommendation unlearning evaluation. CURE4Rec covers four aspects, i.e., unlearning Completeness, recommendation Utility, unleaRning efficiency, and recommendation fairnEss, under three data selection strategies, i.e., core data, edge data, and random data. Specifically, we consider the deeper influence of unlearning on recommendation fairness and robustness towards data with varying impact levels. We construct multiple datasets with CURE4Rec evaluation and conduct extensive experiments on existing recommendation unlearning methods. Our code is released at https://github.com/xiye7lai/CURE4Rec.

Large language models (LLMs) have shown various ability on natural language processing, including problems about causality. It is not intuitive for LLMs to command causality, since pretrained models usually work on statistical associations, and do not focus on causes and effects in sentences. So that probing internal manipulation of causality is necessary for LLMs. This paper proposes a novel approach to probe causality manipulation hierarchically, by providing different shortcuts to models and observe behaviors. We exploit retrieval augmented generation (RAG) and in-context learning (ICL) for models on a designed causality classification task. We conduct experiments on mainstream LLMs, including GPT-4 and some smaller and domain-specific models. Our results suggest that LLMs can detect entities related to causality and recognize direct causal relationships. However, LLMs lack specialized cognition for causality, merely treating them as part of the global semantic of the sentence.

Information engines produce mechanical work through measurement and adaptive control. For information engines, the principal challenge lies in how to store the generated work for subsequent utilization. Here, we report an experimental demonstration where quantized mechanical motion serves as a quantum battery and gets charged in repeated cycles by a single trapped-ion information engine. This is enabled by a key technological advancement in rapid state discrimination, allowing us to suppress measurement-induced disturbances. Consequently, we were able to obtain a charging efficiency over 50\% of the theoretical limit at the optimal temperature. The experimental results substantiate that this approach can render trapped ions a promising platform for microscopic information engines with potential applications in the future upon scaling up.

Multi-instance partial-label learning (MIPL) addresses scenarios where each training sample is represented as a multi-instance bag associated with a candidate label set containing one true label and several false positives. Existing MIPL algorithms have primarily focused on mapping multi-instance bags to candidate label sets for disambiguation, disregarding the intrinsic properties of the label space and the supervised information provided by non-candidate label sets. In this paper, we propose an algorithm named ELIMIPL, i.e., Exploiting conjugate Label Information for Multi-Instance Partial-Label learning, which exploits the conjugate label information to improve the disambiguation performance. To achieve this, we extract the label information embedded in both candidate and non-candidate label sets, incorporating the intrinsic properties of the label space. Experimental results obtained from benchmark and real-world datasets demonstrate the superiority of the proposed ELIMIPL over existing MIPL algorithms and other well-established partial-label learning algorithms., Comment: Accepted at IJCAI 2024. The code can be found at https://github.com/tangw-seu/ELIMIPL

In recent years, foundation models (FMs) such as large language models (LLMs) and latent diffusion models (LDMs) have profoundly impacted diverse sectors, including music. This comprehensive review examines state-of-the-art (SOTA) pre-trained models and foundation models in music, spanning from representation learning, generative learning and multimodal learning. We first contextualise the significance of music in various industries and trace the evolution of AI in music. By delineating the modalities targeted by foundation models, we discover many of the music representations are underexplored in FM development. Then, emphasis is placed on the lack of versatility of previous methods on diverse music applications, along with the potential of FMs in music understanding, generation and medical application. By comprehensively exploring the details of the model pre-training paradigm, architectural choices, tokenisation, finetuning methodologies and controllability, we emphasise the important topics that should have been well explored, like instruction tuning and in-context learning, scaling law and emergent ability, as well as long-sequence modelling etc. A dedicated section presents insights into music agents, accompanied by a thorough analysis of datasets and evaluations essential for pre-training and downstream tasks. Finally, by underscoring the vital importance of ethical considerations, we advocate that following research on FM for music should focus more on such issues as interpretability, transparency, human responsibility, and copyright issues. The paper offers insights into future challenges and trends on FMs for music, aiming to shape the trajectory of human-AI collaboration in the music realm.

The recently discovered spin-triplet superconductor candidate UTe2 with Tc = 2 K has attracted enormous attention because it possesses many interesting properties, such as the extremely high upper critical field Hc2(0), chiral superconductivity and spontaneous time-reversal symmetry breaking, etc., all these suggest that it may be the long-sought spin-triplet superconductor. Here we report the discovery of superconductivity up to Tc = 4K in one of its siblings, i.e., UAs2 under high pressures. Interestingly, the UAs2 shows metallic behavior with an antiferromagnetic (AFM) transition at about 274 K under ambient pressure. Upon applying pressure, this transition is pushed down to lower temperatures with improved electric conductivity. When the pressure rises to about 20-22 GPa, superconductivity occurs together with the emergence of a linear temperature dependence of normal state resistance, the latter is a hallmark of the strange-metal state. The superconductivity with the highest Tc = 4 K is reached under a pressure of about 26.8 GPa, and it is robust against magnetic field with the upper critical field {\mu}0Hc2(0) ~ 12 T, far beyond the Pauli limit. Higher pressures will suppress the superconductivity and bring back the Fermi liquid behavior, showing a clear signature of quantum criticality. Our results open a new avenue for investigating the unconventional superconductivity concerning the mysterious 5f-band electrons in this uranium-based system., Comment: 16 pages of main text, 4 figures + 7 extended data figures

The nitrogen-vacancy (NV) center in diamond is a promising candidate for various quantum applications, such as quantum sensing. High sensitivity in NV-based magnetic sensing requires a diamond sample with a high density of NV centers and a long electron spin dephasing time. In this work, we propose a systematic measurement method for determining the electron spin dephasing time of NV center ensembles and analyze the contributions to the dephasing time from various sources, including NV-NV interactions, strain distribution, $^{13}C$ nuclear spin, and P1 electron spin. We demonstrate the effectiveness of our method on a series of high-performance diamond samples and provide a comprehensive understanding of dephasing sources, enabling the optimization of NV-based quantum sensing applications.

Kagome lattices facilitate various quantum phases, yet in bulk materials, their kagome flat-bands often interact with bulk bands, suppressing kagome electronic characteristics for hosting these phases. Here, we use density-functional-theory calculations to predict the geometric and electronic structures, as well as the topological and magnetic properties, of a series of MoTe$_{2-x}$ kagome monolayers formed by mirror-twin-boundary (MTB) loops. We analyze ten MTB-loop configurations of varying sizes and arrangements to assess their impact on various properties. Within the intrinsic bandgap of MoTe$_{2}$, we identify two sets of kagome bands, primarily originating from in-plane and out-of-plane Mo d-orbitals at MTB-loop edges and -vertices, respectively. Three configurations exhibit superior stability, while three others show comparable stability. Among these, four display bandgaps and potentially non-zero Z$_{2}$ topological invariants, suggesting possible topological phases, while the remaining two are metallic and feature Stoner magnetization. These findings guide the design of kagome-based two-dimensional materials with tunable electronic, topological, and magnetic properties., Comment: 18 pages, 5 figures

Existing works in single-image human reconstruction suffer from weak generalizability due to insufficient training data or 3D inconsistencies for a lack of comprehensive multi-view knowledge. In this paper, we introduce MagicMan, a human-specific multi-view diffusion model designed to generate high-quality novel view images from a single reference image. As its core, we leverage a pre-trained 2D diffusion model as the generative prior for generalizability, with the parametric SMPL-X model as the 3D body prior to promote 3D awareness. To tackle the critical challenge of maintaining consistency while achieving dense multi-view generation for improved 3D human reconstruction, we first introduce hybrid multi-view attention to facilitate both efficient and thorough information interchange across different views. Additionally, we present a geometry-aware dual branch to perform concurrent generation in both RGB and normal domains, further enhancing consistency via geometry cues. Last but not least, to address ill-shaped issues arising from inaccurate SMPL-X estimation that conflicts with the reference image, we propose a novel iterative refinement strategy, which progressively optimizes SMPL-X accuracy while enhancing the quality and consistency of the generated multi-views. Extensive experimental results demonstrate that our method significantly outperforms existing approaches in both novel view synthesis and subsequent 3D human reconstruction tasks., Comment: Project Page: https://thuhcsi.github.io/MagicMan

The rapid progress in Deep Learning (DL) and Large Language Models (LLMs) has exponentially increased demands of computational power and bandwidth. This, combined with the high costs of faster computing chips and interconnects, has significantly inflated High Performance Computing (HPC) construction costs. To address these challenges, we introduce the Fire-Flyer AI-HPC architecture, a synergistic hardware-software co-design framework and its best practices. For DL training, we deployed the Fire-Flyer 2 with 10,000 PCIe A100 GPUs, achieved performance approximating the DGX-A100 while reducing costs by half and energy consumption by 40%. We specifically engineered HFReduce to accelerate allreduce communication and implemented numerous measures to keep our Computation-Storage Integrated Network congestion-free. Through our software stack, including HaiScale, 3FS, and HAI-Platform, we achieved substantial scalability by overlapping computation and communication. Our system-oriented experience from DL training provides valuable insights to drive future advancements in AI-HPC., Comment: This is the preprint version of the paper accepted for presentation at the 2024 International Conference for High Performance Computing, Networking, Storage, and Analysis (SC'24). \c{opyright} 2024 IEEE. Personal use of this material is permitted. For other uses, permission from IEEE must be obtained. Please refer to IEEE Xplore for the final published version

Open Set Object Detection has seen rapid development recently, but it continues to pose significant challenges. Language-based methods, grappling with the substantial modal disparity between textual and visual modalities, require extensive computational resources to bridge this gap. Although integrating visual prompts into these frameworks shows promise for enhancing performance, it always comes with constraints related to textual semantics. In contrast, viusal-only methods suffer from the low-quality fusion of multiple visual prompts. In response, we introduce a strong DETR-based model, Visual Intersection Network for Open Set Object Detection (VINO), which constructs a multi-image visual bank to preserve the semantic intersections of each category across all time steps. Our innovative multi-image visual updating mechanism learns to identify the semantic intersections from various visual prompts, enabling the flexible incorporation of new information and continuous optimization of feature representations. Our approach guarantees a more precise alignment between target category semantics and region semantics, while significantly reducing pre-training time and resource demands compared to language-based methods. Furthermore, the integration of a segmentation head illustrates the broad applicability of visual intersection in various visual tasks. VINO, which requires only 7 RTX4090 GPU days to complete one epoch on the Objects365v1 dataset, achieves competitive performance on par with vision-language models on benchmarks such as LVIS and ODinW35., Comment: 7pages

This paper constructs question answering system for bridge design specification based on large language model. Three implementation schemes are tried: full fine-tuning of the Bert pretrained model, parameter-efficient fine-tuning of the Bert pretrained model, and self-built language model from scratch. Through the self-built question and answer task dataset, based on the tensorflow and keras deep learning platform framework, the model is constructed and trained to predict the start position and end position of the answer in the bridge design specification given by the user. The experimental results show that full fine-tuning of the Bert pretrained model achieves 100% accuracy in the training-dataset, validation-dataset and test-dataset, and the system can extract the answers from the bridge design specification given by the user to answer various questions of the user; While parameter-efficient fine-tuning of the Bert pretrained model and self-built language model from scratch perform well in the training-dataset, their generalization ability in the test-dataset needs to be improved. The research of this paper provides a useful reference for the development of question answering system in professional field., Comment: 10 pages, 7 figures

Existing portrait relighting methods struggle with precise control over facial shadows, particularly when faced with challenges such as handling hard shadows from directional light sources or adjusting shadows while remaining in harmony with existing lighting conditions. In many situations, completely altering input lighting is undesirable for portrait retouching applications: one may want to preserve some authenticity in the captured environment. Existing shadow editing methods typically restrict their application to just the facial region and often offer limited lighting control options, such as shadow softening or rotation. In this paper, we introduce COMPOSE: a novel shadow editing pipeline for human portraits, offering precise control over shadow attributes such as shape, intensity, and position, all while preserving the original environmental illumination of the portrait. This level of disentanglement and controllability is obtained thanks to a novel decomposition of the environment map representation into ambient light and an editable gaussian dominant light source. COMPOSE is a four-stage pipeline that consists of light estimation and editing, light diffusion, shadow synthesis, and finally shadow editing. We define facial shadows as the result of a dominant light source, encoded using our novel gaussian environment map representation. Utilizing an OLAT dataset, we have trained models to: (1) predict this light source representation from images, and (2) generate realistic shadows using this representation. We also demonstrate comprehensive and intuitive shadow editing with our pipeline. Through extensive quantitative and qualitative evaluations, we have demonstrated the robust capability of our system in shadow editing., Comment: Accepted at ECCV 2024

Semantic segmentation of night-time images holds significant importance in computer vision, particularly for applications like night environment perception in autonomous driving systems. However, existing methods tend to parse night-time images from a day-time perspective, leaving the inherent challenges in low-light conditions (such as compromised texture and deceiving matching errors) unexplored. To address these issues, we propose a novel end-to-end optimized approach, named NightFormer, tailored for night-time semantic segmentation, avoiding the conventional practice of forcibly fitting night-time images into day-time distributions. Specifically, we design a pixel-level texture enhancement module to acquire texture-aware features hierarchically with phase enhancement and amplified attention, and an object-level reliable matching module to realize accurate association matching via reliable attention in low-light environments. Extensive experimental results on various challenging benchmarks including NightCity, BDD and Cityscapes demonstrate that our proposed method performs favorably against state-of-the-art night-time semantic segmentation methods., Comment: ECCV 2024

Absorption estimation, the base of spectroscopy, is crucial for probing the composition and dynamics of matter. Conventional methods of estimation rely on coherent laser sources, and in turn suffer from inherent limitations in estimating weak absorption. Here we propose a new measurement strategy with correlated photons to determine the weak absorption by distinguishing the output with and without photons, dubbed as the on-off measurement. Our implementation within the strategy allows the estimation precision to reach the ultimate quantum limit. We demonstrate that absorption spectroscopy that incorporates quantum correlations is capable of estimating weak absorption down to a single-photon level, even in noisy environments, achieving a precision comparable to that obtained through several hundred photons in conventional absorption spectroscopy. By introducing the quantum correlations, our work avoids the occurrence of light-induced damage while breaking the classical inherent limitations in spectroscopy., Comment: 7 pages, 3 figures

Multi-modal Magnetic Resonance Imaging (MRI) is imperative for accurate brain tumor segmentation, offering indispensable complementary information. Nonetheless, the absence of modalities poses significant challenges in achieving precise segmentation. Recognizing the shared anatomical structures between mono-modal and multi-modal representations, it is noteworthy that mono-modal images typically exhibit limited features in specific regions and tissues. In response to this, we present Anatomical Consistency Distillation and Inconsistency Synthesis (ACDIS), a novel framework designed to transfer anatomical structures from multi-modal to mono-modal representations and synthesize modality-specific features. ACDIS consists of two main components: Anatomical Consistency Distillation (ACD) and Modality Feature Synthesis Block (MFSB). ACD incorporates the Anatomical Feature Enhancement Block (AFEB), meticulously mining anatomical information. Simultaneously, Anatomical Consistency ConsTraints (ACCT) are employed to facilitate the consistent knowledge transfer, i.e., the richness of information and the similarity in anatomical structure, ensuring precise alignment of structural features across mono-modality and multi-modality. Complementarily, MFSB produces modality-specific features to rectify anatomical inconsistencies, thereby compensating for missing information in the segmented features. Through validation on the BraTS2018 and BraTS2020 datasets, ACDIS substantiates its efficacy in the segmentation of brain tumors with missing MRI modalities., Comment: Accepted Paper to European Conference on Artificial Intelligence (ECAI 2024)

The Rapid and accurate identification of Gamma-Ray Bursts (GRBs) is crucial for unraveling their origins. However, current burst search algorithms frequently miss low-threshold signals or lack universality for observations. In this study, we propose a novel approach utilizing transfer learning experiment based on convolutional neural network (CNN) to establish a universal GRB identification method, which validated successfully using GECAM-B data. By employing data augmentation techniques, we enhance the diversity and quantity of the GRB sample. We develop a 1D CNN model with a multi-scale feature cross fusion module (MSCFM) to extract features from samples and perform classification. The comparative results demonstrated significant performance improvements following pre-training and transferring on a large-scale dataset. Our optimal model achieved an impressive accuracy of 96.41% on the source dataset of GECAM-B, and identified three previously undiscovered GRBs by contrast with manual analysis of GECAM-B observations. These innovative transfer learning and data augmentation methods presented in this work hold promise for applications in multi-satellite exploration scenarios characterized by limited data sets and a scarcity of labeled samples in high-energy astronomy., Comment: 17 pages, 7 figures

Retrieval-Augmented Generation (RAG) has emerged as a crucial method for addressing hallucinations in large language models (LLMs). While recent research has extended RAG models to complex noisy scenarios, these explorations often confine themselves to limited noise types and presuppose that noise is inherently detrimental to LLMs, potentially deviating from real-world retrieval environments and restricting practical applicability. In this paper, we define seven distinct noise types from a linguistic perspective and establish a Noise RAG Benchmark (NoiserBench), a comprehensive evaluation framework encompassing multiple datasets and reasoning tasks. Through empirical evaluation of eight representative LLMs with diverse architectures and scales, we reveal that these noises can be further categorized into two practical groups: noise that is beneficial to LLMs (aka beneficial noise) and noise that is harmful to LLMs (aka harmful noise). While harmful noise generally impairs performance, beneficial noise may enhance several aspects of model capabilities and overall performance. Our analysis offers insights for developing more robust, adaptable RAG solutions and mitigating hallucinations across diverse retrieval scenarios.

Active learning enhances annotation efficiency by selecting the most revealing samples for labeling, thereby reducing reliance on extensive human input. Previous methods in semantic segmentation have centered on individual pixels or small areas, neglecting the rich patterns in natural images and the power of advanced pre-trained models. To address these challenges, we propose three key contributions: Firstly, we introduce Entity-Superpixel Annotation (ESA), an innovative and efficient active learning strategy which utilizes a class-agnostic mask proposal network coupled with super-pixel grouping to capture local structural cues. Additionally, our method selects a subset of entities within each image of the target domain, prioritizing superpixels with high entropy to ensure comprehensive representation. Simultaneously, it focuses on a limited number of key entities, thereby optimizing for efficiency. By utilizing an annotator-friendly design that capitalizes on the inherent structure of images, our approach significantly outperforms existing pixel-based methods, achieving superior results with minimal queries, specifically reducing click cost by 98% and enhancing performance by 1.71%. For instance, our technique requires a mere 40 clicks for annotation, a stark contrast to the 5000 clicks demanded by conventional methods.

It has been recently demonstrated that lightcone distribution amplitudes (LCDAs) for a heavy meson, defined in the heavy quark limit, can be extracted by simulating quasi-distribution amplitudes on the lattice with a large meson momentum $P^z$. This extraction involves a two-step procedure. In the first-step, the quasi-distribution amplitudes are matched onto the QCD LCDAs in the large $P^z$ limit. In the second step, the QCD LCDAs are matched onto the desired LCDAs defined in HQET. In this work, we present the $m_H^{2n}/(P^z)^{2n}$ and $\Lambda_\text{QCD}^2/(P^z)^2$ corrections in the first step with $m_H$ being the heavy meson mass. To account for $m_H^{2n}/(P^z)^{2n}$ corrections, we employ two methods: the moment relation that can provide mass corrections to all orders, and a leading-twist projector that gives the leading-order mass corrections. For the $\Lambda_\text{QCD}^2/(P^z)^2$ corrections, we employ a renormalon model that addresses the renormalon ambiguity stemming from the divergence of the perturbative series. We adopt two parametrizations for QCD LCDAs to conduct a numerical analysis. The results suggest that these corrections are typically smaller than 20\% in most regions, and should be included in a precision analysis. These findings will serve as a valuable guide for future studies on heavy meson LCDAs, particularly in the context of Lattice QCD, where accounting for such corrections is crucial for improving the accuracy and reliability of the results., Comment: 12 pages,7 figures

Quantum error correction provides a path to reach practical quantum computing by combining multiple physical qubits into a logical qubit, where the logical error rate is suppressed exponentially as more qubits are added. However, this exponential suppression only occurs if the physical error rate is below a critical threshold. In this work, we present two surface code memories operating below this threshold: a distance-7 code and a distance-5 code integrated with a real-time decoder. The logical error rate of our larger quantum memory is suppressed by a factor of $\Lambda$ = 2.14 $\pm$ 0.02 when increasing the code distance by two, culminating in a 101-qubit distance-7 code with 0.143% $\pm$ 0.003% error per cycle of error correction. This logical memory is also beyond break-even, exceeding its best physical qubit's lifetime by a factor of 2.4 $\pm$ 0.3. We maintain below-threshold performance when decoding in real time, achieving an average decoder latency of 63 $\mu$s at distance-5 up to a million cycles, with a cycle time of 1.1 $\mu$s. To probe the limits of our error-correction performance, we run repetition codes up to distance-29 and find that logical performance is limited by rare correlated error events occurring approximately once every hour, or 3 $\times$ 10$^9$ cycles. Our results present device performance that, if scaled, could realize the operational requirements of large scale fault-tolerant quantum algorithms., Comment: 10 pages, 4 figures, Supplementary Information

This paper proposes a new paradigm: generative blockchain, which aims to transform conventional blockchain technology by combining transaction generation and recording, rather than focusing solely on transaction recording. Central to our design is a novel consensus mechanism, Proof-of-Merit (PoM), specifically crafted for environments where businesses must solve complex problems before transactions can be recorded. PoM integrates the generation and recording of transactions within a unified blockchain system, fundamentally differing from prevailing consensus mechanisms that primarily record existing transactions. We demonstrate PoM on a ride service on-demand platform, where the task of solving complex transaction-generating problems is delegated to a pool of independent problem solvers. These solvers generate transactions, and their solutions are selected based on merit. The winning solvers then register these transactions onto the blockchain and are rewarded accordingly. We introduce a Decentralized Control Parameter (DCP) to balance two key performance metrics: efficiency and equity. The applicability of our generative blockchain is illustrated through a ridesharing context, where matchers (solvers) are tasked with matching riders to drivers. We demonstrate PoM's performance and nuanced properties using agent-based simulation, exploring how to find the optimal DCP value to achieve a desirable balance of efficiency and equity in a generative blockchain.

Accurate and reliable prediction has profound implications to a wide range of applications. In this study, we focus on an instance of spatio-temporal learning problem--traffic prediction--to demonstrate an advanced deep learning model developed for making accurate and reliable forecast. Despite the significant progress in traffic prediction, limited studies have incorporated both explicit and implicit traffic patterns simultaneously to improve prediction performance. Meanwhile, the variability nature of traffic states necessitates quantifying the uncertainty of model predictions in a statistically principled way; however, extant studies offer no provable guarantee on the statistical validity of confidence intervals in reflecting its actual likelihood of containing the ground truth. In this paper, we propose an end-to-end traffic prediction framework that leverages three primary components to generate accurate and reliable traffic predictions: dynamic causal structure learning for discovering implicit traffic patterns from massive traffic data, causally-aware spatio-temporal multi-graph convolution network (CASTMGCN) for learning spatio-temporal dependencies, and conformal prediction for uncertainty quantification. CASTMGCN fuses several graphs that characterize different important aspects of traffic networks and an auxiliary graph that captures the effect of exogenous factors on the road network. On this basis, a conformal prediction approach tailored to spatio-temporal data is further developed for quantifying the uncertainty in node-wise traffic predictions over varying prediction horizons. Experimental results on two real-world traffic datasets demonstrate that the proposed method outperforms several state-of-the-art models in prediction accuracy; moreover, it generates more efficient prediction regions than other methods while strictly satisfying the statistical validity in coverage.