arXiv subjects

Wei Li

Publications and source records attributed to Wei Li.

At least 703 records · Page 39Linked to original sources

arXiv2023

Video-Helpful Multimodal Machine Translation

Existing multimodal machine translation (MMT) datasets consist of images and video captions or instructional video subtitles, which rarely contain linguistic ambiguity, making visual information ineffective in generating appropriate translations. Recent work has constructed an ambiguous subtitles dataset to alleviate this problem but is still limited to the problem that videos do not necessarily contribute to disambiguation. We introduce EVA (Extensive training set and Video-helpful evaluation set for Ambiguous subtitles translation), an MMT dataset containing 852k Japanese-English (Ja-En) parallel subtitle pairs, 520k Chinese-English (Zh-En) parallel subtitle pairs, and corresponding video clips collected from movies and TV episodes. In addition to the extensive training set, EVA contains a video-helpful evaluation set in which subtitles are ambiguous, and videos are guaranteed helpful for disambiguation. Furthermore, we propose SAFA, an MMT model based on the Selective Attention model with two novel methods: Frame attention loss and Ambiguity augmentation, aiming to use videos in EVA for disambiguation fully. Experiments on EVA show that visual information and the proposed methods can boost translation performance, and our model performs significantly better than existing MMT models. The EVA dataset and the SAFA model are available at: https://github.com/ku-nlp/video-helpful-MMT.git.

cs.CL↗

arXiv2023

Proximity effect induced intriguing superconductivity in van der Waals heterostructure of magnetic topological insulator and conventional superconductor

Nontrivial topological superconductivity has received enormous research attentions due to its potential for diverse applications in topological quantum computing. The intrinsic issue concerning the correlation between a topological insulator and a superconductor is, however, still widely open. Here, we systemically report an emergent superconductivity in a cross-junction composed of a magnetic topological insulator MnBi2Te4 and a conventional superconductor NbSe2. Remarkably, the interface indicates existence of a reduced superconductivity at surface of NbSe2 and a proximity-effectinduced superconductivity at surface of MnBi2Te4. Furthermore, the in-plane angular-dependent magnetoresistance measurements reveal the fingerprints of the paring symmetry behaviors for these superconducting gaps as a unconventional nature. Our findings extend our views and ideas of topological superconductivity in the superconducting heterostructures with time-reversal symmetry breaking, offering an exciting opportunity to elucidate the cooperative effects on the surface state of a topological insulator aligning a superconductor.

cond-mat.supr-con↗

arXiv2023

Improving Anomaly Segmentation with Multi-Granularity Cross-Domain Alignment

Anomaly segmentation plays a pivotal role in identifying atypical objects in images, crucial for hazard detection in autonomous driving systems. While existing methods demonstrate noteworthy results on synthetic data, they often fail to consider the disparity between synthetic and real-world data domains. Addressing this gap, we introduce the Multi-Granularity Cross-Domain Alignment (MGCDA) framework, tailored to harmonize features across domains at both the scene and individual sample levels. Our contributions are twofold: i) We present the Multi-source Domain Adversarial Training module. This integrates a multi-source adversarial loss coupled with dynamic label smoothing, facilitating the learning of domain-agnostic representations across multiple processing stages. ii) We propose an innovative Cross-domain Anomaly-aware Contrastive Learning methodology.} This method adeptly selects challenging anchor points and images using an anomaly-centric strategy, ensuring precise alignment at the sample level. Extensive evaluations of the Fishyscapes and RoadAnomaly datasets demonstrate MGCDA's superior performance and adaptability. Additionally, its ability to perform parameter-free inference and function with various network architectures highlights its distinctiveness in advancing the frontier of anomaly segmentation.

cs.CV↗

arXiv2023

Cross-head mutual Mean-Teaching for semi-supervised medical image segmentation

Semi-supervised medical image segmentation (SSMIS) has witnessed substantial advancements by leveraging limited labeled data and abundant unlabeled data. Nevertheless, existing state-of-the-art (SOTA) methods encounter challenges in accurately predicting labels for the unlabeled data, giving rise to disruptive noise during training and susceptibility to erroneous information overfitting. Moreover, applying perturbations to inaccurate predictions further reduces consistent learning. To address these concerns, we propose a novel Cross-head mutual mean-teaching Network (CMMT-Net) incorporated strong-weak data augmentation, thereby benefitting both self-training and consistency learning. Specifically, our CMMT-Net consists of both teacher-student peer networks with a share encoder and dual slightly different decoders, and the pseudo labels generated by one mean teacher head are adopted to supervise the other student branch to achieve a mutual consistency. Furthermore, we propose mutual virtual adversarial training (MVAT) to smooth the decision boundary and enhance feature representations. To diversify the consistency training samples, we employ Cross-Set CutMix strategy, which also helps address distribution mismatch issues. Notably, CMMT-Net simultaneously implements data, feature, and network perturbations, amplifying model diversity and generalization performance. Experimental results on three publicly available datasets indicate that our approach yields remarkable improvements over previous SOTA methods across various semi-supervised scenarios. Code and logs will be available at https://github.com/Leesoon1984/CMMT-Net.

cs.CV↗

arXiv2023

MERTech: Instrument Playing Technique Detection Using Self-Supervised Pretrained Model With Multi-Task Finetuning

Instrument playing techniques (IPTs) constitute a pivotal component of musical expression. However, the development of automatic IPT detection methods suffers from limited labeled data and inherent class imbalance issues. In this paper, we propose to apply a self-supervised learning model pre-trained on large-scale unlabeled music data and finetune it on IPT detection tasks. This approach addresses data scarcity and class imbalance challenges. Recognizing the significance of pitch in capturing the nuances of IPTs and the importance of onset in locating IPT events, we investigate multi-task finetuning with pitch and onset detection as auxiliary tasks. Additionally, we apply a post-processing approach for event-level prediction, where an IPT activation initiates an event only if the onset output confirms an onset in that frame. Our method outperforms prior approaches in both frame-level and event-level metrics across multiple IPT benchmark datasets. Further experiments demonstrate the efficacy of multi-task finetuning on each IPT class.

cs.SD↗

arXiv2023

Cross-Domain Policy Adaptation via Value-Guided Data Filtering

Generalizing policies across different domains with dynamics mismatch poses a significant challenge in reinforcement learning. For example, a robot learns the policy in a simulator, but when it is deployed in the real world, the dynamics of the environment may be different. Given the source and target domain with dynamics mismatch, we consider the online dynamics adaptation problem, in which case the agent can access sufficient source domain data while online interactions with the target domain are limited. Existing research has attempted to solve the problem from the dynamics discrepancy perspective. In this work, we reveal the limitations of these methods and explore the problem from the value difference perspective via a novel insight on the value consistency across domains. Specifically, we present the Value-Guided Data Filtering (VGDF) algorithm, which selectively shares transitions from the source domain based on the proximity of paired value targets across the two domains. Empirical results on various environments with kinematic and morphology shifts demonstrate that our method achieves superior performance compared to prior approaches.

cs.LG↗

arXiv2023

Fine-grained Audio-Visual Joint Representations for Multimodal Large Language Models

Audio-visual large language models (LLM) have drawn significant attention, yet the fine-grained combination of both input streams is rather under-explored, which is challenging but necessary for LLMs to understand general video inputs. To this end, a fine-grained audio-visual joint representation (FAVOR) learning framework for multimodal LLMs is proposed in this paper, which extends a text-based LLM to simultaneously perceive speech and audio events in the audio input stream and images or videos in the visual input stream, at the frame level. To fuse the audio and visual feature streams into joint representations and to align the joint space with the LLM input embedding space, we propose a causal Q-Former structure with a causal attention module to enhance the capture of causal relations of the audio-visual frames across time. An audio-visual evaluation benchmark (AVEB) is also proposed which comprises six representative single-modal tasks with five cross-modal tasks reflecting audio-visual co-reasoning abilities. While achieving competitive single-modal performance on audio, speech and image tasks in AVEB, FAVOR achieved over 20% accuracy improvements on the video question-answering task when fine-grained information or temporal causal reasoning is required. FAVOR, in addition, demonstrated remarkable video comprehension and reasoning abilities on tasks that are unprecedented by other multimodal LLMs. An interactive demo of FAVOR is available at https://github.com/BriansIDP/AudioVisualLLM.git, and the training code and model checkpoints will be released soon.

eess.AS↗

arXiv2023

Finite element interpolated neural networks for solving forward and inverse problems

We propose a general framework for solving forward and inverse problems constrained by partial differential equations, where we interpolate neural networks onto finite element spaces to represent the (partial) unknowns. The framework overcomes the challenges related to the imposition of boundary conditions, the choice of collocation points in physics-informed neural networks, and the integration of variational physics-informed neural networks. A numerical experiment set confirms the framework's capability of handling various forward and inverse problems. In particular, the trained neural network generalises well for smooth problems, beating finite element solutions by some orders of magnitude. We finally propose an effective one-loop solver with an initial data fitting step (to obtain a cheap initialisation) to solve inverse problems.

math.NA↗

arXiv2023

Jet tagging algorithm of graph network with HaarPooling message passing

Recently methods of graph neural networks (GNNs) have been applied to solving the problems in high energy physics (HEP) and have shown its great potential for quark-gluon tagging with graph representation of jet events. In this paper, we introduce an approach of GNNs combined with a HaarPooling operation to analyze the events, called HaarPooling Message Passing neural network (HMPNet). In HMPNet, HaarPooling not only extracts the features of graph, but embeds additional information obtained by clustering of k-means of different particle features. We construct Haarpooling from five different features: absolute energy $\log E$, transverse momentum $\log p_T$, relative coordinates $(Δη,Δϕ)$, the mixed ones $(\log E, \log p_T)$ and $(\log E, \log p_T, Δη,Δϕ)$. The results show that an appropriate selection of information for HaarPooling enhances the accuracy of quark-gluon tagging, as adding extra information of $\log P_T$ to the HMPNet outperforms all the others, whereas adding relative coordinates information $(Δη,Δϕ)$ is not very effective. This implies that by adding effective particle features from HaarPooling can achieve much better results than solely pure message passing neutral network (MPNN) can do, which demonstrates significant improvement of feature extraction via the pooling process. Finally we compare the HMPNet study, ordering by $p_T$, with other studies and prove that the HMPNet is also a good choice of GNN algorithms for jet tagging.

hep-ex↗

arXiv2023

IFT: Image Fusion Transformer for Ghost-free High Dynamic Range Imaging

Multi-frame high dynamic range (HDR) imaging aims to reconstruct ghost-free images with photo-realistic details from content-complementary but spatially misaligned low dynamic range (LDR) images. Existing HDR algorithms are prone to producing ghosting artifacts as their methods fail to capture long-range dependencies between LDR frames with large motion in dynamic scenes. To address this issue, we propose a novel image fusion transformer, referred to as IFT, which presents a fast global patch searching (FGPS) module followed by a self-cross fusion module (SCF) for ghost-free HDR imaging. The FGPS searches the patches from supporting frames that have the closest dependency to each patch of the reference frame for long-range dependency modeling, while the SCF conducts intra-frame and inter-frame feature fusion on the patches obtained by the FGPS with linear complexity to input resolution. By matching similar patches between frames, objects with large motion ranges in dynamic scenes can be aligned, which can effectively alleviate the generation of artifacts. In addition, the proposed FGPS and SCF can be integrated into various deep HDR methods as efficient plug-in modules. Extensive experiments on multiple benchmarks show that our method achieves state-of-the-art performance both quantitatively and qualitatively.

cs.CV↗

arXiv2023

Proximal Causal Inference without Uniqueness Assumptions

We consider identification and inference about a counterfactual outcome mean when there is unmeasured confounding using tools from proximal causal inference (Miao et al. [2018], Tchetgen Tchetgen et al. [2020]). Proximal causal inference requires existence of solutions to at least one of two integral equations. We motivate the existence of solutions to the integral equations from proximal causal inference by demonstrating that, assuming the existence of a solution to one of the integral equations, $\sqrt{n}$-estimability of a linear functional (such as its mean) of that solution requires the existence of a solution to the other integral equation. Solutions to the integral equations may not be unique, which complicates estimation and inference. We construct a consistent estimator for the solution set for one of the integral equations and then adapt the theory of extremum estimators to find from the estimated set a consistent estimator for a uniquely defined solution. A debiased estimator for the counterfactual mean is shown to be root-$n$ consistent, regular, and asymptotically semiparametrically locally efficient under additional regularity conditions.

stat.ME↗

arXiv2023

Zero-shot point cloud segmentation by transferring geometric primitives

We investigate transductive zero-shot point cloud semantic segmentation, where the network is trained on seen objects and able to segment unseen objects. The 3D geometric elements are essential cues to imply a novel 3D object type. However, previous methods neglect the fine-grained relationship between the language and the 3D geometric elements. To this end, we propose a novel framework to learn the geometric primitives shared in seen and unseen categories' objects and employ a fine-grained alignment between language and the learned geometric primitives. Therefore, guided by language, the network recognizes the novel objects represented with geometric primitives. Specifically, we formulate a novel point visual representation, the similarity vector of the point's feature to the learnable prototypes, where the prototypes automatically encode geometric primitives via back-propagation. Besides, we propose a novel Unknown-aware InfoNCE Loss to fine-grained align the visual representation with language. Extensive experiments show that our method significantly outperforms other state-of-the-art methods in the harmonic mean-intersection-over-union (hIoU), with the improvement of 17.8\%, 30.4\%, 9.2\% and 7.9\% on S3DIS, ScanNet, SemanticKITTI and nuScenes datasets, respectively. Codes are available (https://github.com/runnanchen/Zero-Shot-Point-Cloud-Segmentation)

cs.CV↗

arXiv2023

Model2Scene: Learning 3D Scene Representation via Contrastive Language-CAD Models Pre-training

Current successful methods of 3D scene perception rely on the large-scale annotated point cloud, which is tedious and expensive to acquire. In this paper, we propose Model2Scene, a novel paradigm that learns free 3D scene representation from Computer-Aided Design (CAD) models and languages. The main challenges are the domain gaps between the CAD models and the real scene's objects, including model-to-scene (from a single model to the scene) and synthetic-to-real (from synthetic model to real scene's object). To handle the above challenges, Model2Scene first simulates a crowded scene by mixing data-augmented CAD models. Next, we propose a novel feature regularization operation, termed Deep Convex-hull Regularization (DCR), to project point features into a unified convex hull space, reducing the domain gap. Ultimately, we impose contrastive loss on language embedding and the point features of CAD models to pre-train the 3D network. Extensive experiments verify the learned 3D scene representation is beneficial for various downstream tasks, including label-free 3D object salient detection, label-efficient 3D scene perception and zero-shot 3D semantic segmentation. Notably, Model2Scene yields impressive label-free 3D object salient detection with an average mAP of 46.08\% and 55.49\% on the ScanNet and S3DIS datasets, respectively. The code will be publicly available.

cs.CV↗

arXiv2023

Optimization on the smallest eigenvalue of grounded Laplacian matrix via edge addition

The grounded Laplacian matrix $\LL_{-S}$ of a graph $\calG=(V,E)$ with $n=|V|$ nodes and $m=|E|$ edges is a $(n-s)\times (n-s)$ submatrix of its Laplacian matrix $\LL$, obtained from $\LL$ by deleting rows and columns corresponding to $s=|S| \ll n $ ground nodes forming set $S\subset V$. The smallest eigenvalue of $\LL_{-S}$ plays an important role in various practical scenarios, such as characterizing the convergence rate of leader-follower opinion dynamics, with a larger eigenvalue indicating faster convergence of opinion. In this paper, we study the problem of adding $k \ll n$ edges among all the nonexistent edges forming the candidate edge set $Q = (V\times V)\backslash E$, in order to maximize the smallest eigenvalue of the grounded Laplacian matrix. We show that the objective function of the combinatorial optimization problem is monotone but non-submodular. To solve the problem, we first simplify the problem by restricting the candidate edge set $Q$ to be $(S\times (V\backslash S))\backslash E$, and prove that it has the same optimal solution as the original problem, although the size of set $Q$ is reduced from $O(n^2)$ to $O(n)$. Then, we propose two greedy approximation algorithms. One is a simple greedy algorithm with an approximation ratio $(1-e^{-αγ})/α$ and time complexity $O(kn^4)$, where $γ$ and $α$ are, respectively, submodularity ratio and curvature, whose bounds are provided for some particular cases. The other is a fast greedy algorithm without approximation guarantee, which has a running time $\tilde{O}(km)$, where $\tilde{O}(\cdot)$ suppresses the ${\rm poly} (\log n)$ factors. Numerous experiments on various real networks are performed to validate the superiority of our algorithms, in terms of effectiveness and efficiency.

cs.MA↗

arXiv2023

MiChao-HuaFen 1.0: A Specialized Pre-trained Corpus Dataset for Domain-specific Large Models

With the advancement of deep learning technologies, general-purpose large models such as GPT-4 have demonstrated exceptional capabilities across various domains. Nevertheless, there remains a demand for high-quality, domain-specific outputs in areas like healthcare, law, and finance. This paper first evaluates the existing large models for specialized domains and discusses their limitations. To cater to the specific needs of certain domains, we introduce the ``MiChao-HuaFen 1.0'' pre-trained corpus dataset, tailored for the news and governmental sectors. The dataset, sourced from publicly available internet data from 2022, underwent multiple rounds of cleansing and processing to ensure high quality and reliable origins, with provisions for consistent and stable updates. This dataset not only supports the pre-training of large models for Chinese vertical domains but also aids in propelling deep learning research and applications in related fields.

cs.CL↗

arXiv2023

Connecting Speech Encoder and Large Language Model for ASR

The impressive capability and versatility of large language models (LLMs) have aroused increasing attention in automatic speech recognition (ASR), with several pioneering studies attempting to build integrated ASR models by connecting a speech encoder with an LLM. This paper presents a comparative study of three commonly used structures as connectors, including fully connected layers, multi-head cross-attention, and Q-Former. Speech encoders from the Whisper model series as well as LLMs from the Vicuna model series with different model sizes were studied. Experiments were performed on the commonly used LibriSpeech, Common Voice, and GigaSpeech datasets, where the LLMs with Q-Formers demonstrated consistent and considerable word error rate (WER) reductions over LLMs with other connector structures. Q-Former-based LLMs can generalise well to out-of-domain datasets, where 12% relative WER reductions over the Whisper baseline ASR model were achieved on the Eval2000 test set without using any in-domain training data from Switchboard. Moreover, a novel segment-level Q-Former is proposed to enable LLMs to recognise speech segments with a duration exceeding the limitation of the encoders, which results in 17% relative WER reductions over other connector structures on 90-second-long speech data.

eess.AS↗

arXiv2023

Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts all text-based language problems into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled data sets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new ``Colossal Clean Crawled Corpus'', we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our data set, pre-trained models, and code.

cs.LG↗

arXiv2023

The Potential of a TeV-Scale Muon-Ion Collider

We propose the development of a novel muon-proton and muon-nucleus collider facility at the TeV scale that is capable of performing precision deep inelastic scattering measurements in new regimes and providing a rich program in nuclear and particle physics. Such a facility could seed, or leverage, the development of a muon-antimuon collider and make use of the existing hadron accelerator infrastructure when sited at a facility such as Brookhaven National Laboratory, Fermilab, or CERN. We discuss the possible energy and luminosity design parameters for several collider configurations, and illustrate the science potential with several studies on deep inelastic scattering kinematics, Higgs and vector boson production, top quark production, and beyond Standard Model leptoquark production. Detector design considerations and a possible road map toward development are also given.

hep-ex↗