arXiv subjects

Lu Lu

Publications and source records attributed to Lu Lu.

At least 127 records · Page 7Linked to original sources

arXiv2024

Self-supervised Gait-based Emotion Representation Learning from Selective Strongly Augmented Skeleton Sequences

Emotion recognition is an important part of affective computing. Extracting emotional cues from human gaits yields benefits such as natural interaction, a nonintrusive nature, and remote detection. Recently, the introduction of self-supervised learning techniques offers a practical solution to the issues arising from the scarcity of labeled data in the field of gait-based emotion recognition. However, due to the limited diversity of gaits and the incompleteness of feature representations for skeletons, the existing contrastive learning methods are usually inefficient for the acquisition of gait emotions. In this paper, we propose a contrastive learning framework utilizing selective strong augmentation (SSA) for self-supervised gait-based emotion representation, which aims to derive effective representations from limited labeled gait data. First, we propose an SSA method for the gait emotion recognition task, which includes upper body jitter and random spatiotemporal mask. The goal of SSA is to generate more diverse and targeted positive samples and prompt the model to learn more distinctive and robust feature representations. Then, we design a complementary feature fusion network (CFFN) that facilitates the integration of cross-domain information to acquire topological structural and global adaptive features. Finally, we implement the distributional divergence minimization loss to supervise the representation learning of the generally and strongly augmented queries. Our approach is validated on the Emotion-Gait (E-Gait) and Emilya datasets and outperforms the state-of-the-art methods under different evaluation protocols.

cs.CV↗

arXiv2024

SALMONN: Towards Generic Hearing Abilities for Large Language Models

Hearing is arguably an essential ability of artificial intelligence (AI) agents in the physical world, which refers to the perception and understanding of general auditory information consisting of at least three types of sounds: speech, audio events, and music. In this paper, we propose SALMONN, a speech audio language music open neural network, built by integrating a pre-trained text-based large language model (LLM) with speech and audio encoders into a single multimodal model. SALMONN enables the LLM to directly process and understand general audio inputs and achieve competitive performances on a number of speech and audio tasks used in training, such as automatic speech recognition and translation, auditory-information-based question answering, emotion recognition, speaker verification, and music and audio captioning etc. SALMONN also has a diverse set of emergent abilities unseen in the training, which includes but is not limited to speech translation to untrained languages, speech-based slot filling, spoken-query-based question answering, audio-based storytelling, and speech audio co-reasoning etc. The presence of cross-modal emergent abilities is studied, and a novel few-shot activation tuning approach is proposed to activate such abilities. To our knowledge, SALMONN is the first model of its type and can be regarded as a step towards AI with generic hearing abilities. The source code, model checkpoints and data are available at https://github.com/bytedance/SALMONN.

cs.SD↗

arXiv2024

SA-SOT: Speaker-Aware Serialized Output Training for Multi-Talker ASR

Multi-talker automatic speech recognition plays a crucial role in scenarios involving multi-party interactions, such as meetings and conversations. Due to its inherent complexity, this task has been receiving increasing attention. Notably, the serialized output training (SOT) stands out among various approaches because of its simplistic architecture and exceptional performance. However, the frequent speaker changes in token-level SOT (t-SOT) present challenges for the autoregressive decoder in effectively utilizing context to predict output sequences. To address this issue, we introduce a masked t-SOT label, which serves as the cornerstone of an auxiliary training loss. Additionally, we utilize a speaker similarity matrix to refine the self-attention mechanism of the decoder. This strategic adjustment enhances contextual relationships within the same speaker's tokens while minimizing interactions between different speakers' tokens. We denote our method as speaker-aware SOT (SA-SOT). Experiments on the Librispeech datasets demonstrate that our SA-SOT obtains a relative cpWER reduction ranging from 12.75% to 22.03% on the multi-talker test sets. Furthermore, with more extensive training, our method achieves an impressive cpWER of 3.41%, establishing a new state-of-the-art result on the LibrispeechMix dataset.

cs.SD↗

arXiv2024

Conformalized-DeepONet: A Distribution-Free Framework for Uncertainty Quantification in Deep Operator Networks

In this paper, we adopt conformal prediction, a distribution-free uncertainty quantification (UQ) framework, to obtain confidence prediction intervals with coverage guarantees for Deep Operator Network (DeepONet) regression. Initially, we enhance the uncertainty quantification frameworks (B-DeepONet and Prob-DeepONet) previously proposed by the authors by using split conformal prediction. By combining conformal prediction with our Prob- and B-DeepONets, we effectively quantify uncertainty by generating rigorous confidence intervals for DeepONet prediction. Additionally, we design a novel Quantile-DeepONet that allows for a more natural use of split conformal prediction. We refer to this distribution-free effective uncertainty quantification framework as split conformal Quantile-DeepONet regression. Finally, we demonstrate the effectiveness of the proposed methods using various ordinary, partial differential equation numerical examples, and multi-fidelity learning.

cs.LG↗

arXiv2024

CoAVT: A Cognition-Inspired Unified Audio-Visual-Text Pre-Training Model for Multimodal Processing

There has been a long-standing quest for a unified audio-visual-text model to enable various multimodal understanding tasks, which mimics the listening, seeing and reading process of human beings. Humans tends to represent knowledge using two separate systems: one for representing verbal (textual) information and one for representing non-verbal (visual and auditory) information. These two systems can operate independently but can also interact with each other. Motivated by this understanding of human cognition, in this paper, we introduce CoAVT -- a novel cognition-inspired Correlated Audio-Visual-Text pre-training model to connect the three modalities. It contains a joint audio-visual encoder that learns to encode audio-visual synchronization information together with the audio and visual content for non-verbal information, and a text encoder to handle textual input for verbal information. To bridge the gap between modalities, CoAVT employs a query encoder, which contains a set of learnable query embeddings, and extracts the most informative audiovisual features of the corresponding text. Additionally, to leverage the correspondences between audio and vision with language respectively, we also establish the audio-text and visual-text bi-modal alignments upon the foundational audiovisual-text tri-modal alignment to enhance the multimodal representation learning. Finally, we jointly optimize CoAVT model with three multimodal objectives: contrastive loss, matching loss and language modeling loss. Extensive experiments show that CoAVT can learn strong multimodal correlations and be generalized to various downstream tasks. CoAVT establishes new state-of-the-art performance on text-video retrieval task on AudioCaps for both zero-shot and fine-tuning settings, audio-visual event classification and audio-visual retrieval tasks on AudioSet and VGGSound.

eess.AS↗

arXiv2024

DIMON: Learning Solution Operators of Partial Differential Equations on a Diffeomorphic Family of Domains

The solution of a PDE over varying initial/boundary conditions on multiple domains is needed in a wide variety of applications, but it is computationally expensive if the solution is computed de novo whenever the initial/boundary conditions of the domain change. We introduce a general operator learning framework, called DIffeomorphic Mapping Operator learNing (DIMON) to learn approximate PDE solutions over a family of domains $\{Ω_θ}_θ$, that learns the map from initial/boundary conditions and domain $Ω_θ$ to the solution of the PDE, or to specified functionals thereof. DIMON is based on transporting a given problem (initial/boundary conditions and domain $Ω_θ$) to a problem on a reference domain $Ω_{0}$, where training data from multiple problems is used to learn the map to the solution on $Ω_{0}$, which is then re-mapped to the original domain $Ω_θ$. We consider several problems to demonstrate the performance of the framework in learning both static and time-dependent PDEs on non-rigid geometries; these include solving the Laplace equation, reaction-diffusion equations, and a multiscale PDE that characterizes the electrical propagation on the left ventricle. This work paves the way toward the fast prediction of PDE solutions on a family of domains and the application of neural operators in engineering and precision medicine.

cs.LG↗

arXiv2024

Speeding up and reducing memory usage for scientific machine learning via mixed precision

Scientific machine learning (SciML) has emerged as a versatile approach to address complex computational science and engineering problems. Within this field, physics-informed neural networks (PINNs) and deep operator networks (DeepONets) stand out as the leading techniques for solving partial differential equations by incorporating both physical equations and experimental data. However, training PINNs and DeepONets requires significant computational resources, including long computational times and large amounts of memory. In search of computational efficiency, training neural networks using half precision (float16) rather than the conventional single (float32) or double (float64) precision has gained substantial interest, given the inherent benefits of reduced computational time and memory consumed. However, we find that float16 cannot be applied to SciML methods, because of gradient divergence at the start of training, weight updates going to zero, and the inability to converge to a local minima. To overcome these limitations, we explore mixed precision, which is an approach that combines the float16 and float32 numerical formats to reduce memory usage and increase computational speed. Our experiments showcase that mixed precision training not only substantially decreases training times and memory demands but also maintains model accuracy. We also reinforce our empirical observations with a theoretical analysis. The research has broad implications for SciML in various computational applications.

cs.LG↗

arXiv2024

Multi-resolution partial differential equations preserved learning framework for spatiotemporal dynamics

Traditional data-driven deep learning models often struggle with high training costs, error accumulation, and poor generalizability in complex physical processes. Physics-informed deep learning (PiDL) addresses these challenges by incorporating physical principles into the model. Most PiDL approaches regularize training by embedding governing equations into the loss function, yet this depends heavily on extensive hyperparameter tuning to weigh each loss term. To this end, we propose to leverage physics prior knowledge by ``baking'' the discretized governing equations into the neural network architecture via the connection between the partial differential equations (PDE) operators and network structures, resulting in a PDE-preserved neural network (PPNN). This method, embedding discretized PDEs through convolutional residual networks in a multi-resolution setting, largely improves the generalizability and long-term prediction accuracy, outperforming conventional black-box models. The effectiveness and merit of the proposed methods have been demonstrated across various spatiotemporal dynamical systems governed by spatiotemporal PDEs, including reaction-diffusion, Burgers', and Navier-Stokes equations.

cs.LG↗

arXiv2024

Theoretical efficiency limit of diffractive input couplers in augmented reality waveguides

Considerable efforts have been devoted into augmented reality (AR) displays to enable the immersive user experience in the wearable glasses form factor. Transparent waveguide combiners offer a compact solution to guide light from the microdisplay to the front of eyes while maintaining the see-through optical path to view the real world simultaneously. To deliver a realistic virtual image with low power consumption, the waveguide combiners need to have high efficiency and good image quality. One important limiting factor for the efficiency of diffractive waveguide combiners is the out-coupling problem in the input couplers, where the guided light interacts with the input gratings again and get partially out-coupled. In this study, we introduce a theoretical model to deterministically find the upper bound of the input efficiency of a uniform input grating. Our model considers the polarization management at the input coupler and can work for arbitrary input polarization state ensemble. Our model also provides the corresponding characteristics of the input coupler, such as the grating diffraction efficiencies and the Jones matrix of the polarization management components, to achieve the optimal input efficiency. Equipped with this theoretical model, we investigate how the upper bound of input efficiency varies with geometric parameters including the waveguide thickness, the projector pupil size, and the projector pupil relief distance. Our study shines light on the fundamental efficiency limits on input couplers in diffractive waveguide combiners and highlights the benefits of polarization control in improving the input efficiency.

physics.optics↗

arXiv2024

A unified multichannel far-field speech recognition system: combining neural beamforming with attention based end-to-end model

Far-field speech recognition is a challenging task that conventionally uses signal processing beamforming to attack noise and interference problem. But the performance has been found usually limited due to heavy reliance on environmental assumption. In this paper, we propose a unified multichannel far-field speech recognition system that combines the neural beamforming and transformer-based Listen, Spell, Attend (LAS) speech recognition system, which extends the end-to-end speech recognition system further to include speech enhancement. Such framework is then jointly trained to optimize the final objective of interest. Specifically, factored complex linear projection (fCLP) has been adopted to form the neural beamforming. Several pooling strategies to combine look directions are then compared in order to find the optimal approach. Moreover, information of the source direction is also integrated in the beamforming to explore the usefulness of source direction as a prior, which is usually available especially in multi-modality scenario. Experiments on different microphone array geometry are conducted to evaluate the robustness against spacing variance of microphone array. Large in-house databases are used to evaluate the effectiveness of the proposed framework and the proposed method achieve 19.26\% improvement when compared with a strong baseline.

eess.AS↗

arXiv2023

A graph discretization of vector Laplace operator

In this paper, we study the graph-theoretic analogues of vector Laplacian (or Helmholtz operator) and vector Laplace equation. We determine the graph matrix representation of vector Laplacian and obtain the dimension of solution space of vector Laplace equation on graphs.

math.CO↗

arXiv2023

Planning Reliability Assurance Tests for Autonomous Vehicles

Artificial intelligence (AI) technology has become increasingly prevalent and transforms our everyday life. One important application of AI technology is the development of autonomous vehicles (AV). However, the reliability of an AV needs to be carefully demonstrated via an assurance test so that the product can be used with confidence in the field. To plan for an assurance test, one needs to determine how many AVs need to be tested for how many miles and the standard for passing the test. Existing research has made great efforts in developing reliability demonstration tests in the other fields of applications for product development and assessment. However, statistical methods have not been utilized in AV test planning. This paper aims to fill in this gap by developing statistical methods for planning AV reliability assurance tests based on recurrent events data. We explore the relationship between multiple criteria of interest in the context of planning AV reliability assurance tests. Specifically, we develop two test planning strategies based on homogeneous and non-homogeneous Poisson processes while balancing multiple objectives with the Pareto front approach. We also offer recommendations for practical use. The disengagement events data from the California Department of Motor Vehicles AV testing program is used to illustrate the proposed assurance test planning methods.

stat.AP↗

arXiv2023

PF-DMD: Physics-fusion dynamic mode decomposition for accurate and robust forecasting of dynamical systems with imperfect data and physics

The DMD (Dynamic Mode Decomposition) method has attracted widespread attention as a representative modal-decomposition method and can build a predictive model. However, the DMD may give predicted results that deviate from physical reality in some scenarios, such as dealing with translation problems or noisy data. Therefore, this paper proposes a physics-fusion dynamic mode decomposition (PFDMD) method to address this issue. The proposed PFDMD method first obtains a data-driven model using DMD, then calculates the residual of the physical equations, and finally corrects the predicted results using Kalman filtering and gain coefficients. In this way, the PFDMD method can integrate the physics-informed equations with the data-driven model generated by DMD. Numerical experiments are conducted using the PFDMD, including the Allen-Cahn, advection-diffusion, and Burgers' equations. The results demonstrate that the proposed PFDMD method can significantly reduce the reconstruction and prediction errors by incorporating physics-informed equations, making it usable for translation and shock problems where the standard DMD method has failed.

physics.comp-ph↗

arXiv2023

Improving Large-scale Deep Biasing with Phoneme Features and Text-only Data in Streaming Transducer

Deep biasing for the Transducer can improve the recognition performance of rare words or contextual entities, which is essential in practical applications, especially for streaming Automatic Speech Recognition (ASR). However, deep biasing with large-scale rare words remains challenging, as the performance drops significantly when more distractors exist and there are words with similar grapheme sequences in the bias list. In this paper, we combine the phoneme and textual information of rare words in Transducers to distinguish words with similar pronunciation or spelling. Moreover, the introduction of training with text-only data containing more rare words benefits large-scale deep biasing. The experiments on the LibriSpeech corpus demonstrate that the proposed method achieves state-of-the-art performance on rare word error rate for different scales and levels of bias lists.

cs.CL↗

arXiv2023

Distance-regular Cayley graphs over (pseudo-) semi-dihedral groups

Distance-regular graphs are a class of regualr graphs with pretty combinatorial symmetry. In 2007, Miklavič and Potočnik proposed the problem of charaterizing distance-regular Cayley graphs, which can be viewed as a natural extension of the problem of characterizing strongly-regular Cayley graphs (or equivalently, regular partial difference sets). In this paper, we provide a partial characterization for distance-regular Cayley graphs over semi-dihedral groups and pseudo-semi-dihedral groups, both of which are $2$-groups with a cyclic subgroup of index $2$.

math.CO↗

arXiv2023

Nonparametric Estimation of Conditional Copula using Smoothed Checkerboard Bernstein Sieves

Conditional copulas are useful tools for modeling the dependence between multiple response variables that may vary with a given set of predictor variables. Conditional dependence measures such as conditional Kendall's tau and Spearman's rho that can be expressed as functionals of the conditional copula are often used to evaluate the strength of dependence conditioning on the covariates. In general, semiparametric estimation methods of conditional copulas rely on an assumed parametric copula family where the copula parameter is assumed to be a function of the covariates. The functional relationship can be estimated nonparametrically using different techniques but it is required to choose an appropriate copula model from various candidate families. In this paper, by employing the empirical checkerboard Bernstein copula (ECBC) estimator we propose a fully nonparametric approach for estimating conditional copulas, which doesn't require any selection of parametric copula models. Closed-form estimates of the conditional dependence measures are derived directly from the proposed ECBC-based conditional copula estimator. We provide the large-sample consistency of the proposed estimator as well as the estimates of conditional dependence measures. The finite-sample performance of the proposed estimator and comparison with semiparametric methods are investigated through simulation studies. An application to real case studies is also provided.

stat.ME↗

arXiv2023

Spectral radius of graphs forbidden $C_7$ or $C_6^{\triangle}$

Let $C_k^{\triangle}$ be the graph obtained from a cycle $C_{k}$ by adding a new vertex connecting two adjacent vertices in $C_{k}$. In this note, we obtain the graph maximizing the spectral radius among all graphs with size $m$ and containing no subgraph isomorphic to $C_6^{\triangle}$. As a byproduct, we will show that if the spectral radius $λ(G)\ge1+\sqrt{m-2}$, then $G$ must contains all the cycles $C_i$ for $3\le i\le 7$ unless $G\cong K_3\nabla \left(\frac{m-3}{3}K_1\right)$.

math.CO↗

arXiv2023

ROAM: memory-efficient large DNN training via optimized operator ordering and memory layout

As deep learning models continue to increase in size, the memory requirements for training have surged. While high-level techniques like offloading, recomputation, and compression can alleviate memory pressure, they also introduce overheads. However, a memory-efficient execution plan that includes a reasonable operator execution order and tensor memory layout can significantly increase the models' memory efficiency and reduce overheads from high-level techniques. In this paper, we propose ROAM which operates on computation graph level to derive memory-efficient execution plan with optimized operator order and tensor memory layout for models. We first propose sophisticated theories that carefully consider model structure and training memory load to support optimization for large complex graphs that have not been well supported in the past. An efficient tree-based algorithm is further proposed to search task divisions automatically, along with delivering high performance and effectiveness to solve the problem. Experiments show that ROAM achieves a substantial memory reduction of 35.7%, 13.3%, and 27.2% compared to Pytorch and two state-of-the-art methods and offers a remarkable 53.7x speedup. The evaluation conducted on the expansive GPT2-XL further validates ROAM's scalability.

cs.LG↗