Searcharxiv⌕ Search

arXiv subjects

Yi Zhou

Publications and source records attributed to Yi Zhou.

At least 325 records · Page 18Linked to original sources
arXiv2023

Spatio-Temporal Calibration for Omni-Directional Vehicle-Mounted Event Cameras

We present a solution to the problem of spatio-temporal calibration for event cameras mounted on an onmi-directional vehicle. Different from traditional methods that typically determine the camera's pose with respect to the vehicle's body frame using alignment of trajectories, our approach leverages the kinematic correlation of two sets of linear velocity estimates from event data and wheel odometers, respectively. The overall calibration task consists of estimating the underlying temporal offset between the two heterogeneous sensors, and furthermore, recovering the extrinsic rotation that defines the linear relationship between the two sets of velocity estimates. The first sub-problem is formulated as an optimization one, which looks for the optimal temporal offset that maximizes a correlation measurement invariant to arbitrary linear transformation. Once the temporal offset is compensated, the extrinsic rotation can be worked out with an iterative closed-form solver that incrementally registers associated linear velocity estimates. The proposed algorithm is proved effective on both synthetic data and real data, outperforming traditional methods based on alignment of trajectories.

cs.RO↗
arXiv2023

Generalized-Smooth Nonconvex Optimization is As Efficient As Smooth Nonconvex Optimization

Various optimal gradient-based algorithms have been developed for smooth nonconvex optimization. However, many nonconvex machine learning problems do not belong to the class of smooth functions and therefore the existing algorithms are sub-optimal. Instead, these problems have been shown to satisfy certain generalized-smooth conditions, which have not been well understood in the existing literature. In this paper, we propose a notion of $α$-symmetric generalized-smoothness that extends the existing notions and covers many important functions such as high-order polynomials and exponential functions. We study the fundamental properties and establish descent lemmas for the functions in this class. Then, to solve such a large class of nonconvex problems, we design a special deterministic normalized gradient descent algorithm that achieves the optimal iteration complexity $\mathcal{O}(ε^{-2})$, and also prove that the popular SPIDER variance reduction algorithm achieves the optimal sample complexity $\mathcal{O}(ε^{-3})$ in the stochastic setting. Our results show that solving generalized-smooth nonconvex problems is as efficient as solving smooth nonconvex problems.

math.OC↗
arXiv2023

Fast Complementary Dynamics via Skinning Eigenmodes

We propose a reduced-space elasto-dynamic solver that is well suited for augmenting rigged character animations with secondary motion. At the core of our method is a novel deformation subspace based on Linear Blend Skinning that overcomes many of the shortcomings prior subspace methods face. Our skinning subspace is parameterized entirely by a set of scalar weights, which we can obtain through a small, material-aware and rig-sensitive generalized eigenvalue problem. The resulting subspace can easily capture rotational motion and guarantees that the resulting simulation is rotation equivariant. We further propose a simple local-global solver for linear co-rotational elasticity and propose a clustering method to aggregate per-tetrahedra non-linear energetic quantities. The result is a compact simulation that is fully decoupled from the complexity of the mesh.

cs.GR↗
arXiv2023

Multi-scale Transformer Network with Edge-aware Pre-training for Cross-Modality MR Image Synthesis

Cross-modality magnetic resonance (MR) image synthesis can be used to generate missing modalities from given ones. Existing (supervised learning) methods often require a large number of paired multi-modal data to train an effective synthesis model. However, it is often challenging to obtain sufficient paired data for supervised training. In reality, we often have a small number of paired data while a large number of unpaired data. To take advantage of both paired and unpaired data, in this paper, we propose a Multi-scale Transformer Network (MT-Net) with edge-aware pre-training for cross-modality MR image synthesis. Specifically, an Edge-preserving Masked AutoEncoder (Edge-MAE) is first pre-trained in a self-supervised manner to simultaneously perform 1) image imputation for randomly masked patches in each image and 2) whole edge map estimation, which effectively learns both contextual and structural information. Besides, a novel patch-wise loss is proposed to enhance the performance of Edge-MAE by treating different masked patches differently according to the difficulties of their respective imputations. Based on this proposed pre-training, in the subsequent fine-tuning stage, a Dual-scale Selective Fusion (DSF) module is designed (in our MT-Net) to synthesize missing-modality images by integrating multi-scale features extracted from the encoder of the pre-trained Edge-MAE. Further, this pre-trained encoder is also employed to extract high-level features from the synthesized image and corresponding ground-truth image, which are required to be similar (consistent) in the training. Experimental results show that our MT-Net achieves comparable performance to the competing methods even using $70\%$ of all available paired data. Our code will be publicly available at https://github.com/lyhkevin/MT-Net.

eess.IV↗
arXiv2023

Cross-view Geo-localization via Learning Disentangled Geometric Layout Correspondence

Cross-view geo-localization aims to estimate the location of a query ground image by matching it to a reference geo-tagged aerial images database. As an extremely challenging task, its difficulties root in the drastic view changes and different capturing time between two views. Despite these difficulties, recent works achieve outstanding progress on cross-view geo-localization benchmarks. However, existing methods still suffer from poor performance on the cross-area benchmarks, in which the training and testing data are captured from two different regions. We attribute this deficiency to the lack of ability to extract the spatial configuration of visual feature layouts and models' overfitting on low-level details from the training set. In this paper, we propose GeoDTR which explicitly disentangles geometric information from raw features and learns the spatial correlations among visual features from aerial and ground pairs with a novel geometric layout extractor module. This module generates a set of geometric layout descriptors, modulating the raw features and producing high-quality latent representations. In addition, we elaborate on two categories of data augmentations, (i) Layout simulation, which varies the spatial configuration while keeping the low-level details intact. (ii) Semantic augmentation, which alters the low-level details and encourages the model to capture spatial configurations. These augmentations help to improve the performance of the cross-view geo-localization models, especially on the cross-area benchmarks. Moreover, we propose a counterfactual-based learning process to benefit the geometric layout extractor in exploring spatial information. Extensive experiments show that GeoDTR not only achieves state-of-the-art results but also significantly boosts the performance on same-area and cross-area benchmarks.

cs.CV↗
arXiv2023

Multi-Loss Convolutional Network with Time-Frequency Attention for Speech Enhancement

The Dual-Path Convolution Recurrent Network (DPCRN) was proposed to effectively exploit time-frequency domain information. By combining the DPRNN module with Convolution Recurrent Network (CRN), the DPCRN obtained a promising performance in speech separation with a limited model size. In this paper, we explore self-attention in the DPCRN module and design a model called Multi-Loss Convolutional Network with Time-Frequency Attention(MNTFA) for speech enhancement. We use self-attention modules to exploit the long-time information, where the intra-chunk self-attentions are used to model the spectrum pattern and the inter-chunk self-attention are used to model the dependence between consecutive frames. Compared to DPRNN, axial self-attention greatly reduces the need for memory and computation, which is more suitable for long sequences of speech signals. In addition, we propose a joint training method of a multi-resolution STFT loss and a WavLM loss using a pre-trained WavLM network. Experiments show that with only 0.23M parameters, the proposed model achieves a better performance than DPCRN.

cs.SD↗
arXiv2023

Learning Dynamic Contextualised Word Embeddings via Template-based Temporal Adaptation

Dynamic contextualised word embeddings (DCWEs) represent the temporal semantic variations of words. We propose a method for learning DCWEs by time-adapting a pretrained Masked Language Model (MLM) using time-sensitive templates. Given two snapshots $C_1$ and $C_2$ of a corpus taken respectively at two distinct timestamps $T_1$ and $T_2$, we first propose an unsupervised method to select (a) \emph{pivot} terms related to both $C_1$ and $C_2$, and (b) \emph{anchor} terms that are associated with a specific pivot term in each individual snapshot. We then generate prompts by filling manually compiled templates using the extracted pivot and anchor terms. Moreover, we propose an automatic method to learn time-sensitive templates from $C_1$ and $C_2$, without requiring any human supervision. Next, we use the generated prompts to adapt a pretrained MLM to $T_2$ by fine-tuning using those prompts. Multiple experiments show that our proposed method reduces the perplexity of test sentences in $C_2$, outperforming the current state-of-the-art.

cs.CL↗
arXiv2023

Why Did the Chicken Cross the Road? Rephrasing and Analyzing Ambiguous Questions in VQA

Natural language is ambiguous. Resolving ambiguous questions is key to successfully answering them. Focusing on questions about images, we create a dataset of ambiguous examples. We annotate these, grouping answers by the underlying question they address and rephrasing the question for each group to reduce ambiguity. Our analysis reveals a linguistically-aligned ontology of reasons for ambiguity in visual questions. We then develop an English question-generation model which we demonstrate via automatic and human evaluation produces less ambiguous questions. We further show that the question generation objective we use allows the model to integrate answer group information without any direct supervision.

cs.CL↗
arXiv2023

Sensitivity analysis for publication bias on the time-dependent summary ROC analysis in meta-analysis of prognosis studies

In the analysis of prognosis studies with time-to-event outcomes, dichotomization of patients is often made. As the evaluations of prognostic capacity, the survivals of groups with high/low expression of the biomarker are often estimated by the Kaplan-Meier method, and the difference between groups is summarized via the hazard ratio (HR). The high/low expressions are usually determined by study-specific cutoff values, which brings heterogeneity over multiple prognosis studies and difficulty to synthesizing the results in a simple way. In meta-analysis of diagnostic studies with binary outcomes, the summary receiver operating characteristics (SROC) analysis provides a useful cutoff-free summary over studies. Recently, this methodology has been extended to the time-dependent SROC analysis for time-to-event outcomes in meta-analysis of prognosis studies. In this paper, we propose a sensitivity analysis method for evaluating the impact of publication bias on the time-dependent SROC analysis. Our proposal extends the recently introduced sensitivity analysis method for meta-analysis of diagnostic studies based on the bivariate normal model on sensitivity and specificity pairs. To model the selective publication process specific to prognosis studies, we introduce a trivariate model on the time-dependent sensitivity and specificity and the log-transformed HR. Based on the proved asymptotic property of the trivariate model, we introduce a likelihood based sensitivity analysis method based on the conditional likelihood constrained by the expected proportion of published studies. We illustrate the proposed sensitivity analysis method through the meta-analysis of Ki67 for breast cancer. Simulation studies are conducted to evaluate the performance of the proposed method.

stat.ME↗
arXiv2023

Together We Make Sense -- Learning Meta-Sense Embeddings from Pretrained Static Sense Embeddings

Sense embedding learning methods learn multiple vectors for a given ambiguous word, corresponding to its different word senses. For this purpose, different methods have been proposed in prior work on sense embedding learning that use different sense inventories, sense-tagged corpora and learning methods. However, not all existing sense embeddings cover all senses of ambiguous words equally well due to the discrepancies in their training resources. To address this problem, we propose the first-ever meta-sense embedding method -- Neighbour Preserving Meta-Sense Embeddings, which learns meta-sense embeddings by combining multiple independently trained source sense embeddings such that the sense neighbourhoods computed from the source embeddings are preserved in the meta-embedding space. Our proposed method can combine source sense embeddings that cover different sets of word senses. Experimental results on Word Sense Disambiguation (WSD) and Word-in-Context (WiC) tasks show that the proposed meta-sense embedding method consistently outperforms several competitive baselines.

cs.CL↗
arXiv2023

Variance-Reduced Off-Policy TDC Learning: Non-Asymptotic Convergence Analysis

Variance reduction techniques have been successfully applied to temporal-difference (TD) learning and help to improve the sample complexity in policy evaluation. However, the existing work applied variance reduction to either the less popular one time-scale TD algorithm or the two time-scale GTD algorithm but with a finite number of i.i.d.\ samples, and both algorithms apply to only the on-policy setting. In this work, we develop a variance reduction scheme for the two time-scale TDC algorithm in the off-policy setting and analyze its non-asymptotic convergence rate over both i.i.d.\ and Markovian samples. In the i.i.d.\ setting, our algorithm {matches the best-known lower bound $\tilde{O}(ε^{-1}$).} In the Markovian setting, our algorithm achieves the state-of-the-art sample complexity $O(ε^{-1} \log ε^{-1})$ that is near-optimal. Experiments demonstrate that the proposed variance-reduced TDC achieves a smaller asymptotic convergence error than both the conventional TDC and the variance-reduced TD.

cs.LG↗
arXiv2023

Super-Resolution Imaging via Angular Magnification

The far-field resolution of optical imaging systems is restricted by the Abbe diffraction limit, a direct result of the wave nature of light. One successful technological approach to circumventing this limit is to reduce the effective size of a point-spread-function. In the past decades, great endeavors have been made to engineer an effective point-spread-function by exploiting different mechanisms, including optical nonlinearities and structured light illumination. However, these methods are hard to be applied to objects in a far distance. Here, we propose a new way to achieve super-resolution in a far field by utilizing angular magnification. We present the first proof-of-concept demonstration of such an idea and demonstrate a new class of lenses with angular magnification for far-field super-resolution imaging. Both theoretical and experimental results demonstrate a more than two-fold enhancement beyond the angular-resolution limit in the far-field imaging. The proposed approach can be applied to super-resolution imaging of objects in far distance. It has promising potential applications in super-resolution telescopes and remote sensing.

physics.optics↗
arXiv2023

Solving Cosine Similarity Underestimation between High Frequency Words by L2 Norm Discounting

Cosine similarity between two words, computed using their contextualised token embeddings obtained from masked language models (MLMs) such as BERT has shown to underestimate the actual similarity between those words (Zhou et al., 2022). This similarity underestimation problem is particularly severe for highly frequent words. Although this problem has been noted in prior work, no solution has been proposed thus far. We observe that the L2 norm of contextualised embeddings of a word correlates with its log-frequency in the pretraining corpus. Consequently, the larger L2 norms associated with the highly frequent words reduce the cosine similarity values measured between them, thus underestimating the similarity scores. To solve this issue, we propose a method to discount the L2 norm of a contextualised word embedding by the frequency of that word in a corpus when measuring the cosine similarities between words. We show that the so called stop words behave differently from the rest of the words, which require special consideration during their discounting process. Experimental results on a contextualised word similarity dataset show that our proposed discounting method accurately solves the similarity underestimation problem.

cs.CL↗
arXiv2023

Accented Text-to-Speech Synthesis with Limited Data

This paper presents an accented text-to-speech (TTS) synthesis framework with limited training data. We study two aspects concerning accent rendering: phonetic (phoneme difference) and prosodic (pitch pattern and phoneme duration) variations. The proposed accented TTS framework consists of two models: an accented front-end for grapheme-to-phoneme (G2P) conversion and an accented acoustic model with integrated pitch and duration predictors for phoneme-to-Mel-spectrogram prediction. The accented front-end directly models the phonetic variation, while the accented acoustic model explicitly controls the prosodic variation. Specifically, both models are first pre-trained on a large amount of data, then only the accent-related layers are fine-tuned on a limited amount of data for the target accent. In the experiments, speech data of three English accents, i.e., General American English, Irish English, and British English Received Pronunciation, are used for pre-training. The pre-trained models are then fine-tuned with Scottish and General Australian English accents, respectively. Both objective and subjective evaluation results show that the accented TTS front-end fine-tuned with a small accented phonetic lexicon (5k words) effectively handles the phonetic variation of accents, while the accented TTS acoustic model fine-tuned with a limited amount of accented speech data (approximately 3 minutes) effectively improves the prosodic rendering including pitch and duration. The overall accent modeling contributes to improved speech quality and accent similarity.

eess.AS↗
arXiv2023

Dual Residual Attention Network for Image Denoising

In image denoising, deep convolutional neural networks (CNNs) can obtain favorable performance on removing spatially invariant noise. However, many of these networks cannot perform well on removing the real noise (i.e. spatially variant noise) generated during image acquisition or transmission, which severely sets back their application in practical image denoising tasks. Instead of continuously increasing the network depth, many researchers have revealed that expanding the width of networks can also be a useful way to improve model performance. It also has been verified that feature filtering can promote the learning ability of the models. Therefore, in this paper, we propose a novel Dual-branch Residual Attention Network (DRANet) for image denoising, which has both the merits of a wide model architecture and attention-guided feature learning. The proposed DRANet includes two different parallel branches, which can capture complementary features to enhance the learning ability of the model. We designed a new residual attention block (RAB) and a novel hybrid dilated residual attention block (HDRAB) for the upper and the lower branches, respectively. The RAB and HDRAB can capture rich local features through multiple skip connections between different convolutional layers, and the unimportant features are dropped by the residual attention modules. Meanwhile, the long skip connections in each branch, and the global feature fusion between the two parallel branches can capture the global features as well. Moreover, the proposed DRANet uses downsampling operations and dilated convolutions to increase the size of the receptive field, which can enable DRANet to capture more image context information. Extensive experiments demonstrate that compared with other state-of-the-art denoising methods, our DRANet can produce competitive denoising performance both on synthetic and real-world noise removal.

eess.IV↗
arXiv2023

LESS-VFL: Communication-Efficient Feature Selection for Vertical Federated Learning

We propose LESS-VFL, a communication-efficient feature selection method for distributed systems with vertically partitioned data. We consider a system of a server and several parties with local datasets that share a sample ID space but have different feature sets. The parties wish to collaboratively train a model for a prediction task. As part of the training, the parties wish to remove unimportant features in the system to improve generalization, efficiency, and explainability. In LESS-VFL, after a short pre-training period, the server optimizes its part of the global model to determine the relevant outputs from party models. This information is shared with the parties to then allow local feature selection without communication. We analytically prove that LESS-VFL removes spurious features from model training. We provide extensive empirical evidence that LESS-VFL can achieve high accuracy and remove spurious features at a fraction of the communication cost of other feature selection approaches.

cs.LG↗
arXiv2023

Two-stream instability with a growth rate insensitive to collisions in a dissipative plasma jet

The two-stream instability (Buneman instability) is traditionally derived as a collisionless instability with the presumption that collisions inhibit this instability. We show here via a combination of a collisional two-fluid model and associated experimental observations made in the Caltech plasma jet experiment, that in fact, a low frequency mode of the two-stream instability is indifferent to collisions. Despite the collision frequency greatly exceeding the growth rate of the instability, the instability can still cause an exponential growth of electron velocity and a rapid depletion of particle density. High collisionality nevertheless has an important effect as it enables the development of a double layer when the cross-section of the plasma jet is constricted by a kink-instigated Rayleigh-Taylor instability.

physics.plasm-ph↗
arXiv2023

Lacing topological orders in two dimensions: exactly solvable models for Kitaev's sixteen-fold way

A family of two-dimensional (2D) spin-1/2 models have been constructed to realize Kitaev's sixteen-fold way of anyon theories. Defining a one-dimensional (1D) path through all the lattice sites, and performing the Jordan-Wigner transformation with the help of the 1D path, we find that such a spin-1/2 model is equivalent to a model with $ν$ species of Majorana fermions coupled to a static $\mathbb{Z}_2$ gauge field. Here each specie of Majorana fermions gives rise to an energy band that carries a Chern number $\mathcal{C}=1$, yielding a total Chern number $\mathcal{C}=ν$. It has been shown that the ground states are three (four)-fold topologically degenerate on a torus, when $ν$ is an odd (even) number. These exactly solvable models can be achieved by quantum simulations.

cond-mat.str-el↗