arXiv subjects

Xiangyu Zhang

Publications and source records attributed to Xiangyu Zhang.

At least 253 records · Page 14Linked to original sources

arXiv2023

Channel Estimation with Dynamic Metasurface Antennas via Model-Based Learning

Dynamic Metasurface Antenna (DMA) is a cutting-edge antenna technology offering scalable and sustainable solutions for large antenna arrays. The effectiveness of DMAs stems from their inherent configurable analog signal processing capabilities, which facilitate cost-limited implementations. However, when DMAs are used in multiple input multiple output (MIMO) communication systems, they pose challenges in channel estimation due to their analog compression. In this paper, we propose two model-based learning methods to overcome this challenge. Our approach starts by casting channel estimation as a compressed sensing problem. Here, the sensing matrix is formed using a random DMA weighting matrix combined with a spatial gridding dictionary. We then employ the learned iterative shrinkage and thresholding algorithm (LISTA) to recover the sparse channel parameters. LISTA unfolds the iterative shrinkage and thresholding algorithm into a neural network and trains the neural network into a highly efficient channel estimator fitting with the previous channel. As the sensing matrix is crucial to the accuracy of LISTA recovery, we introduce another data-aided method, LISTA-sensing matrix optimization (LISTA-SMO), to jointly optimize the sensing matrix. LISTA-SMO takes LISTA as a backbone and embeds the sensing matrix optimization layers in LISTA's neural network, allowing for the optimization of the sensing matrix along with the training of LISTA. Furthermore, we propose a self-supervised learning technique to tackle the difficulty of acquiring noise-free data. Our numerical results demonstrate that LISTA outperforms traditional sparse recovery methods regarding channel estimation accuracy and efficiency. Besides, LISTA-SMO achieves better channel accuracy than LISTA, demonstrating the effectiveness in optimizing the sensing matrix.

cs.IT↗

arXiv2023

Assessing and Improving Syntactic Adversarial Robustness of Pre-trained Models for Code Translation

Context: Pre-trained models (PTMs) have demonstrated significant potential in automatic code translation. However, the vulnerability of these models in translation tasks, particularly in terms of syntax, has not been extensively investigated. Objective: To fill this gap, our study aims to propose a novel approach CoTR to assess and improve the syntactic adversarial robustness of PTMs in code translation. Method: CoTR consists of two components: CoTR-A and CoTR-D. CoTR-A generates adversarial examples by transforming programs, while CoTR-D proposes a semantic distance-based sampling data augmentation method and adversarial training method to improve the model's robustness and generalization capabilities. The Pass@1 metric is used by CoTR to assess the performance of PTMs, which is more suitable for code translation tasks and offers a more precise evaluation in real world scenarios. Results: The effectiveness of CoTR is evaluated through experiments on real world Java to Python datasets. The results demonstrate that CoTR-A can significantly reduce the performance of existing PTMs, while CoTR-D effectively improves the robustness of PTMs. Conclusion: Our study identifies the limitations of current PTMs, including large language models, in code translation tasks. It highlights the potential of CoTR as an effective solution to enhance the robustness of PTMs for code translation tasks.

cs.SE↗

arXiv2023

ParaFuzz: An Interpretability-Driven Technique for Detecting Poisoned Samples in NLP

Backdoor attacks have emerged as a prominent threat to natural language processing (NLP) models, where the presence of specific triggers in the input can lead poisoned models to misclassify these inputs to predetermined target classes. Current detection mechanisms are limited by their inability to address more covert backdoor strategies, such as style-based attacks. In this work, we propose an innovative test-time poisoned sample detection framework that hinges on the interpretability of model predictions, grounded in the semantic meaning of inputs. We contend that triggers (e.g., infrequent words) are not supposed to fundamentally alter the underlying semantic meanings of poisoned samples as they want to stay stealthy. Based on this observation, we hypothesize that while the model's predictions for paraphrased clean samples should remain stable, predictions for poisoned samples should revert to their true labels upon the mutations applied to triggers during the paraphrasing process. We employ ChatGPT, a state-of-the-art large language model, as our paraphraser and formulate the trigger-removal task as a prompt engineering problem. We adopt fuzzing, a technique commonly used for unearthing software vulnerabilities, to discover optimal paraphrase prompts that can effectively eliminate triggers while concurrently maintaining input semantics. Experiments on 4 types of backdoor attacks, including the subtle style backdoors, and 4 distinct datasets demonstrate that our approach surpasses baseline methods, including STRIP, RAP, and ONION, in precision and recall.

cs.CR↗

arXiv2023

Hierarchical Semi-Implicit Variational Inference with Application to Diffusion Model Acceleration

Semi-implicit variational inference (SIVI) has been introduced to expand the analytical variational families by defining expressive semi-implicit distributions in a hierarchical manner. However, the single-layer architecture commonly used in current SIVI methods can be insufficient when the target posterior has complicated structures. In this paper, we propose hierarchical semi-implicit variational inference, called HSIVI, which generalizes SIVI to allow more expressive multi-layer construction of semi-implicit distributions. By introducing auxiliary distributions that interpolate between a simple base distribution and the target distribution, the conditional layers can be trained by progressively matching these auxiliary distributions one layer after another. Moreover, given pre-trained score networks, HSIVI can be used to accelerate the sampling process of diffusion models with the score matching objective. We show that HSIVI significantly enhances the expressiveness of SIVI on several Bayesian inference problems with complicated target distributions. When used for diffusion model acceleration, we show that HSIVI can produce high quality samples comparable to or better than the existing fast diffusion model based samplers with a small number of function evaluations on various datasets.

cs.LG↗

arXiv2023

LAMP: Learn A Motion Pattern for Few-Shot-Based Video Generation

With the impressive progress in diffusion-based text-to-image generation, extending such powerful generative ability to text-to-video raises enormous attention. Existing methods either require large-scale text-video pairs and a large number of training resources or learn motions that are precisely aligned with template videos. It is non-trivial to balance a trade-off between the degree of generation freedom and the resource costs for video generation. In our study, we present a few-shot-based tuning framework, LAMP, which enables text-to-image diffusion model Learn A specific Motion Pattern with 8~16 videos on a single GPU. Specifically, we design a first-frame-conditioned pipeline that uses an off-the-shelf text-to-image model for content generation so that our tuned video diffusion model mainly focuses on motion learning. The well-developed text-to-image techniques can provide visually pleasing and diverse content as generation conditions, which highly improves video quality and generation freedom. To capture the features of temporal dimension, we expand the pretrained 2D convolution layers of the T2I model to our novel temporal-spatial motion learning layers and modify the attention blocks to the temporal level. Additionally, we develop an effective inference trick, shared-noise sampling, which can improve the stability of videos with computational costs. Our method can also be flexibly applied to other tasks, e.g. real-world image animation and video editing. Extensive experiments demonstrate that LAMP can effectively learn the motion pattern on limited data and generate high-quality videos. The code and models are available at https://rq-wu.github.io/projects/LAMP.

cs.CV↗

arXiv2023

Enhancing Code-switching Speech Recognition with Interactive Language Biases

Languages usually switch within a multilingual speech signal, especially in a bilingual society. This phenomenon is referred to as code-switching (CS), making automatic speech recognition (ASR) challenging under a multilingual scenario. We propose to improve CS-ASR by biasing the hybrid CTC/attention ASR model with multi-level language information comprising frame- and token-level language posteriors. The interaction between various resolutions of language biases is subsequently explored in this work. We conducted experiments on datasets from the ASRU 2019 code-switching challenge. Compared to the baseline, the proposed interactive language biases (ILB) method achieves higher performance and ablation studies highlight the effects of different language biases and their interactions. In addition, the results presented indicate that language bias implicitly enhances internal language modeling, leading to performance degradation after employing an external language model.

eess.AS↗

arXiv2023

Cold & Warm Net: Addressing Cold-Start Users in Recommender Systems

Cold-start recommendation is one of the major challenges faced by recommender systems (RS). Herein, we focus on the user cold-start problem. Recently, methods utilizing side information or meta-learning have been used to model cold-start users. However, it is difficult to deploy these methods to industrial RS. There has not been much research that pays attention to the user cold-start problem in the matching stage. In this paper, we propose Cold & Warm Net based on expert models who are responsible for modeling cold-start and warm-up users respectively. A gate network is applied to incorporate the results from two experts. Furthermore, dynamic knowledge distillation acting as a teacher selector is introduced to assist experts in better learning user representation. With comprehensive mutual information, features highly relevant to user behavior are selected for the bias net which explicitly models user behavior bias. Finally, we evaluate our Cold & Warm Net on public datasets in comparison to models commonly applied in the matching stage and it outperforms other models on all user types. The proposed model has also been deployed on an industrial short video platform and achieves a significant increase in app dwell time and user retention rate.

cs.IR↗

arXiv2023

Unidirectional brain-computer interface: Artificial neural network encoding natural images to fMRI response in the visual cortex

While significant advancements in artificial intelligence (AI) have catalyzed progress across various domains, its full potential in understanding visual perception remains underexplored. We propose an artificial neural network dubbed VISION, an acronym for "Visual Interface System for Imaging Output of Neural activity," to mimic the human brain and show how it can foster neuroscientific inquiries. Using visual and contextual inputs, this multimodal model predicts the brain's functional magnetic resonance imaging (fMRI) scan response to natural images. VISION successfully predicts human hemodynamic responses as fMRI voxel values to visual inputs with an accuracy exceeding state-of-the-art performance by 45%. We further probe the trained networks to reveal representational biases in different visual areas, generate experimentally testable hypotheses, and formulate an interpretable metric to associate these hypotheses with cortical functions. With both a model and evaluation metric, the cost and time burdens associated with designing and implementing functional analysis on the visual cortex could be reduced. Our work suggests that the evolution of computational models may shed light on our fundamental understanding of the visual cortex and provide a viable approach toward reliable brain-machine interfaces.

cs.CV↗

arXiv2023

Cross Modal Transformer: Towards Fast and Robust 3D Object Detection

In this paper, we propose a robust 3D detector, named Cross Modal Transformer (CMT), for end-to-end 3D multi-modal detection. Without explicit view transformation, CMT takes the image and point clouds tokens as inputs and directly outputs accurate 3D bounding boxes. The spatial alignment of multi-modal tokens is performed by encoding the 3D points into multi-modal features. The core design of CMT is quite simple while its performance is impressive. It achieves 74.1\% NDS (state-of-the-art with single model) on nuScenes test set while maintaining fast inference speed. Moreover, CMT has a strong robustness even if the LiDAR is missing. Code is released at https://github.com/junjie18/CMT.

cs.CV↗

arXiv2023

RevColV2: Exploring Disentangled Representations in Masked Image Modeling

Masked image modeling (MIM) has become a prevalent pre-training setup for vision foundation models and attains promising performance. Despite its success, existing MIM methods discard the decoder network during downstream applications, resulting in inconsistent representations between pre-training and fine-tuning and can hamper downstream task performance. In this paper, we propose a new architecture, RevColV2, which tackles this issue by keeping the entire autoencoder architecture during both pre-training and fine-tuning. The main body of RevColV2 contains bottom-up columns and top-down columns, between which information is reversibly propagated and gradually disentangled. Such design enables our architecture with the nice property: maintaining disentangled low-level and semantic information at the end of the network in MIM pre-training. Our experimental results suggest that a foundation model with decoupled features can achieve competitive performance across multiple downstream vision tasks such as image classification, semantic segmentation and object detection. For example, after intermediate fine-tuning on ImageNet-22K dataset, RevColV2-L attains 88.4% top-1 accuracy on ImageNet-1K classification and 58.6 mIoU on ADE20K semantic segmentation. With extra teacher and large scale dataset, RevColv2-L achieves 62.1 box AP on COCO detection and 60.4 mIoU on ADE20K semantic segmentation. Code and models are released at https://github.com/megvii-research/RevCol

cs.CV↗

arXiv2023

PEM: Representing Binary Program Semantics for Similarity Analysis via a Probabilistic Execution Model

Binary similarity analysis determines if two binary executables are from the same source program. Existing techniques leverage static and dynamic program features and may utilize advanced Deep Learning techniques. Although they have demonstrated great potential, the community believes that a more effective representation of program semantics can further improve similarity analysis. In this paper, we propose a new method to represent binary program semantics. It is based on a novel probabilistic execution engine that can effectively sample the input space and the program path space of subject binaries. More importantly, it ensures that the collected samples are comparable across binaries, addressing the substantial variations of input specifications. Our evaluation on 9 real-world projects with 35k functions, and comparison with 6 state-of-the-art techniques show that PEM can achieve a precision of 96% with common settings, outperforming the baselines by 10-20%.

cs.SE↗

arXiv2023

Scene-Driven Exploration and GUI Modeling for Android Apps

Due to the competitive environment, mobile apps are usually produced under pressure with lots of complicated functionality and UI pages. Therefore, it is challenging for various roles to design, understand, test, and maintain these apps. The extracted transition graphs for apps such as ATG, WTG, and STG have a low transition coverage and coarse-grained granularity, which limits the existing methods of graphical user interface (GUI) modeling by UI exploration. To solve these problems, in this paper, we propose SceneDroid, a scene-driven exploration approach to extracting the GUI scenes dynamically by integrating a series of novel techniques including smart exploration, state fuzzing, and indirect launching strategies. We present the GUI scenes as a scene transition graph (SceneTG) to model the GUI of apps with high transition coverage and fine? grained granularity. Compared with the existing GUI modeling tools, SceneDroid has improved by 168.74% in the coverage of transition pairs and 162.42% in scene extraction. Apart from the effectiveness evaluation of SceneDroid, we also illustrate the future potential of SceneDroid as a fundamental capability to support app development, reverse engineering, and GUI regression testing.

cs.SE↗

arXiv2023

On-the-fly Improving Performance of Deep Code Models via Input Denoising

Deep learning has been widely adopted to tackle various code-based tasks by building deep code models based on a large amount of code snippets. While these deep code models have achieved great success, even state-of-the-art models suffer from noise present in inputs leading to erroneous predictions. While it is possible to enhance models through retraining/fine-tuning, this is not a once-and-for-all approach and incurs significant overhead. In particular, these techniques cannot on-the-fly improve performance of (deployed) models. There are currently some techniques for input denoising in other domains (such as image processing), but since code input is discrete and must strictly abide by complex syntactic and semantic constraints, input denoising techniques in other fields are almost not applicable. In this work, we propose the first input denoising technique (i.e., CodeDenoise) for deep code models. Its key idea is to localize noisy identifiers in (likely) mispredicted inputs, and denoise such inputs by cleansing the located identifiers. It does not need to retrain or reconstruct the model, but only needs to cleanse inputs on-the-fly to improve performance. Our experiments on 18 deep code models (i.e., three pre-trained models with six code-based datasets) demonstrate the effectiveness and efficiency of CodeDenoise. For example, on average, CodeDenoise successfully denoises 21.91% of mispredicted inputs and improves the original models by 2.04% in terms of the model accuracy across all the subjects in an average of 0.48 second spent on each input, substantially outperforming the widely-used fine-tuning strategy.

cs.SE↗

arXiv2023

SCSC: Spatial Cross-scale Convolution Module to Strengthen both CNNs and Transformers

This paper presents a module, Spatial Cross-scale Convolution (SCSC), which is verified to be effective in improving both CNNs and Transformers. Nowadays, CNNs and Transformers have been successful in a variety of tasks. Especially for Transformers, increasing works achieve state-of-the-art performance in the computer vision community. Therefore, researchers start to explore the mechanism of those architectures. Large receptive fields, sparse connections, weight sharing, and dynamic weight have been considered keys to designing effective base models. However, there are still some issues to be addressed: large dense kernels and self-attention are inefficient, and large receptive fields make it hard to capture local features. Inspired by the above analyses and to solve the mentioned problems, in this paper, we design a general module taking in these design keys to enhance both CNNs and Transformers. SCSC introduces an efficient spatial cross-scale encoder and spatial embed module to capture assorted features in one layer. On the face recognition task, FaceResNet with SCSC can improve 2.7% with 68% fewer FLOPs and 79% fewer parameters. On the ImageNet classification task, Swin Transformer with SCSC can achieve even better performance with 22% fewer FLOPs, and ResNet with CSCS can improve 5.3% with similar complexity. Furthermore, a traditional network (e.g., ResNet) embedded with SCSC can match Swin Transformer's performance.

cs.CV↗

arXiv2023

A Syntax-Guided Multi-Task Learning Approach for Turducken-Style Code Generation

Due to the development of pre-trained language models, automated code generation techniques have shown great promise in recent years. However, the generated code is difficult to meet the syntactic constraints of the target language, especially in the case of Turducken-style code, where declarative code snippets are embedded within imperative programs. In this study, we summarize the lack of syntactic constraints into three significant challenges: (1) the efficient representation of syntactic constraints, (2) the effective integration of syntactic information, and (3) the scalable syntax-first decoding algorithm. To address these challenges, we propose a syntax-guided multi-task learning approach TurduckenGen. Specifically, we first explicitly append the type information to the code tokens to capture the representation of syntactic constraints. Then we formalize code generation with syntactic constraint representation as an auxiliary task to enable the model to learn the syntactic constraints of the code. Finally, the syntactically correct code is selected accurately from the multiple candidates with the help of the compiler feedback. Extensive experiments and comprehensive analysis demonstrate the effectiveness and general applicability of our approach after being compared with six state-of-the-art baselines on two Turducken-style code datasets. Finally, we conducted a human study and found the code quality generated by our approach is better than baselines in terms of code readability and semantic similarity.

cs.SE↗

arXiv2023

OnlineRefer: A Simple Online Baseline for Referring Video Object Segmentation

Referring video object segmentation (RVOS) aims at segmenting an object in a video following human instruction. Current state-of-the-art methods fall into an offline pattern, in which each clip independently interacts with text embedding for cross-modal understanding. They usually present that the offline pattern is necessary for RVOS, yet model limited temporal association within each clip. In this work, we break up the previous offline belief and propose a simple yet effective online model using explicit query propagation, named OnlineRefer. Specifically, our approach leverages target cues that gather semantic information and position prior to improve the accuracy and ease of referring predictions for the current frame. Furthermore, we generalize our online model into a semi-online framework to be compatible with video-based backbones. To show the effectiveness of our method, we evaluate it on four benchmarks, \ie, Refer-Youtube-VOS, Refer-DAVIS17, A2D-Sentences, and JHMDB-Sentences. Without bells and whistles, our OnlineRefer with a Swin-L backbone achieves 63.5 J&F and 64.8 J&F on Refer-Youtube-VOS and Refer-DAVIS17, outperforming all other offline methods.

cs.CV↗

arXiv2023

GroupLane: End-to-End 3D Lane Detection with Channel-wise Grouping

Efficiency is quite important for 3D lane detection due to practical deployment demand. In this work, we propose a simple, fast, and end-to-end detector that still maintains high detection precision. Specifically, we devise a set of fully convolutional heads based on row-wise classification. In contrast to previous counterparts, ours supports recognizing both vertical and horizontal lanes. Besides, our method is the first one to perform row-wise classification in bird-eye-view. In the heads, we split feature into multiple groups and every group of feature corresponds to a lane instance. During training, the predictions are associated with lane labels using the proposed single-win one-to-one matching to compute loss, and no post-processing operation is demanded for inference. In this way, our proposed fully convolutional detector, GroupLane, realizes end-to-end detection like DETR. Evaluated on 3 real world 3D lane benchmarks, OpenLane, Once-3DLanes, and OpenLane-Huawei, GroupLane adopting ConvNext-Base as the backbone outperforms the published state-of-the-art PersFormer by 13.6% F1 score in the OpenLane validation set. Besides, GroupLane with ResNet18 still surpasses PersFormer by 4.9% F1 score, while the inference speed is nearly 7x faster and the FLOPs is only 13.3% of it.

cs.CV↗

arXiv2023

ChatSpot: Bootstrapping Multimodal LLMs via Precise Referring Instruction Tuning

Human-AI interactivity is a critical aspect that reflects the usability of multimodal large language models (MLLMs). However, existing end-to-end MLLMs only allow users to interact with them through language instructions, leading to the limitation of the interactive accuracy and efficiency. In this study, we present precise referring instructions that utilize diverse reference representations such as points and boxes as referring prompts to refer to the special region. This enables MLLMs to focus on the region of interest and achieve finer-grained interaction. Based on precise referring instruction, we propose ChatSpot, a unified end-to-end multimodal large language model that supports diverse forms of interactivity including mouse clicks, drag-and-drop, and drawing boxes, which provides a more flexible and seamless interactive experience. We also construct a multi-grained vision-language instruction-following dataset based on existing datasets and GPT-4 generating. Furthermore, we design a series of evaluation tasks to assess the effectiveness of region recognition and interaction. Experimental results showcase ChatSpot's promising performance.

cs.CL↗