arXiv subjects

Wei Huang

Publications and source records attributed to Wei Huang.

At least 253 records · Page 14Linked to original sources

arXiv2024

Embedding Generalized Semantic Knowledge into Few-Shot Remote Sensing Segmentation

Few-shot segmentation (FSS) for remote sensing (RS) imagery leverages supporting information from limited annotated samples to achieve query segmentation of novel classes. Previous efforts are dedicated to mining segmentation-guiding visual cues from a constrained set of support samples. However, they still struggle to address the pronounced intra-class differences in RS images, as sparse visual cues make it challenging to establish robust class-specific representations. In this paper, we propose a holistic semantic embedding (HSE) approach that effectively harnesses general semantic knowledge, i.e., class description (CD) embeddings.Instead of the naive combination of CD embeddings and visual features for segmentation decoding, we investigate embedding the general semantic knowledge during the feature extraction stage.Specifically, in HSE, a spatial dense interaction module allows the interaction of visual support features with CD embeddings along the spatial dimension via self-attention.Furthermore, a global content modulation module efficiently augments the global information of the target category in both support and query features, thanks to the transformative fusion of visual features and CD embeddings.These two components holistically synergize general CD embeddings and visual cues, constructing a robust class-specific representation.Through extensive experiments on the standard FSS benchmark, the proposed HSE approach demonstrates superior performance compared to peer work, setting a new state-of-the-art.

cs.CV↗

arXiv2024

Know in AdVance: Linear-Complexity Forecasting of Ad Campaign Performance with Evolving User Interest

Real-time Bidding (RTB) advertisers wish to \textit{know in advance} the expected cost and yield of ad campaigns to avoid trial-and-error expenses. However, Campaign Performance Forecasting (CPF), a sequence modeling task involving tens of thousands of ad auctions, poses challenges of evolving user interest, auction representation, and long context, making coarse-grained and static-modeling methods sub-optimal. We propose \textit{AdVance}, a time-aware framework that integrates local auction-level and global campaign-level modeling. User preference and fatigue are disentangled using a time-positioned sequence of clicked items and a concise vector of all displayed items. Cross-attention, conditioned on the fatigue vector, captures the dynamics of user interest toward each candidate ad. Bidders compete with each other, presenting a complete graph similar to the self-attention mechanism. Hence, we employ a Transformer Encoder to compress each auction into embedding by solving auxiliary tasks. These sequential embeddings are then summarized by a conditional state space model (SSM) to comprehend long-range dependencies while maintaining global linear complexity. Considering the irregular time intervals between auctions, we make SSM's parameters dependent on the current auction embedding and the time interval. We further condition SSM's global predictions on the accumulation of local results. Extensive evaluations and ablation studies demonstrate its superiority over state-of-the-art methods. AdVance has been deployed on the Tencent Advertising platform, and A/B tests show a remarkable 4.5\% uplift in Average Revenue per User (ARPU).

cs.IR↗

arXiv2024

BiLLM: Pushing the Limit of Post-Training Quantization for LLMs

Pretrained large language models (LLMs) exhibit exceptional general language processing capabilities but come with significant demands on memory and computational resources. As a powerful compression technology, binarization can extremely reduce model weights to a mere 1 bit, lowering the expensive computation and memory requirements. However, existing quantization techniques fall short of maintaining LLM performance under ultra-low bit-widths. In response to this challenge, we present BiLLM, a groundbreaking 1-bit post-training quantization scheme tailored for pretrained LLMs. Based on the weight distribution of LLMs, BiLLM first identifies and structurally selects salient weights, and minimizes the compression loss through an effective binary residual approximation strategy. Moreover, considering the bell-shaped distribution of the non-salient weights, we propose an optimal splitting search to group and binarize them accurately. BiLLM achieving for the first time high-accuracy inference (e.g. 8.41 perplexity on LLaMA2-70B) with only 1.08-bit weights across various LLMs families and evaluation metrics, outperforms SOTA quantization methods of LLM by significant margins. Moreover, BiLLM enables the binarization process of the LLM with 7 billion weights within 0.5 hours on a single GPU, demonstrating satisfactory time efficiency. Our code is available at https://github.com/Aaronhuang-778/BiLLM.

cs.LG↗

arXiv2024

Dynamic Self-adaptive Multiscale Distillation from Pre-trained Multimodal Large Model for Efficient Cross-modal Representation Learning

In recent years, pre-trained multimodal large models have attracted widespread attention due to their outstanding performance in various multimodal applications. Nonetheless, the extensive computational resources and vast datasets required for their training present significant hurdles for deployment in environments with limited computational resources. To address this challenge, we propose a novel dynamic self-adaptive multiscale distillation from pre-trained multimodal large model for efficient cross-modal representation learning for the first time. Unlike existing distillation methods, our strategy employs a multiscale perspective, enabling the extraction structural knowledge across from the pre-trained multimodal large model. Ensuring that the student model inherits a comprehensive and nuanced understanding of the teacher knowledge. To optimize each distillation loss in a balanced and efficient manner, we propose a dynamic self-adaptive distillation loss balancer, a novel component eliminating the need for manual loss weight adjustments and dynamically balances each loss item during the distillation process. Our methodology streamlines pre-trained multimodal large models using only their output features and original image-level information, requiring minimal computational resources. This efficient approach is suited for various applications and allows the deployment of advanced multimodal technologies even in resource-limited settings. Extensive experiments has demonstrated that our method maintains high performance while significantly reducing model complexity and training costs. Moreover, our distilled student model utilizes only image-level information to achieve state-of-the-art performance on cross-modal retrieval tasks, surpassing previous methods that relied on region-level information.

cs.CV↗

arXiv2024

Neural Causal Graph Collaborative Filtering

Graph collaborative filtering (GCF) has gained considerable attention in recommendation systems by leveraging graph learning techniques to enhance collaborative filtering (CF). One classical approach in GCF is to learn user and item embeddings with Graph Convolutional Network (GCN) and utilize these embeddings for CF models. However, existing GCN-based methods are insufficient in generating satisfactory embeddings for CF models. This is because they fail to model complex node dependencies and variable relation dependencies from a given graph, making the learned embeddings fragile to uncover the root causes of user interests. In this work, we propose to integrate causal modeling with the learning process of GCN-based GCF models, leveraging causality-aware graph embeddings to capture complex causal relations in recommendations. We complete the task by 1) Causal Graph conceptualization, 2) Neural Causal Model parameterization and 3) Variational inference for Neural Causal Model. Our Neural Causal Model, called Neural Causal Graph Collaborative Filtering (NCGCF), enables causal modeling for GCN-based GCF to facilitate accurate recommendations. Extensive experiments show that NCGCF provides precise recommendations that align with user preferences. We release our code and processed datasets at https://github.com/Chrystalii/CNGCF.

cs.IR↗

arXiv2024

Global and Local Prompts Cooperation via Optimal Transport for Federated Learning

Prompt learning in pretrained visual-language models has shown remarkable flexibility across various downstream tasks. Leveraging its inherent lightweight nature, recent research attempted to integrate the powerful pretrained models into federated learning frameworks to simultaneously reduce communication costs and promote local training on insufficient data. Despite these efforts, current federated prompt learning methods lack specialized designs to systematically address severe data heterogeneities, e.g., data distribution with both label and feature shifts involved. To address this challenge, we present Federated Prompts Cooperation via Optimal Transport (FedOTP), which introduces efficient collaborative prompt learning strategies to capture diverse category traits on a per-client basis. Specifically, for each client, we learn a global prompt to extract consensus knowledge among clients, and a local prompt to capture client-specific category characteristics. Unbalanced Optimal Transport is then employed to align local visual features with these prompts, striking a balance between global consensus and local personalization. By relaxing one of the equality constraints, FedOTP enables prompts to focus solely on the core regions of image patches. Extensive experiments on datasets with various types of heterogeneities have demonstrated that our FedOTP outperforms the state-of-the-art methods.

cs.LG↗

arXiv2024

Formal Verification of Robustness and Resilience of Learning-Enabled State Estimation Systems

This paper presents a formal verification guided approach for a principled design and implementation of robust and resilient learning-enabled systems. We focus on learning-enabled state estimation systems (LE-SESs), which have been widely used in robotics applications to determine the current state (e.g., location, speed, direction, etc.) of a complex system. The LE-SESs are networked systems, composed of a set of connected components including: Bayes filters for state estimation, and neural networks for processing sensory input. We study LE-SESs from the perspective of formal verification, which determines the satisfiabilty of a system model against the specified properties. Over LE-SESs, we investigate two key properties -- robustness and resilience -- and provide their formal definitions. To enable formal verification, we reduce the LE-SESs to a novel class of labelled transition systems, named {PO}^2-LTS in the paper, and formally express the properties as constrained optimisation objectives. We prove that the verification problems are NP-complete. Based on {PO}^2-LTS and the optimisation objectives, practical verification algorithms are developed to check the satisfiability of the properties on the LE-SESs. As a major case study, we interrogate a real-world dynamic tracking system which uses a single Kalman Filter (KF) -- a special case of Bayes filter -- to localise and track a ground vehicle. Its perception system, based on convolutional neural networks, processes a high-resolution Wide Area Motion Imagery (WAMI) data stream. Experimental results show that our algorithms can not only verify the properties of the WAMI tracking system but also provide representative examples, the latter of which inspired us to take an enhanced LE-SESs design where runtime monitors or joint-KFs are required. Experimental results confirm the improvement in the robustness of the enhanced design.

cs.RO↗

arXiv2024

Ensemble Adversarial Defense via Integration of Multiple Dispersed Low Curvature Models

The integration of an ensemble of deep learning models has been extensively explored to enhance defense against adversarial attacks. The diversity among sub-models increases the attack cost required to deceive the majority of the ensemble, thereby improving the adversarial robustness. While existing approaches mainly center on increasing diversity in feature representations or dispersion of first-order gradients with respect to input, the limited correlation between these diversity metrics and adversarial robustness constrains the performance of ensemble adversarial defense. In this work, we aim to enhance ensemble diversity by reducing attack transferability. We identify second-order gradients, which depict the loss curvature, as a key factor in adversarial robustness. Computing the Hessian matrix involved in second-order gradients is computationally expensive. To address this, we approximate the Hessian-vector product using differential approximation. Given that low curvature provides better robustness, our ensemble model was designed to consider the influence of curvature among different sub-models. We introduce a novel regularizer to train multiple more-diverse low-curvature network models. Extensive experiments across various datasets demonstrate that our ensemble model exhibits superior robustness against a range of attacks, underscoring the effectiveness of our approach.

cs.LG↗

arXiv2024

Privacy-Preserving End-to-End Spoken Language Understanding

Spoken language understanding (SLU), one of the key enabling technologies for human-computer interaction in IoT devices, provides an easy-to-use user interface. Human speech can contain a lot of user-sensitive information, such as gender, identity, and sensitive content. New types of security and privacy breaches have thus emerged. Users do not want to expose their personal sensitive information to malicious attacks by untrusted third parties. Thus, the SLU system needs to ensure that a potential malicious attacker cannot deduce the sensitive attributes of the users, while it should avoid greatly compromising the SLU accuracy. To address the above challenge, this paper proposes a novel SLU multi-task privacy-preserving model to prevent both the speech recognition (ASR) and identity recognition (IR) attacks. The model uses the hidden layer separation technique so that SLU information is distributed only in a specific portion of the hidden layer, and the other two types of information are removed to obtain a privacy-secure hidden layer. In order to achieve good balance between efficiency and privacy, we introduce a new mechanism of model pre-training, namely joint adversarial training, to further enhance the user privacy. Experiments over two SLU datasets show that the proposed method can reduce the accuracy of both the ASR and IR attacks close to that of a random guess, while leaving the SLU performance largely unaffected.

cs.CR↗

arXiv2024

Highly stable power control for chip-based continuous-variable quantum key distribution system

Quantum key distribution allows secret key generation with information theoretical security. It can be realized with photonic integrated circuits to benefit the tiny footprints and the large-scale manufacturing capacity. Continuous-variable quantum key distribution is suitable for chip-based integration due to its compatibility with mature optical communication devices. However, the quantum signal power control compatible with the mature photonic integration process faces difficulties on stability, which limits the system performance and causes the overestimation of secret key rate that opens practical security loopholes. Here, a highly stable chip-based quantum signal power control scheme based on a biased Mach-Zehnder interferometer structure is proposed, theoretically analyzed and experimentally implemented with standard silicon photonic techniques. Simulations and experimental results show that the proposed scheme significantly improves the system stability, where the standard deviation of the secret key rate is suppressed by an order of magnitude compared with the system using traditional designs, showing a promising and practicable way to realize highly stable continuous-variable quantum key distribution system on chip.

quant-ph↗

arXiv2024

Shear-enhanced Liquid Crystal Spinning of Conjugated Polymer Fibers

Conjugated polymer fibers can be used to manufacture various soft fibrous optoelectronic devices, significantly advancing wearable devices and smart textiles. Recently, conjugated polymer-based fibrous electronic devices have been widely used in energy conversion, electrochemical sensing, and human-machine interaction. However, the insufficient mechanical properties of conjugated polymer fibers, the difficulty in solution processing semiconductors with rigid main chains, and the challenges in large-scale continuous production have limited their further development in the wearable field. We regulated the pi - pi stacking interactions in conjugated polymer molecules below their critical liquid crystal concentration by applying fluid shear stress. We implemented secondary orientation, leading to the continuous fabrication of anisotropic semiconductor fibers. This strategy enables conjugated polymers with rigid backbones to synergistically enhance the mechanical and semiconductor properties of fibers through liquid crystal spinning. Furthermore, conjugated polymer fibers, exhibiting excellent electrochemical performance and high mechanical strength (600 MPa) that essentially meet the requirements for industrialized preparation, maintain stability under extreme temperatures, radiation, and chemical reagents. Lastly, we have demonstrated logic circuits using semiconductor fiber organic electrochemical transistors, showcasing its application potential in the field of wearable fabric-style logic processing. These findings confirm the importance of the liquid crystalline state and solution control in optimizing the performance of conjugated polymer fibers, thus paving the way for developing a new generation of soft fiber semiconductor devices.

cond-mat.soft↗

arXiv2024

High-Rate 16-node quantum access network based on passive optical network

Quantum key distribution can provide information-theoretical secure communication, which is now heading towards building the quantum secure network for real-world applications. In most built quantum secure networks, point-to-multipoint (PTMP) topology is one of the most popular schemes, especially for quantum access networks. However, due to the lack of custom protocols with high secret key rate and compatible with classical optical networks for PTMP scheme, there is still no efficient way for a high-performance quantum access network with a multitude of users. Here, we report an experimental demonstration of a high-rate 16-nodes quantum access network based on passive optical network, where a high-efficient coherent-state PTMP protocol is novelly designed to allow independent secret key generation between one transmitter and multiple receivers concurrently. Such accomplishment is attributed to a well-designed real-time shot-noise calibration method, a series of advanced digital signal processing algorithms and a flexible post-processing strategy with high success probability. Finally, the experimental results show that the average secret key rate is around 2.086 Mbps between the transmitter and each user, which is two orders of magnitude higher than previous demonstrations. With the advantages of low cost, excellent compatibility, and wide bandwidth, our work paves the way for building practical PTMP quantum access networks, thus constituting an important step towards scalable quantum secure networks.

quant-ph↗

arXiv2024

DiffSal: Joint Audio and Video Learning for Diffusion Saliency Prediction

Audio-visual saliency prediction can draw support from diverse modality complements, but further performance enhancement is still challenged by customized architectures as well as task-specific loss functions. In recent studies, denoising diffusion models have shown more promising in unifying task frameworks owing to their inherent ability of generalization. Following this motivation, a novel Diffusion architecture for generalized audio-visual Saliency prediction (DiffSal) is proposed in this work, which formulates the prediction problem as a conditional generative task of the saliency map by utilizing input audio and video as the conditions. Based on the spatio-temporal audio-visual features, an extra network Saliency-UNet is designed to perform multi-modal attention modulation for progressive refinement of the ground-truth saliency map from the noisy map. Extensive experiments demonstrate that the proposed DiffSal can achieve excellent performance across six challenging audio-visual benchmarks, with an average relative improvement of 6.3\% over the previous state-of-the-art results by six metrics.

cs.CV↗

arXiv2024

Continuous-variable quantum key distribution over 28.6 km fiber with an integrated silicon photonic receiver chip

Quantum key distribution, which ensures information-theoretically secret key generation, is currently advancing through photonic integration to achieve high performance, cost reduction and compact size, thereby facilitating the large-scale deployment. Continuous-variable quantum key distribution is an attractive approach for photonic integrations due to its compatibility with off-the-shelf optical communication devices. However, its chip-based systems have encountered significant limitations primarily related to the shot-noise-limited receiver design, which demands low noise, wide bandwidth, high clearance and well stability. Here, we report the implementation of a real local oscillator continuous-variable quantum key distribution system with an integrated silicon photonic receiver chip. Thanks to the well-designed chip-based homodyne detectors with a bandwidth up to 1.5 GHz and a clearance up to 7.42 dB, the transmission distance of the system has been extended to 28.6 km, achieving a secret key generation rate of Mbps level. This technological advancement enables the quantum key distribution systems with photonic integrated receivers to achieve the coverage in both access network scenarios and short-distance metropolitan interconnections, paving the way for the development of the next-generation quantum key distribution networks on a large scale.

quant-ph↗

arXiv2024

A Unified Causal View of Instruction Tuning

Instruction tuning on a mixture of tasks has improved zero-shot capabilities in natural language processing (NLP). Nevertheless, existing methods often learn features that exhibit correlations between instruction-formatted samples and target labels, rather than causal relationships. Termed as ``spurious correlation'' in statistics, such a correlation may change drastically in a new task, making the effect from the learned features to be misleading. To this end, we develop a meta Structural Causal Model (meta-SCM) to integrate different NLP tasks under a single causal structure of the data. Specifically, the meta-SCM introduces multiple latent factors that represent properties of source context, only some of which causally influence the target labels for a specific task. The key idea is to learn task-required causal factors and only use those to make predictions for a given task. Theoretically, we prove the causal factor can be identified without mixing information from others. Guided by the identifiability, we propose a Structural Instruction Tuning (SIT) method to learn the task-required causal representations that can mimic the causal factors for each task. The utility of our approach is verified by improvements of zero-shot ability on a range of unseen datasets and tasks.

cs.CL↗

arXiv2024

Nonparametric Estimation of Mediation Effects with A General Treatment

To investigate causal mechanisms, causal mediation analysis decomposes the total treatment effect into the natural direct and indirect effects. This paper examines the estimation of the direct and indirect effects in a general treatment effect model, where the treatment can be binary, multi-valued, continuous, or a mixture. We propose generalized weighting estimators with weights estimated by solving an expanding set of equations. Under some sufficient conditions, we show that the proposed estimators are consistent and asymptotically normal. Specifically, when the treatment is discrete, the proposed estimators attain the semiparametric efficiency bounds. Meanwhile, when the treatment is continuous, the convergence rates of the proposed estimators are slower than $N^{-1/2}$; however, they are still more efficient than that constructed from the true weighting function. A simulation study reveals that our estimators exhibit a satisfactory finite-sample performance, while an application shows their practical value

math.ST↗

arXiv2024

Generator Identification for Linear SDEs with Additive and Multiplicative Noise

In this paper, we present conditions for identifying the generator of a linear stochastic differential equation (SDE) from the distribution of its solution process with a given fixed initial state. These identifiability conditions are crucial in causal inference using linear SDEs as they enable the identification of the post-intervention distributions from its observational distribution. Specifically, we derive a sufficient and necessary condition for identifying the generator of linear SDEs with additive noise, as well as a sufficient condition for identifying the generator of linear SDEs with multiplicative noise. We show that the conditions derived for both types of SDEs are generic. Moreover, we offer geometric interpretations of the derived identifiability conditions to enhance their understanding. To validate our theoretical results, we perform a series of simulations, which support and substantiate the established findings.

math.ST↗

arXiv2024

A Fast, Performant, Secure Distributed Training Framework For Large Language Model

The distributed (federated) LLM is an important method for co-training the domain-specific LLM using siloed data. However, maliciously stealing model parameters and data from the server or client side has become an urgent problem to be solved. In this paper, we propose a secure distributed LLM based on model slicing. In this case, we deploy the Trusted Execution Environment (TEE) on both the client and server side, and put the fine-tuned structure (LoRA or embedding of P-tuning v2) into the TEE. Then, secure communication is executed in the TEE and general environments through lightweight encryption. In order to further reduce the equipment cost as well as increase the model performance and accuracy, we propose a split fine-tuning scheme. In particular, we split the LLM by layers and place the latter layers in a server-side TEE (the client does not need a TEE). We then combine the proposed Sparsification Parameter Fine-tuning (SPF) with the LoRA part to improve the accuracy of the downstream task. Numerous experiments have shown that our method guarantees accuracy while maintaining security.

cs.LG↗