arXiv subjects

Han Wang

Publications and source records attributed to Han Wang.

At least 253 records · Page 14Linked to original sources

arXiv2024

Learning Dynamic Local Context Representations for Infrared Small Target Detection

Infrared small target detection (ISTD) is challenging due to complex backgrounds, low signal-to-clutter ratios, and varying target sizes and shapes. Effective detection relies on capturing local contextual information at the appropriate scale. However, small-kernel CNNs have limited receptive fields, leading to false alarms, while transformer models, with global receptive fields, often treat small targets as noise, resulting in miss-detections. Hybrid models struggle to bridge the semantic gap between CNNs and transformers, causing high complexity.To address these challenges, we propose LCRNet, a novel method that learns dynamic local context representations for ISTD. The model consists of three components: (1) C2FBlock, inspired by PDE solvers, for efficient small target information capture; (2) DLC-Attention, a large-kernel attention mechanism that dynamically builds context and reduces feature redundancy; and (3) HLKConv, a hierarchical convolution operator based on large-kernel decomposition that preserves sparsity and mitigates the drawbacks of dilated convolutions. Despite its simplicity, with only 1.65M parameters, LCRNet achieves state-of-the-art (SOTA) performance.Experiments on multiple datasets, comparing LCRNet with 33 SOTA methods, demonstrate its superior performance and efficiency.

cs.CV↗

arXiv2024

Apollo-Forecast: Overcoming Aliasing and Inference Speed Challenges in Language Models for Time Series Forecasting

Encoding time series into tokens and using language models for processing has been shown to substantially augment the models' ability to generalize to unseen tasks. However, existing language models for time series forecasting encounter several obstacles, including aliasing distortion and prolonged inference times, primarily due to the limitations of quantization processes and the computational demands of large models. This paper introduces Apollo-Forecast, a novel framework that tackles these challenges with two key innovations: the Anti-Aliasing Quantization Module (AAQM) and the Race Decoding (RD) technique. AAQM adeptly encodes sequences into tokens while mitigating high-frequency noise in the original signals, thus enhancing both signal fidelity and overall quantization efficiency. RD employs a draft model to enable parallel processing and results integration, which markedly accelerates the inference speed for long-term predictions, particularly in large-scale models. Extensive experiments on various real-world datasets show that Apollo-Forecast outperforms state-of-the-art methods by 35.41\% and 18.99\% in WQL and MASE metrics, respectively, in zero-shot scenarios. Furthermore, our method achieves a 1.9X-2.7X acceleration in inference speed over baseline methods.

cs.LG↗

arXiv2024

Dynamic-VLM: Simple Dynamic Visual Token Compression for VideoLLM

The application of Large Vision-Language Models (LVLMs) for analyzing images and videos is an exciting and rapidly evolving field. In recent years, we've seen significant growth in high-quality image-text datasets for fine-tuning image understanding, but there is still a lack of comparable datasets for videos. Additionally, many VideoLLMs are extensions of single-image VLMs, which may not efficiently handle the complexities of longer videos. In this study, we introduce a large-scale synthetic dataset created from proprietary models, using carefully designed prompts to tackle a wide range of questions. We also explore a dynamic visual token compression architecture that strikes a balance between computational efficiency and performance. Our proposed \model{} achieves state-of-the-art results across various video tasks and shows impressive generalization, setting new baselines in multi-image understanding. Notably, \model{} delivers an absolute improvement of 2.7\% over LLaVA-OneVision on VideoMME and 10.7\% on MuirBench. Codes are available at https://github.com/Hon-Wong/ByteVideoLLM

cs.CV↗

arXiv2024

Strategizing Equitable Transit Evacuations: A Data-Driven Reinforcement Learning Approach

As natural disasters become increasingly frequent, the need for efficient and equitable evacuation planning has become more critical. This paper proposes a data-driven, reinforcement learning-based framework to optimize bus-based evacuations with an emphasis on improving both efficiency and equity. We model the evacuation problem as a Markov Decision Process solved by reinforcement learning, using real-time transit data from General Transit Feed Specification and transportation networks extracted from OpenStreetMap. The reinforcement learning agent dynamically reroutes buses from their scheduled location to minimize total passengers' evacuation time while prioritizing equity-priority communities. Simulations on the San Francisco Bay Area transportation network indicate that the proposed framework achieves significant improvements in both evacuation efficiency and equitable service distribution compared to traditional rule-based and random strategies. These results highlight the potential of reinforcement learning to enhance system performance and urban resilience during emergency evacuations, offering a scalable solution for real-world applications in intelligent transportation systems.

cs.LG↗

arXiv2024

Constructive RNNs: An Error-Recurrence Perspective on Time-Variant Zero Finding Problem Solving Under Uncertainty

When facing time-variant problems in analog computing, the desirable RNN design requires finite-time convergence and robustness with respect to various types of uncertainties, due to the time-variant nature and difficulties in implementation. It is very worthwhile to explore terminal zeroing neural networks, through examining and applying available attracting laws. In this paper, from a control-theoretic point of view, an error recurrence system approach is presented by equipping with uncertainty compensation in the pre-specified error dynamics, capable of enhancing robustness properly. Novel rectifying actions are designed to make finite-time settling so that the convergence speed and the computing accuracy of time-variant computing can be improved. Double-power and power-exponential rectifying actions are respectively formed to construct specific models, while the particular expressions of settling time function for the former are presented, and for the latter the proximate settling-time estimations are given, with which the fixed-time convergence of the corresponding models is in turn established. Moreover, the uncertainty compensation by the signum/smoothing-signum techniques are adopted for finite-duration stabilization. Theoretical results are presented to demonstrate effectiveness (involving fixed-time convergence and robustness) of the proposed computing schemes for the time-variant QP problem solving.

eess.SY↗

arXiv2024

A Multimodal Fusion Network For Student Emotion Recognition Based on Transformer and Tensor Product

This paper introduces a new multi-modal model based on the Transformer architecture and tensor product fusion strategy, combining BERT's text vectors and ViT's image vectors to classify students' psychological conditions, with an accuracy of 93.65%. The purpose of the study is to accurately analyze the mental health status of students from various data sources. This paper discusses modal fusion methods, including early, late and intermediate fusion, to overcome the challenges of integrating multi-modal information. Ablation studies compare the performance of different models and fusion techniques, showing that the proposed model outperforms existing methods such as CLIP and ViLBERT in terms of accuracy and inference speed. Conclusions indicate that while this model has significant advantages in emotion recognition, its potential to incorporate other data modalities provides areas for future research.

cs.CV↗

arXiv2024

CRTRE: Causal Rule Generation with Target Trial Emulation Framework

Causal inference and model interpretability are gaining increasing attention, particularly in the biomedical domain. Despite recent advance, decorrelating features in nonlinear environments with human-interpretable representations remains underexplored. In this study, we introduce a novel method called causal rule generation with target trial emulation framework (CRTRE), which applies randomize trial design principles to estimate the causal effect of association rules. We then incorporate such association rules for the downstream applications such as prediction of disease onsets. Extensive experiments on six healthcare datasets, including synthetic data, real-world disease collections, and MIMIC-III/IV, demonstrate the model's superior performance. Specifically, our method achieved a $β$ error of 0.907, outperforming DWR (1.024) and SVM (1.141). On real-world datasets, our model achieved accuracies of 0.789, 0.920, and 0.300 for Esophageal Cancer, Heart Disease, and Cauda Equina Syndrome prediction task, respectively, consistently surpassing baseline models. On the ICD code prediction tasks, it achieved AUC Macro scores of 92.8 on MIMIC-III and 96.7 on MIMIC-IV, outperforming the state-of-the-art models KEPT and MSMN. Expert evaluations further validate the model's effectiveness, causality, and interpretability.

cs.LG↗

arXiv2024

DERs-Aided Blackstart and Load Restoration Framework for Distribution Systems Considering Synchronization and Frequency Security Constraints

Extreme weather events have led to long-duration outages in the distribution system (DS), necessitating novel approaches to blackstart and restore the system. Existing blackstart solutions utilize blackstart units to establish multiple microgrids, sequentially energize non-blackstart units, and restore loads. However, these approaches often result in isolated microgrids. In DER-aided blackstart, the continuous operation of these microgrids is uncertain due to the finite energy capacity of commonly used blackstart units, such as battery energy storage (BES)-based grid-forming inverters (GFMIs). To address this issue, this article proposes a holistic blackstart and restoration framework that incorporates synchronization between microgrids and the entire DS with the transmission grid (TG). To support synchronization, we leveraged virtual synchronous generator-based control for GFMIs to estimate their frequency response to load pick-up events using only initial/final quasi-steady-state points. Subsequently, a synchronization switching condition was developed to model synchronizing switches, aligning them seamlessly with a linearized branch flow problem. Finally, we designed a bottom-up blackstart and restoration framework that considers the switching structure of the DS, energizing/synchronizing switches, DERs with grid-following inverters, and BES-based GFMIs with frequency security constraints. The proposed framework is validated in IEEE-123-bus system, considering cases with two and four GFMIs under various TG recovery instants.

eess.SY↗

arXiv2024

ALI-Agent: Assessing LLMs' Alignment with Human Values via Agent-based Evaluation

Large Language Models (LLMs) can elicit unintended and even harmful content when misaligned with human values, posing severe risks to users and society. To mitigate these risks, current evaluation benchmarks predominantly employ expert-designed contextual scenarios to assess how well LLMs align with human values. However, the labor-intensive nature of these benchmarks limits their test scope, hindering their ability to generalize to the extensive variety of open-world use cases and identify rare but crucial long-tail risks. Additionally, these static tests fail to adapt to the rapid evolution of LLMs, making it hard to evaluate timely alignment issues. To address these challenges, we propose ALI-Agent, an evaluation framework that leverages the autonomous abilities of LLM-powered agents to conduct in-depth and adaptive alignment assessments. ALI-Agent operates through two principal stages: Emulation and Refinement. During the Emulation stage, ALI-Agent automates the generation of realistic test scenarios. In the Refinement stage, it iteratively refines the scenarios to probe long-tail risks. Specifically, ALI-Agent incorporates a memory module to guide test scenario generation, a tool-using module to reduce human labor in tasks such as evaluating feedback from target LLMs, and an action module to refine tests. Extensive experiments across three aspects of human values--stereotypes, morality, and legality--demonstrate that ALI-Agent, as a general evaluation framework, effectively identifies model misalignment. Systematic analysis also validates that the generated test scenarios represent meaningful use cases, as well as integrate enhanced measures to probe long-tail risks. Our code is available at https://github.com/SophieZheng998/ALI-Agent.git

cs.AI↗

arXiv2024

Kernel Looping: Eliminating Synchronization Boundaries for Peak Inference Performance

Token generation speed is critical to power the next wave of AI inference applications. GPUs significantly underperform during token generation due to synchronization overheads at kernel boundaries, utilizing only 21% of their peak memory bandwidth. While recent dataflow architectures mitigate these overheads by enabling aggressive fusion of decoder layers into a single kernel, they too leave performance on the table due to synchronization penalties at layer boundaries. This paper presents kernel looping, a specialized global optimization technique which exploits an optimization opportunity brought by combining the unique layer-level fusion possible in modern dataflow architectures with the repeated layer structure found in language models. Kernel looping eliminates synchronization costs between consecutive calls to the same kernel by transforming these calls into a single call to a modified kernel containing a pipelined outer loop. We evaluate kernel looping on the SambaNova SN40L Reconfigurable Dataflow Unit (RDU), a commercial dataflow accelerator for AI. Experiments demonstrate that kernel looping speeds up the decode phase of a wide array of powerful open-source models by up to 2.2$\times$ on SN40L. Kernel looping allows scaling of decode performance over multiple SN40L sockets, achieving speedups of up to 2.5$\times$. Finally, kernel looping enables SN40L to achieve over 90% of peak performance on 8 and 16 sockets and achieve a speedup of up to 3.7$\times$ over DGX H100. Kernel looping, as well as the models evaluated in this paper, are deployed in production in a commercial AI inference cloud.

cs.CL↗

arXiv2024

Improving Pinterest Search Relevance Using Large Language Models

To improve relevance scoring on Pinterest Search, we integrate Large Language Models (LLMs) into our search relevance model, leveraging carefully designed text representations to predict the relevance of Pins effectively. Our approach uses search queries alongside content representations that include captions extracted from a generative visual language model. These are further enriched with link-based text data, historically high-quality engaged queries, user-curated boards, Pin titles and Pin descriptions, creating robust models for predicting search relevance. We use a semi-supervised learning approach to efficiently scale up the amount of training data, expanding beyond the expensive human labeled data available. By utilizing multilingual LLMs, our system extends training data to include unseen languages and domains, despite initial data and annotator expertise being confined to English. Furthermore, we distill from the LLM-based model into real-time servable model architectures and features. We provide comprehensive offline experimental validation for our proposed techniques and demonstrate the gains achieved through the final deployed system at scale.

cs.IR↗

arXiv2024

REGNet V2: End-to-End REgion-based Grasp Detection Network for Grippers of Different Sizes in Point Clouds

Grasping has been a crucial but challenging problem in robotics for many years. One of the most important challenges is how to make grasping generalizable and robust to novel objects as well as grippers in unstructured environments. We present \regnet, a robotic grasping system that can adapt to different parallel jaws to grasp diversified objects. To support different grippers, \regnet embeds the gripper parameters into point clouds, based on which it predicts suitable grasp configurations. It includes three components: Score Network (SN), Grasp Region Network (GRN), and Refine Network (RN). In the first stage, SN is used to filter suitable points for grasping by grasp confidence scores. In the second stage, based on the selected points, GRN generates a set of grasp proposals. Finally, RN refines the grasp proposals for more accurate and robust predictions. We devise an analytic policy to choose the optimal grasp to be executed from the predicted grasp set. To train \regnet, we construct a large-scale grasp dataset containing collision-free grasp configurations using different parallel-jaw grippers. The experimental results demonstrate that \regnet with the analytic policy achieves the highest success rate of $74.98\%$ in real-world clutter scenes with $20$ objects, significantly outperforming several state-of-the-art methods, including GPD, PointNetGPD, and S4G. The code and dataset are available at https://github.com/zhaobinglei/REGNet-V2.

cs.RO↗

arXiv2024

Data-Driven Stable Neural Feedback Loop Design

This paper proposes a data-driven approach to design a feedforward Neural Network (NN) controller with a stability guarantee for plants with unknown dynamics. We first introduce data-driven representations of stability conditions for Neural Feedback Loops (NFLs) with linear plants, which can be formulated into a semidefinite program (SDP). Subsequently, this SDP constraint is integrated into the NN training process to ensure stability of the feedback loop. The whole NN controller design problem can be solved by an iterative algorithm. Finally, we illustrate the effectiveness of the proposed method compared to model-based methods via numerical examples.

math.OC↗

arXiv2024

The next step in galaxy cluster strong lensing: modeling the surface brightness of multiply-imaged sources

Overcoming both modeling and computational challenges, we present, for the first time, the extended surface-brightness distribution model of a strongly-lensed source in a complex galaxy-cluster-scale system. We exploit the high-resolution Hubble Space Telescope (HST) imaging and extensive Multi Unit Spectroscopic Explorer spectroscopy to build an extended strong-lensing model, in a full multi-plane formalism, of SDSS J1029+2623, a lens cluster at $z = 0.588$ with three multiple images of a background quasar ($z = 2.1992$). Going beyond typical cluster strong-lensing modeling techniques, we include as observables both the positions of 26 pointlike multiple images from seven background sources, spanning a wide redshift range between 1.02 and 5.06, and the extended surface-brightness distribution of the strongly-lensed quasar host galaxy, over $\sim78000$ HST pixels. In addition, we model the light distribution of seven objects, angularly close to the strongly-lensed quasar host, over $\sim9300$ HST pixels. Our extended lens model reproduces well both the observed intensity and morphology of the quasar host galaxy in the HST F160W band (with a 0''.03 pixel scale). The reconstructed source shows a single, compact, and smooth surface-brightness distribution, for which we estimate an intrinsic magnitude of 23.3 $\pm$ 0.1 in the F160W band and a half-light radius of (2.39 $\pm$ 0.03) kpc. The increased number of observables enables the accurate determination of the total mass of line-of-sight halos lying angularly close to the extended arc. This work paves the way for a new generation of galaxy cluster strong-lens models, where additional, complementary lensing observables are directly incorporated as model constraints.

astro-ph.GA↗

arXiv2024

Bridging OOD Detection and Generalization: A Graph-Theoretic View

In the context of modern machine learning, models deployed in real-world scenarios often encounter diverse data shifts like covariate and semantic shifts, leading to challenges in both out-of-distribution (OOD) generalization and detection. Despite considerable attention to these issues separately, a unified framework for theoretical understanding and practical usage is lacking. To bridge the gap, we introduce a graph-theoretic framework to jointly tackle both OOD generalization and detection problems. By leveraging the graph formulation, data representations are obtained through the factorization of the graph's adjacency matrix, enabling us to derive provable error quantifying OOD generalization and detection performance. Empirical results showcase competitive performance in comparison to existing methods, thereby validating our theoretical underpinnings. Code is publicly available at https://github.com/deeplearning-wisc/graph-spectral-ood.

cs.LG↗

arXiv2024

Scattering of surface waves by ocean currents: the U2H map

Ocean turbulence at meso- and submesocales affects the propagation of surface waves through refraction and scattering, inducing spatial modulations in significant wave height (SWH). We develop a theoretical framework that relates these modulations to the current that induces them. We exploit the asymptotic smallness of the ratio of typical current speed to wave group speed to derive a linear map -- the U2H map -- between surface current velocity and SWH anomaly. The U2H map is a convolution, non-local in space, expressible as a product in Fourier space by a factor independent of the magnitude of the wavenumber vector. Analytic expressions of the U2H map show how the SWH responds differently to the vortical and divergent parts of the current, and how the anisotropy of the wave spectrum is key to large current-induced SWH anomalies. We implement the U2H map numerically and test its predictions against WAVEWATCH III numerical simulations for both idealised and realistic current configurations.

physics.flu-dyn↗

arXiv2024

Enhancing Visual Question Answering through Ranking-Based Hybrid Training and Multimodal Fusion

Visual Question Answering (VQA) is a challenging task that requires systems to provide accurate answers to questions based on image content. Current VQA models struggle with complex questions due to limitations in capturing and integrating multimodal information effectively. To address these challenges, we propose the Rank VQA model, which leverages a ranking-inspired hybrid training strategy to enhance VQA performance. The Rank VQA model integrates high-quality visual features extracted using the Faster R-CNN model and rich semantic text features obtained from a pre-trained BERT model. These features are fused through a sophisticated multimodal fusion technique employing multi-head self-attention mechanisms. Additionally, a ranking learning module is incorporated to optimize the relative ranking of answers, thus improving answer accuracy. The hybrid training strategy combines classification and ranking losses, enhancing the model's generalization ability and robustness across diverse datasets. Experimental results demonstrate the effectiveness of the Rank VQA model. Our model significantly outperforms existing state-of-the-art models on standard VQA datasets, including VQA v2.0 and COCO-QA, in terms of both accuracy and Mean Reciprocal Rank (MRR). The superior performance of Rank VQA is evident in its ability to handle complex questions that require understanding nuanced details and making sophisticated inferences from the image and text. This work highlights the effectiveness of a ranking-based hybrid training strategy in improving VQA performance and lays the groundwork for further research in multimodal learning methods.

cs.CV↗

arXiv2024

Active robustness against the detuning-error for Rydberg quantum gates

Error suppression to the experimental imperfections is a central challenge for useful quantum computing. Recent studies have shown the advantages of using single-modulated pulses based on optimal control which can realize high-fidelity two-qubit gates in neutral-atom arrays. However, typical optimization only minimizes the ideal gate error in the absence of any decay, which allows the gate to be passively influenced by all error sources leading to an exponential increase of sensitivity when the error becomes larger. In the present work, we propose the realization of two-qubit CZ gates with active robustness against two-photon detuning errors. Our method depends on a modified cost function in numerical optimization for shaping gate pulses, which can minimize, not only the ideal gate error but also the fluctuations of gate infidelity over a wide error range. We introduce a family of Rydberg blockade gates with active robustness towards the impacts of versatile noise sources such as Doppler dephasing and ac Stark shifts. The resulting gates with robust pulses can significantly increase the insensitivity to any type of errors acting on the two-photon detuning, benefiting from a relaxed requirement of colder atomic temperatures or more stable lasers for current experimental technology.

quant-ph↗